0.838273	-8.03218	either | or
0.336142	-8.46853	both | and
0.514212	-8.73898	more | than
0.600743	-8.79049	source | target
0.655213	-8.85977	positive | negative
0.335286	-9.02588	such as | and
0.367906	-9.13069	word | word
0.93871	-9.14109	(a) | (b)
0.777778	-9.20879	at | level
0.449216	-9.25419	not | but
0.415704	-9.33821	features | features
0.610368	-9.34523	precision | recall
0.40933	-9.34742	words | words
0.736181	-9.37963	first | then
0.636187	-9.41556	(1) | (2)
0.375262	-9.44726	and | respectively.
0.391482	-9.52136	can be | by
0.33783	-9.52625	– | –
0.562172	-9.55516	sentences | sentences
0.899543	-9.57477	left | right
0.67013	-9.59891	other | such as
0.397612	-9.60315	the same | as
0.775801	-9.62266	(2) | (3)
0.389091	-9.63718	number of | number
0.641089	-9.64126	extract | from
0.930851	-9.65683	at | level,
0.409326	-9.66768	if | then
0.314917	-9.6685	model | model
0.970414	-9.67836	(i) | (ii)
0.40102	-9.69208	translation | translation
0.415011	-9.70158	feature | feature
0.656069	-9.74786	context | context
0.448509	-9.75164	was | was
0.854271	-9.77006	does not | any
0.571749	-9.77108	topic | topic
0.853535	-9.77624	at | time,
0.245409	-9.7817	+ | +
0.409091	-9.78506	system | system
0.623978	-9.78835	syntactic | semantic
0.468413	-9.79459	different | different
0.789474	-9.79647	one | another
0.655488	-9.80337	text | text
0.731518	-9.82783	extracting | from
0.496403	-9.83212	event | event
0.577017	-9.83769	better | than
0.632047	-9.8508	relation | relation
0.597855	-9.86072	only | also
0.321429	-9.86187	where | are
0.285185	-9.87081	can | be
0.83871	-9.87949	training | testing
0.448331	-9.91088	whether | or
0.755656	-9.91377	treat | as
0.581081	-9.9215	need | be
0.62069	-9.94035	higher | than
0.492095	-9.94142	one | other
0.273394	-9.94727	can be | as
0.659498	-9.95479	document | document
0.954545	-9.95601	neither | nor
0.396846	-9.96415	sentence | sentence
0.512304	-9.98504	have | been
0.411594	-9.98897	phrase | phrase
0.328096	-9.99419	i | i
0.346392	-9.99544	language | language
0.644928	-10.0065	alignment | alignment
0.883562	-10.0106	4 | 4.1
0.834356	-10.0208	part | speech
0.410448	-10.0272	semantic | semantic
0.544974	-10.0281	and | respectively,
0.397472	-10.0282	one | one
0.557423	-10.0405	score | score
0.863946	-10.0493	at | time.
0.59164	-10.0619	the first | second
0.571429	-10.0732	features | features.
0.604811	-10.0841	pairs | pairs
0.221657	-10.0874	these | are
0.981651	-10.0906	state | art
0.300514	-10.0948	and | them
0.271388	-10.0974	• | •
0.30477	-10.0982	data | data
0.919355	-10.0985	precision | recall.
0.26699	-10.1206	or | or
0.546784	-10.1238	verb | verb
0.578947	-10.1257	new | new
0.295377	-10.1271	including | and
0.48046	-10.1402	its | its
0.423488	-10.1405	has | been
0.559375	-10.1462	node | node
0.803922	-10.152	input | output
0.223058	-10.1556	based on | and
0.832168	-10.1576	all | except
0.35567	-10.1651	more | more
0.962264	-10.1671	play | role
0.525568	-10.1726	figure 2: | of
0.416071	-10.1782	corpus | corpus
0.579861	-10.1784	query | query
0.560261	-10.1814	baseline | baseline
0.392628	-10.1861	proceedings of the | june
0.374636	-10.1862	apply | to
0.383077	-10.1926	may | be
0.215362	-10.1933	based on | of
0.511111	-10.2095	annotation | annotation
0.192414	-10.2115	a | set of
0.835821	-10.2134	(och | ney,
0.610442	-10.2184	algorithm | algorithm
0.497326	-10.2211	assign | to
0.234663	-10.2252	was | by
0.802817	-10.2299	share | same
0.407609	-10.2326	let | be
0.209069	-10.2365	when | is
0.743902	-10.241	labeled | unlabeled
0.586207	-10.2572	rule | rule
0.211984	-10.2622	all | are
0.375594	-10.2625	we use | as
0.445434	-10.2628	english | english
0.389365	-10.2703	in | case,
0.788732	-10.2716	high | low
0.573034	-10.2754	with | highest
0.452214	-10.2778	, | .
0.81203	-10.2793	let | denote
0.579151	-10.2834	number | times
0.473008	-10.2861	compare | with
0.465	-10.2911	models | models
0.513846	-10.2956	at | time
0.641148	-10.2965	reordering | reordering
0.656566	-10.3006	class | class
0.632075	-10.3075	at | level.
0.613333	-10.315	whose | are
0.553114	-10.3232	focus | on
0.45122	-10.3285	mapping | to
0.864865	-10.3311	incorporate | into
0.226044	-10.3323	it is | that
0.321696	-10.3325	, | ,
0.519608	-10.3334	noun | noun
0.607143	-10.3337	label | label
0.779412	-10.3366	in | setting,
0.398833	-10.3498	all | were
0.593074	-10.3501	in | way.
0.151575	-10.3502	where | is
0.286437	-10.3558	s | s
0.836207	-10.3559	replacing | with
0.71519	-10.3585	improve | performance
0.251962	-10.3607	in | context
0.717949	-10.366	department | computer science
0.560784	-10.3679	allowing | to
0.46594	-10.37	applying | to
0.392996	-10.3748	in | cases
0.438725	-10.3853	information | information
0.385633	-10.3874	consists of | and
0.785714	-10.3907	in | ways.
0.709677	-10.3957	instances | instances
0.383019	-10.3964	table 1: | of
0.322237	-10.3969	joint | and
0.264865	-10.3985	in | case
0.684848	-10.3988	value | value
0.790323	-10.3993	subject | object
0.492163	-10.4062	lexical | lexical
0.795082	-10.4069	3 | 3.1
0.306373	-10.412	all | all
0.392354	-10.4149	provide | for
0.45961	-10.4233	tree | tree
0.730496	-10.4241	following | 1.
0.703947	-10.4269	assigning | to
0.552846	-10.4324	less | than
0.852941	-10.4385	at | beginning
0.498328	-10.4443	sense | sense
0.295294	-10.447	we | them
0.658824	-10.4488	length | length
0.452778	-10.4502	target | source
0.539683	-10.4506	how | can be
0.404922	-10.4551	much | than
0.688312	-10.4572	needs | be
0.683871	-10.4643	in | literature
0.537849	-10.4665	candidate | candidate
0.402685	-10.468	words | their
0.76	-10.4708	first | last
0.576744	-10.4758	we define | as
0.574074	-10.4832	negative | positive
0.419753	-10.4832	adding | to
0.477564	-10.4841	with | without
0.212834	-10.4864	the | described in
0.691275	-10.4864	seems | be
0.3	-10.4866	has been | in
0.37827	-10.4866	x | y
0.771186	-10.4917	replace | with
0.527559	-10.4934	evaluation | evaluation
0.410628	-10.5022	if | then | is
0.452941	-10.5026	a | version of
0.449275	-10.5091	a | called
0.785714	-10.51	for | purposes
0.590909	-10.5115	about | about
0.492958	-10.513	learning | learning
0.734375	-10.5178	anonymous reviewers | their
0.451039	-10.5194	natural | processing
0.636905	-10.5241	grammar | grammar
0.585	-10.5253	syntactic | syntactic
0.373469	-10.5273	one | more
0.475083	-10.5341	(e.g., | (e.g.,
0.849462	-10.5349	global | local
0.538462	-10.5355	recall | precision
0.125058	-10.5371	null | null
0.354478	-10.5418	the | provided by
0.666667	-10.545	in | step,
0.747899	-10.5479	we refer to | as
0.480836	-10.554	sentiment | sentiment
0.934211	-10.5572	plays | role
0.428969	-10.5572	they | they
0.634146	-10.5577	sentences | sentences.
0.767857	-10.559	treating | as
0.237245	-10.5597	in | case | the
0.385488	-10.5671	i | j
0.829787	-10.5726	on | side,
0.770642	-10.5726	if | otherwise.
0.822917	-10.574	general | specific
0.275379	-10.5766	not | any
0.249042	-10.5796	1 | 2 | 3
0.645161	-10.5803	take | account
0.535398	-10.5804	accuracy | accuracy
0.436578	-10.5828	they | their
0.705426	-10.5842	on | scale
0.518828	-10.5885	opinion | opinion
0.252736	-10.5898	due to | of
0.615385	-10.5913	combining | with
0.691729	-10.5922	term | term
0.286632	-10.5925	since | are
0.768519	-10.5936	category | category
0.314241	-10.5937	were | were
0.54717	-10.5962	parsing | parsing
0.568528	-10.5968	english | chinese
0.334507	-10.5969	p | p
0.29805	-10.5978	t | t
0.440367	-10.5992	significantly | than
0.778846	-10.5993	conclusions | future work
0.792079	-10.6036	figure | illustrates
0.55665	-10.6036	vector | vector
0.407407	-10.6109	k | k
0.897436	-10.6118	start | end
0.427326	-10.6124	≤ | ≤
0.918919	-10.6141	convert | into
0.364807	-10.6205	have | same
0.0956586	-10.6226	the | between | and
0.276543	-10.6242	in | form
0.426866	-10.636	each | pair
0.572193	-10.6379	distribution | distribution
0.107311	-10.6382	the | can be
0.984127	-10.6386	2. | 3. | 4.
0.535211	-10.6421	some | some
0.244597	-10.6422	include | and
0.620253	-10.6422	classification | classification
0.894737	-10.6432	at | same time
0.315876	-10.6433	between | two
0.312903	-10.6435	and | 2010)
0.350202	-10.6475	a | amount of
0.470588	-10.649	d | d
0.433022	-10.6527	bleu | bleu
0.820225	-10.6577	several | including
0.56383	-10.659	referred | as
0.619355	-10.6591	terms | terms
0.431677	-10.6591	this | due to
0.223993	-10.6594	while | is
0.927536	-10.6594	integrate | into
0.603659	-10.6623	in | cases.
0.381418	-10.6653	their | their
0.55102	-10.6659	assigns | to
0.715517	-10.6679	however, | still
0.377724	-10.6725	similarity | similarity
0.588235	-10.6731	enables | to
0.331471	-10.6756	training | test
0.44186	-10.6792	extend | to
0.360987	-10.6853	these | were
0.631944	-10.6929	bilingual | bilingual
0.526316	-10.6929	parallel | parallel
0.389474	-10.6933	v | v
0.314433	-10.6987	allow | to
0.556757	-10.6996	noun | verb
0.368171	-10.7033	have been | for
0.857143	-10.7055	5 | 5.1
0.362587	-10.7072	test | test
0.193548	-10.7092	we use | to
0.397222	-10.7104	whose | is
0.402857	-10.7138	∈ | ∈
0.75	-10.7142	articles | articles
0.635714	-10.715	order | order
0.590062	-10.717	time | time
0.479508	-10.7182	search | search
0.316163	-10.7182	includes | and
0.265	-10.7219	proceedings of the | on
0.644444	-10.7227	topics | topics
0.561798	-10.724	lexical | syntactic
0.53	-10.7245	to | extent
0.835443	-10.7304	second | third
0.37659	-10.7313	see | for
0.875	-10.7323	transform | into
0.487179	-10.7333	effect | on
0.287202	-10.7344	the | process.
0.266325	-10.7351	of | natural language
0.592357	-10.7357	pair | pair
0.16576	-10.7363	the | used in
0.809524	-10.7376	in | ways:
0.567251	-10.738	tokens | tokens
0.203593	-10.7393	= | where | is
0.639098	-10.7471	events | events
0.0925261	-10.7479	between | and
0.130885	-10.7518	the | based on
0.451493	-10.752	work | et al.
0.37931	-10.7547	task | task
0.769231	-10.7551	takes | input
0.438163	-10.7568	f | f
0.681034	-10.7593	reference | reference
0.641221	-10.7599	input | input
0.517413	-10.7624	average | average
0.703704	-10.7633	meaning | meaning
0.31016	-10.7635	two | two
0.363636	-10.7639	as follows: | and
0.406154	-10.7688	table 3: | of
0.540984	-10.7718	extraction | extraction
0.551136	-10.7726	chinese | english
0.38764	-10.774	could | be
0.734694	-10.777	2 | 2.1
0.5	-10.7777	it | difficult
0.355609	-10.778	defined as | of
0.571429	-10.7808	head | head
0.661157	-10.7814	paraphrase | paraphrase
0.192551	-10.7844	an | an
0.828947	-10.7855	with | exception
0.758242	-10.7861	weight | weight
0.440741	-10.7872	following | •
0.197339	-10.7878	x | x
0.141714	-10.7884	a | its
0.302977	-10.7902	note that | is
0.457831	-10.7908	since | only
0.666667	-10.7918	segmentation | segmentation
0.373333	-10.7965	note that | are
0.228686	-10.799	were | by
0.488479	-10.7995	conference | empirical | in
0.545977	-10.7999	bleu | ter
0.362025	-10.8	add | to
0.356265	-10.8029	source | source
0.350711	-10.8029	w | w
0.770115	-10.8046	unsupervised | supervised
0.497608	-10.8059	figure 1: | for
0.382436	-10.8064	entity | entity
0.755556	-10.8071	pattern | pattern
0.459016	-10.8076	we apply | to
0.479821	-10.808	figure 3: | of
0.208263	-10.8092	each | each
0.253117	-10.8128	each | its
0.337778	-10.8128	table 2: | of
0.981132	-10.8133	special | group
0.327083	-10.8147	dependency | dependency
0.3575	-10.8159	should | be
0.776471	-10.8159	for | reasons.
0.518519	-10.8187	indicates that | is
0.807692	-10.8202	translating | into
0.642276	-10.8204	sequence | words
0.54386	-10.8224	we | expect
0.485981	-10.824	word | word.
0.252513	-10.8263	r | r
0.356962	-10.8296	proceedings | workshop on
0.764706	-10.8336	illustrated | figure
0.639344	-10.8343	labels | labels
0.405229	-10.836	since | have
0.395639	-10.8366	over | over
0.678899	-10.8368	but | still
0.465217	-10.8373	more | less
0.399361	-10.8375	can be | into
0.758621	-10.8384	concept | concept
0.663717	-10.8408	similar | similar
0.962264	-10.8412	(c) | (d)
0.168565	-10.8415	the | does not
0.587413	-10.8472	adds | to
0.196276	-10.8472	since | is
0.875	-10.8489	positive | negative.
0.777778	-10.8498	string | string
0.794872	-10.8537	(4) | (5)
0.875	-10.8543	acl-hlt | usa,
0.485577	-10.8566	map | to
0.528409	-10.8569	in | experiment,
0.574324	-10.8591	graph | graph
0.376812	-10.8605	and | 2007)
0.624	-10.8626	u | u
0.649123	-10.8691	errors | errors
0.260504	-10.8695	training | training
0.738636	-10.8733	count | count
0.67619	-10.8742	unigram | bigram
0.438247	-10.8753	# | #
0.51087	-10.879	any | any
0.591241	-10.88	one | per
0.657658	-10.8802	list | list
0.559211	-10.882	users | their
0.960784	-10.885	sigdial | discourse
0.554839	-10.8852	respectively | and
0.32043	-10.887	were | using
0.651786	-10.8871	structure | structure
0.616	-10.8877	values | values
0.69	-10.8903	texts | texts
0.25884	-10.8906	combination | and
0.654545	-10.8916	languages | languages
0.473934	-10.8917	on | dataset
0.717391	-10.892	trigger | trigger
0.573427	-10.8931	classifier | classifier
0.0720869	-10.8948	the | et al.,
0.582734	-10.8949	for | extraction.
0.338164	-10.8958	in addition, | of
0.443515	-10.8961	like | or
0.146978	-10.8969	from | from
0.455752	-10.8971	at | end
0.673077	-10.8972	clustering | clustering
0.846154	-10.8986	stanford | manning,
0.251682	-10.8996	whether | is
0.540373	-10.9023	parser | parser
0.315011	-10.9032	a | containing
0.252033	-10.9058	if | if
0.387097	-10.9072	will | will
0.387097	-10.9084	and | 2008)
0.240695	-10.9091	we use | for
0.459091	-10.9098	o | o
0.446352	-10.91	such | can be
0.786667	-10.9111	best of | knowledge,
0.492147	-10.9122	train | model
0.186012	-10.9123	in | such as
0.590909	-10.9132	proposed | method
0.228766	-10.9184	however, | not
0.534161	-10.9197	patterns | patterns
0.724138	-10.9199	average number | per
0.823529	-10.9225	6th | 30–31,
0.252078	-10.9229	while | are
0.242038	-10.9232	the | described | section
0.248656	-10.9248	most | are
0.287523	-10.9252	using | features
0.267296	-10.932	will | be
0.494624	-10.9327	an important | in
0.25578	-10.9365	c | c
0.392491	-10.9369	on | side
0.583333	-10.9403	corpus | million
0.36858	-10.9408	probability | probability
0.349593	-10.9417	the | proposed by
0.275804	-10.943	in particular, | the
0.369697	-10.943	when | when
0.370031	-10.9441	provides | for
0.907407	-10.9468	precision | recall,
0.124913	-10.9469	the | of | given
1	-10.947	2011: | 17-18,
1	-10.947	12th | dialogue,
0.452055	-10.9471	best | best
0.171018	-10.9495	is | to be
0.531646	-10.9501	human | human
0.84127	-10.9517	concepts | concepts
0.740741	-10.9541	extent | which
0.324582	-10.959	4 | 5
0.443946	-10.962	user | user
0.585938	-10.9621	french | english
0.592	-10.9671	before | after
0.37785	-10.9693	|
0.288973	-10.9703	features | feature
0.977778	-10.9709	point | view,
0.7375	-10.9725	it is | clear
0.66	-10.9736	mention | mention
0.653465	-10.9756	side | side
0.220358	-10.9789	but | not
0.673684	-10.9796	1: | 2:
0.888889	-10.9805	t1 | t2
0.74359	-10.9818	in | fashion
0.663265	-10.9825	second | first
0.201311	-10.9835	when | are
0.647059	-10.9853	for | machine translation.
0.272414	-10.9889	the | contained
0.505952	-10.9894	for | languages,
0.353801	-10.9895	— | —
0.774648	-10.9897	support | machines
0.174066	-10.9898	for example, | a
0.398524	-10.9904	found | be
0.417004	-10.991	sections | and
0.380471	-10.9911	for | languages.
0.8125	-10.9914	c1 | c2
0.42735	-10.9917	m | m
0.290196	-10.9929	and | (2010)
0.525974	-10.993	on | dataset.
0.80303	-11.0019	thought | as
0.603448	-11.0019	0.2 | 0.8
0.162414	-11.003	the | task.
0.37013	-11.0033	and | 2005)
0.381443	-11.0039	similarity between | and
0.725	-11.0042	combines | with
0.519231	-11.0045	results | reported
0.208633	-11.005	training data | the
0.29	-11.0085	and | 2009)
0.37884	-11.0129	for | task,
0.526667	-11.0168	topic | topics
0.560606	-11.0168	document | documents
0.485876	-11.0173	words, | words
0.434389	-11.0174	in | setting
0.485876	-11.0176	features, | features
0.310905	-11.0189	a | framework
0.437788	-11.0201	on | basis
0.0925508	-11.0212	the | to be
0.730769	-11.0216	instance | instance
0.473118	-11.0219	in | order.
0.142857	-11.0235	is | number of
0.819672	-11.0248	metrics | metrics
0.134523	-11.0254	a | number of
0.530612	-11.0261	in | field
0.736842	-11.0264	in | stage,
0.918367	-11.03	beyond | scope
0.571429	-11.0307	wikipedia | wikipedia
0.616822	-11.0346	verbs | verbs
0.161087	-11.0347	for example, | and
0.191756	-11.0347	set | set
0.50625	-11.035	table 2: | on
0.649485	-11.0366	do not | they
0.616822	-11.0393	temporal | temporal
0.37037	-11.0394	sentence | sentences
0.64	-11.0395	decoding | decoding
0.421739	-11.0417	error | error
0.611111	-11.042	question | question
0.625	-11.0424	output | output
0.191031	-11.0449	are | their
0.596491	-11.0457	like | like
0.916667	-11.0464	at | levels.
0.627451	-11.0473	work | done
0.246617	-11.0493	the | algorithm.
0.75	-11.0496	improves | performance
0.440191	-11.0503	namely | and
0.527778	-11.0505	events | event
0.723684	-11.052	inference | inference
0.219309	-11.053	used | et al.,
0.175305	-11.0538	is | it is
0.816667	-11.0545	mitchell | lapata
0.784615	-11.0552	matching | matching
0.332418	-11.0557	and | 2006)
0.531915	-11.0559	2010; | 2010).
0.453608	-11.0561	and | (2010).
0.264808	-11.0583	feature | features
0.144572	-11.0605	of | such as
0.227684	-11.0609	in order to | a
0.639175	-11.061	developing | for
0.708861	-11.0621	higher | lower
0.72973	-11.0633	assumed | be
0.953488	-11.0637	forward | backward
1	-11.0647	klein | manning
0.851852	-11.0657	s1 | s2
0.276392	-11.067	word | its
0.220195	-11.0671	and | 2008).
0.682353	-11.0673	points | points
0.682353	-11.0674	selection | selection
0.290254	-11.0674	difference between | the
0.352201	-11.0701	use | information
0.639175	-11.0703	english | spanish
0.803279	-11.0706	in | way:
0.255776	-11.0719	the | (i.e.
0.97561	-11.0723	8: | 9:
0.206674	-11.0725	in | corpus.
0.682353	-11.0729	vectors | vectors
0.2512	-11.0742	the | corresponding to
0.652174	-11.0749	7 | 9
0.393701	-11.0762	from | web
0.395257	-11.0767	a | problem.
0.736111	-11.0769	taking | account
0.78125	-11.0789	(the | (the
0.354839	-11.0825	j | j
0.513514	-11.0827	limit | to
0.520833	-11.0837	sentences | contain
0.203175	-11.0849	the | i.e.
0.686747	-11.0851	nodes | edges
0.8	-11.0853	treats | as
0.691358	-11.0861	for | purpose of
0.645161	-11.0872	predicate | predicate
0.769231	-11.0876	correct | incorrect
0.345679	-11.0886	also | other
0.369718	-11.0891	this | has been
0.1251	-11.0891	the | corpus.
0.638298	-11.0919	automatic | automatic
0.769231	-11.092	extracts | from
0.207778	-11.0928	the | trained on
0.242792	-11.0933	of | according to
0.678571	-11.0934	manual | automatic
0.381132	-11.0942	refer to | as
0.521127	-11.0943	used | generate
0.215403	-11.0963	this | can be
0.244204	-11.0975	: | :
0.269303	-11.0982	et | 2009)
0.975	-11.0999	dirichlet | (lda)
0.8	-11.1	solve | problem
0.13569	-11.1003	can be | in
0.22807	-11.1006	n | n
0.537879	-11.1012	evaluate | performance
0.0773782	-11.1024	as | as
0.458564	-11.103	parse | parse
0.732394	-11.1052	every | every
0.631579	-11.1091	sense | senses
0.73913	-11.1094	at | cost
0.849057	-11.1094	incorporating | information
0.670588	-11.1096	multiple | multiple
0.296984	-11.1112	predictions | the
0.765625	-11.1125	select | best
0.930233	-11.1128	(pang | lee,
0.473373	-11.1146	cluster | cluster
0.473373	-11.1152	make | more
0.651685	-11.119	braceleftbigg | if
0.786885	-11.1205	improve | performance.
0.473054	-11.1209	top | top
0.712329	-11.123	improve | accuracy
0.37594	-11.1254	with | being
0.774194	-11.1258	6 conclusion | have
0.171228	-11.1263	like | and
0.386454	-11.1274	a | technique
0.446809	-11.1277	might | be
0.138407	-11.1281	1 | 2
0.818182	-11.129	for | sake
0.783333	-11.1313	depicted | figure
0.490323	-11.1328	for | parsing.
0.47561	-11.1381	english | french
0.455056	-11.1388	it | clear
0.455056	-11.139	while | only
0.450549	-11.1392	in | context,
0.30198	-11.1402	the | itself
0.71831	-11.1412	question | answer
0.195046	-11.1415	the | training data
0.621053	-11.1415	for | detection.
0.42233	-11.1454	size | size
0.260148	-11.1461	the | used.
0.0844143	-11.1461	is | and | is
0.119029	-11.1466	in order | the
0.490196	-11.1477	approach | approach
0.21916	-11.1483	→ | →
0.3663	-11.149	an | based on
0.284141	-11.1491	y | y
0.560345	-11.1493	note that | only
0.374046	-11.15	and | (2007)
0.675	-11.1523	multiple | single
0.742424	-11.1526	achieves | accuracy
0.560345	-11.1533	chinese | chinese
0.430769	-11.1542	separate | for
0.708333	-11.1563	small | large
0.322857	-11.1564	the | community
0.28833	-11.1569	for | languages
0.390756	-11.1579	a | distribution over
0.928571	-11.1579	first, | second,
0.215013	-11.1583	words | word
0.87234	-11.162	annual meeting | interest
0.714286	-11.1636	fewer | than
0.724638	-11.1636	by | annotators.
0.647059	-11.1636	is | challenging
0.555556	-11.164	tags | tags
0.758065	-11.1656	nouns | verbs
0.647059	-11.166	adapting | to
0.87234	-11.1665	inside | outside
0.233333	-11.1666	in | domain
0.84	-11.1674	shown | table 4.
0.383673	-11.1678	given | set of
0.12383	-11.1687	can be | a
0.758065	-11.1687	(papineni | 2002)
0.10352	-11.1695	are | by
0.349153	-11.1703	pos | pos
0.386555	-11.1716	and | 2004)
0.195722	-11.1728	in | et al.
0.536585	-11.1738	improves | by
0.160316	-11.1746	for each | in
0.0609427	-11.1753	an | of
0.854167	-11.1758	predicates | arguments
0.183712	-11.1769	directly | the
0.311475	-11.1773	tables | and
0.36	-11.1774	on | data,
0.527559	-11.1785	viewed | a
0.315642	-11.1792	and | 2004).
0.684211	-11.1798	question | how
0.811321	-11.1798	task | identifying
0.906977	-11.1799	subjective | objective
0.6625	-11.1828	low | high
0.318052	-11.1831	evaluate | on
0.358974	-11.1837	however, | only
0.379592	-11.1845	that | more than
0.5625	-11.1845	translated | translated
0.523438	-11.1855	combine | with
0.126531	-11.1863	we also | the
0.622222	-11.1874	edges | edges
0.493056	-11.1874	are | ones
0.308108	-11.1881	b | b
0.258004	-11.1883	made | of
0.972973	-11.1888	point | view
0.13714	-11.1906	three | and
0.775862	-11.1909	monolingual | monolingual
0.523438	-11.1929	vs. | vs.
0.632184	-11.1931	svm | svm
0.145884	-11.1937	does not | the
0.220527	-11.1953	each | one
0.694444	-11.1963	on | web.
0.34589	-11.1993	final | is
0.606383	-11.2009	within | same
0.642857	-11.201	achieve | performance
0.886364	-11.2011	instances | instances.
0.507463	-11.2011	function | function
0.823529	-11.2015	full | full
0.384615	-11.2028	would | would
0.972222	-11.2033	received | attention
0.367188	-11.2042	a | threshold
0.507463	-11.2044	relation | relations
0.694444	-11.2049	sets | sets
0.465409	-11.2056	references | and
0.51145	-11.2075	our | system,
0.671053	-11.208	output | input
0.0801187	-11.2087	we | our
0.229358	-11.2102	using | using
0.33121	-11.2108	same | different
0.12172	-11.2112	0 | 0
0.376033	-11.2118	a | approach.
0.330159	-11.2118	& | &
0.617978	-11.2132	minimum | training
0.866667	-11.2134	from a | perspective,
0.271552	-11.2135	the | one.
0.524194	-11.2138	not | enough
0.346154	-11.2148	perform | on
0.658228	-11.2153	measures | measures
0.342466	-11.2154	sentences | sentence
0.591837	-11.2154	edge | edge
0.496403	-11.2161	in | domain,
0.374486	-11.2168	the | itself.
0.347518	-11.2177	< | <
0.658228	-11.218	morphological | morphological
0.741935	-11.2189	transition | transition
0.572816	-11.2205	betwen | and
0.578431	-11.2217	in | future.
0.425532	-11.2221	2009; | 2010).
0.675676	-11.2225	specific | specific
0.173605	-11.2256	it | it
0.758621	-11.2282	remainder | paper
0.307263	-11.2285	same | same
0.30411	-11.2303	in | cases,
0.114608	-11.2313	by | by
0.52459	-11.2324	documents | document
0.902439	-11.2349	decompose | into
0.737705	-11.235	trained | tested
0.129289	-11.2355	a | can be
0.464516	-11.2357	features | features,
0.111316	-11.236	we can | the
0.630952	-11.2363	using | techniques
0.237374	-11.237	although | is
0.604396	-11.2371	sentences, | sentences
0.680556	-11.2371	achieves | performance
0.362205	-11.2374	in | language,
0.285024	-11.238	for instance, | a
0.567308	-11.2411	relations | relation
0.71875	-11.2412	achieves | best
0.264151	-11.2424	in | corpus,
0.389908	-11.2426	in | way,
0.428571	-11.2436	f1 | f1
0.902439	-11.2448	under | assumption
0.792453	-11.2482	performance | measured
0.167513	-11.2488	are | they
0.534483	-11.2489	score | scores
0.453416	-11.2491	tend | be
0.507812	-11.2493	each | represents
0.520661	-11.2499	even though | is
0.737705	-11.2504	probabilities | probabilities
0.0709413	-11.2506	from | to
0.180473	-11.2517	a | model.
0.290816	-11.2517	how | are
0.245009	-11.2532	participants | the
0.550459	-11.2538	we train | on
0.591398	-11.2546	which | turn
0.745763	-11.2547	metric | metric
0.179764	-11.2553	have been | and
0.671233	-11.2561	influence | on
0.733333	-11.2574	he | his
0.970588	-11.2575	recommendations | views
0.27191	-11.2584	each | has
0.489051	-11.2585	advantage of | that
0.201733	-11.2599	word | words
0.29443	-11.2606	annual meeting | association
0.619048	-11.264	table 2: | results
0.281553	-11.2655	to evaluate | of
0.237847	-11.266	applied | to
0.191441	-11.2662	goal | to
0.243194	-11.2678	this | was | by
0.463576	-11.2706	similar | but
0.219911	-11.2707	two | one
0.154185	-11.2723	a | approach
0.57	-11.2724	has | effect
0.452229	-11.2725	detecting | in
0.347015	-11.2736	the | hypothesis.
0.829787	-11.2738	at | step,
0.860465	-11.2746	japanese | japanese
0.8	-11.2748	as | increases.
0.4	-11.2757	for each | its
0.260504	-11.2765	are | those
0.69697	-11.277	regard | as
0.773585	-11.2771	splitting | into
0.609195	-11.2774	for | applications.
0.878049	-11.2776	prediction | prediction
0.226115	-11.2777	at | at
0.0540492	-11.2778	= | = | =
0.69697	-11.2791	bag | words
0.860465	-11.2815	primary | secondary
0.220211	-11.2821	represent | of
0.351351	-11.2827	let | be | and
0.826087	-11.2838	most | focused on
0.265351	-11.2839	the | obtained by
0.368644	-11.284	results | obtained
0.292225	-11.2844	a | component
0.198044	-11.2851	the | would be
0.741379	-11.2857	in | fashion.
0.141777	-11.2867	has been | the
0.512397	-11.2875	coreference | coreference
0.644737	-11.2877	and | (2002)
0.496124	-11.2892	methods | methods
0.358871	-11.2893	the | metric.
0.434524	-11.2901	then | these
0.292225	-11.2908	dialogue | dialogue
0.75	-11.291	our | learns
0.304985	-11.2913	the | feature.
0.360656	-11.2919	the | result.
0.260128	-11.2921	for | task.
0.657534	-11.2931	three | (1)
0.451613	-11.2952	enable | to
0.338182	-11.2954	is | source | target
0.150754	-11.2955	the | has been
0.588889	-11.2969	higher | higher
0.548077	-11.2978	is | this paper.
0.112625	-11.2989	number of | and
0.373333	-11.2993	5 | 6
0.537037	-11.2993	entities | entities
0.512605	-11.2993	run | on
0.243856	-11.3008	the | used by
0.0645119	-11.3017	the | was
0.75	-11.3019	synset | synset
0.736842	-11.304	in | article,
0.56	-11.3047	the | project.
0.408602	-11.307	1 introduction | are
0.38835	-11.3071	their | they
0.195122	-11.3073	these | are | to
0.542857	-11.3077	total | total
0.33574	-11.3111	although | not
0.385714	-11.3117	in | languages,
0.468085	-11.3135	rules | rule
0.464789	-11.3177	role | role
0.232112	-11.3181	this | because
0.296296	-11.3184	although | are
1	-11.3196	air | prime
0.323129	-11.3202	direct | of
0.96875	-11.3207	nlp | educational
0.78	-11.321	examples | examples.
0.372197	-11.3226	de | de
0.837209	-11.3233	adjectives | nouns
0.504132	-11.3236	multiple | same
0.246063	-11.3239	our | approach
0.496	-11.3244	phrases | phrase
0.189095	-11.3252	the | task,
0.593023	-11.3254	label | labels
0.157767	-11.3256	if | are
0.25261	-11.3258	would | be
0.360169	-11.3269	the | class,
0.445161	-11.327	only | single
0.569892	-11.3278	view | as
0.286863	-11.3288	10 | 10
0.822222	-11.3289	we compare | against
0.423529	-11.3307	and | (2005)
0.142193	-11.3359	the | contains
0.194067	-11.336	uses | to
0.484375	-11.3362	could be | with
0.364035	-11.3362	only | only
0.543689	-11.3365	nodes | node
0.631579	-11.3365	find | best
0.96875	-11.3368	5. | 6.
0.43949	-11.3369	corpus | annotated
0.764706	-11.3404	measure | quality
0.491935	-11.3411	that | training data.
0.286104	-11.3421	we | apply
0.219904	-11.3425	which | used
0.195929	-11.3429	both | are
0.586207	-11.3432	comparison between | and
0.740741	-11.3447	parenleftbigg | parenrightbigg
0.791667	-11.3455	divided | into
0.373832	-11.3462	then | into
0.484375	-11.3468	correlation | correlation
0.37799	-11.3493	on | level
0.818182	-11.3496	she | her
0.122024	-11.3502	is | et al.,
0.7	-11.3529	rules, | rules
0.488	-11.3536	translation | translations
0.647887	-11.3546	new | existing
0.362832	-11.3549	take | as
0.35443	-11.3555	method | method
0.188319	-11.3576	the | results.
0.661765	-11.3578	person | person
0.222591	-11.358	a | system.
0.166202	-11.3585	( | )
0.234317	-11.3589	many | are
0.495868	-11.3603	translations | translation
0.301538	-11.3607	the | process,
0.914286	-11.3609	at | point,
0.607595	-11.3618	(1) | (4)
0.68254	-11.363	models, | models
0.557895	-11.3641	arguments | arguments
0.7	-11.3648	documents | documents.
0.23133	-11.3656	the second | is
0.320557	-11.3669	increasing | of
0.433121	-11.3679	each | contains
0.939394	-11.3696	international conference | resources
0.272959	-11.3705	based on | from
0.483871	-11.3718	a | approximation
0.288136	-11.3727	similar to | in
0.534653	-11.3729	any opinions, | or
0.337209	-11.3731	other | other
0.174869	-11.376	a large | of
0.217532	-11.3775	because | are
0.714286	-11.3777	uses | uses
0.508929	-11.3778	is | suitable
0.0335985	-11.3783	the | of | is
0.813953	-11.3784	across | languages.
0.27957	-11.3785	first | second
0.393617	-11.3793	evaluated | on
0.76	-11.3794	short | long
0.813953	-11.3797	expression | expression
0.396739	-11.3799	knowledge | knowledge
0.540816	-11.3821	problem | problem
0.364055	-11.3833	but | also
0.182339	-11.3848	number of | that
0.161991	-11.3853	) | )
1	-11.3863	learning, | 23–24
0.735849	-11.3867	proved | be
0.255656	-11.389	was | using
0.966667	-11.3899	order | magnitude
0.179088	-11.3899	2 | 3
0.384615	-11.3904	results | experiments
0.39779	-11.3908	over | baseline
0.34728	-11.3914	aim | to
0.583333	-11.3931	types | type
0.171079	-11.3952	observed | the
0.231776	-11.3953	a | way
0.495652	-11.398	we add | to
0.647059	-11.3985	penn | treebank
0.262019	-11.3987	two | same
0.357143	-11.3988	if | is not
0.585366	-11.399	convert | to
0.328302	-11.399	in | 6.
0.701754	-11.3991	scores | scores.
0.364486	-11.3994	represent | as
0.340081	-11.3996	the | below.
0.140679	-11.4006	using | as
0.602564	-11.4017	in | scenario,
0.647059	-11.4019	being | being
0.232824	-11.4023	represents | of
0.417178	-11.403	distance | distance
0.677419	-11.4042	article | article
0.364486	-11.4049	those | which
0.795455	-11.4049	from | sources.
0.613333	-11.4049	experiments | performed
0.569767	-11.4055	on | hand
0.336	-11.4061	the | determined by
0.0917458	-11.4063	on | on
0.224199	-11.4066	and | do not
0.438356	-11.4074	and | 1998).
0.514019	-11.4085	in | table,
0.722222	-11.4086	structures | structures
0.722222	-11.4089	achieve | accuracy
0.140954	-11.409	is | set of
0.102842	-11.4098	1 | 1
0.462121	-11.4101	measure | similarity
0.235644	-11.413	that | does not
0.333333	-11.4133	assess | of
0.730769	-11.4156	nodes | nodes.
0.440559	-11.4161	a | tool
0.651515	-11.4168	image | image
0.443662	-11.4169	words | appear
0.265	-11.4177	annotator | the
0.265152	-11.4182	definition | a
0.689655	-11.4182	document | document.
0.405882	-11.4184	is | accurate
0.174242	-11.4188	because | is
0.935484	-11.4192	zettlemoyer | collins
0.455224	-11.421	for | systems,
0.578313	-11.4215	unlabeled | labeled
0.145234	-11.4226	the | process
0.571429	-11.4232	average | length
0.359813	-11.4246	g | g
0.5875	-11.4256	queries | query
0.23506	-11.4267	our approach | and
0.672131	-11.4274	cost | cost
0.13707	-11.4276	the | are not
0.42953	-11.4284	restrict | to
0.689655	-11.4286	future work | presented
0.25908	-11.4286	et | 2009),
0.0881775	-11.4296	all | in
0.205167	-11.4305	the | shared task
0.965517	-11.4306	transforming | into
0.47541	-11.4307	with | help
0.965517	-11.431	answers | answers
0.407186	-11.4317	6 | 7
0.74	-11.432	within | within
0.571429	-11.4328	wordnet | wordnet
0.217094	-11.4331	set | all
0.411043	-11.4347	for example, | sentence
0.589744	-11.4348	improvement | improvement
0.460938	-11.4364	in | document,
0.514563	-11.4364	split | into | and
0.368159	-11.4365	correlation | human
0.261194	-11.4369	l | l
0.74	-11.4373	learner | learner
0.352941	-11.4379	reason | that
0.205247	-11.4396	the | score.
0.482759	-11.4411	40 | 80
0.374359	-11.4414	performance | different
0.334711	-11.4415	a | algorithm.
0.491071	-11.4417	we | here.
0.191176	-11.4417	set of | that
0.573171	-11.4424	categories: | and
0.261307	-11.4425	takes | and
0.504673	-11.4426	dev | test
0.909091	-11.4429	6 7 | .
0.208267	-11.4432	portland, oregon, | 2011.
0.646154	-11.4439	two | 1)
0.397661	-11.4445	news | news
0.580247	-11.4454	probability | being
0.366337	-11.4476	this | done
0.191375	-11.448	in | way
0.611111	-11.4487	15 | 25
0.191375	-11.4499	with | same
0.328	-11.4511	same | but
0.55814	-11.4545	results | discussion
0.231225	-11.4551	if | is | it
0.965517	-11.4565	(petrov | klein,
0.358852	-11.4579	for | tasks.
0.447761	-11.4583	mean | mean
0.582278	-11.4589	generated | generated
0.460317	-11.4611	applies | to
0.336134	-11.4617	speech | speech
0.964286	-11.4636	part | speech,
0.322957	-11.4636	no | or
0.382514	-11.4637	we will | this
0.26087	-11.4637	for | we used
0.339056	-11.4645	for | systems.
0.666667	-11.4647	several | methods
0.495413	-11.4654	words | phrases
0.614286	-11.4662	11 | 12
0.241758	-11.4676	gives | of
0.243304	-11.4676	and | 2005).
0.837838	-11.4685	bootstrapping | seed
0.684211	-11.4688	class | class.
0.305263	-11.4694	while | more
0.655738	-11.47	experiments | conducted
0.655738	-11.4705	would be | if
0.228346	-11.4716	an | its
0.415584	-11.4727	it | useful
0.39645	-11.4728	(kim | al.,
0.305263	-11.4742	the | annotations.
0.266846	-11.4744	have | their
0.672414	-11.4749	0.4 | 0.6 0.7
0.200608	-11.4752	of | words,
0.388571	-11.4763	here, | is
0.785714	-11.4778	sentences | parsed
0.617647	-11.4778	each | pair.
0.785714	-11.478	swsd | contextual
0.115191	-11.4781	use | for
0.166144	-11.4782	such | or
0.559524	-11.4792	far | than
0.108156	-11.4793	the | et al.
0.605634	-11.4796	idea | using
0.640625	-11.4797	for | disambiguation.
0.329218	-11.4798	the | surrounding
0.369792	-11.48	used | train
0.576923	-11.4804	query | queries
0.453125	-11.4808	2006; | 2007;
0.495327	-11.4811	section | discusses
0.466667	-11.4818	one | three
0.837838	-11.4823	in | fashion,
0.678571	-11.4828	for | detection,
0.262467	-11.4831	a | language.
0.532609	-11.4833	in | settings.
0.317829	-11.4853	that | certain
0.456	-11.4855	for | annotation.
0.744681	-11.4855	distinguishing | from
0.532609	-11.4858	classifier | trained
0.474138	-11.487	provide | information
0.903226	-11.4882	huang | sagae
0.446154	-11.4887	providing | for
0.21492	-11.4887	another | is
0.661017	-11.4891	given | sentence,
0.72	-11.4898	percentage | correctly
0.0840541	-11.4899	if | is
0.526882	-11.4916	three | three
0.605634	-11.4935	resources | resources
0.188705	-11.4949	the main | of
0.154696	-11.4956	can be | for
0.962963	-11.4957	north | chapter
0.666667	-11.4963	thank | anonymous | helpful
0.131039	-11.4967	we use | and
0.406452	-11.4973	lm | lm
0.645161	-11.4974	emotion | emotion
0.356436	-11.4974	measure | between
0.423611	-11.4987	a | problem,
0.211806	-11.4995	here, | the
0.306569	-11.4998	also | but
0.755556	-11.4999	correlate | human
0.236443	-11.5002	the | methods.
0.126394	-11.5003	we | two | and
0.30855	-11.5012	with | features.
0.349282	-11.5027	using | algorithm
0.433824	-11.5055	a | rule.
0.0875149	-11.5062	for example, | the
0.685185	-11.5063	right | left
0.397516	-11.5065	as | example,
0.247596	-11.5066	q | q
0.331897	-11.5066	but | only
0.65	-11.5075	cluster | clusters
0.589041	-11.5099	whenever | is
0.172293	-11.51	simply | the
0.4	-11.5101	learn | from
0.780488	-11.5111	figure | depicts
0.729167	-11.5128	science foundation | grant
0.108059	-11.5133	a | which | a
0.236726	-11.5135	has | has
0.289474	-11.5135	in | task,
0.729167	-11.5137	question | whether
0.389222	-11.5147	our | model,
0.280374	-11.5151	and | (2009)
0.5625	-11.5153	in this paper, we | approach
0.12781	-11.5153	on | data
0.571429	-11.5162	task | task.
0.495146	-11.5163	and | respectively).
0.462185	-11.5165	lower | than
0.297203	-11.5165	too | to
0.903226	-11.5172	future work | explore
0.345972	-11.5172	asked | to
0.634921	-11.5183	variables | variables
0.73913	-11.5185	since | already
0.123792	-11.5186	in | corpus
0.833333	-11.5187	tutor | student
0.337838	-11.519	refers to | of
0.347619	-11.5192	in | case, | is
0.202265	-11.5196	we have | that
0.655172	-11.52	each | independently
0.75	-11.5202	p1 | p2
0.461538	-11.5203	this | true
0.962963	-11.5205	(brown | 1993)
0.298932	-11.5206	a | representing
0.75	-11.5214	turn | into
0.714286	-11.5218	child | child
0.293515	-11.5222	combining | and
0.625	-11.5227	policy | policy
0.372222	-11.5228	natural | edinburgh, scotland,
0.75	-11.5235	key | key
0.962963	-11.5247	recommendations | view
0.473214	-11.5248	it | easy to
0.372222	-11.526	consider | which
0.324895	-11.5262	from | original
0.377143	-11.5264	considered | be
0.196899	-11.5269	also | which
0.255875	-11.5272	a | lexicon
0.534091	-11.5274	propose | new
0.234136	-11.5276	key | of
0.962963	-11.5277	allocation | (blei
0.541176	-11.5278	learning | learn
0.193694	-11.5279	a small | of
0.521739	-11.5289	science | technology
0.301818	-11.5292	identifying | in
0.54878	-11.5294	vocabulary | words
0.274096	-11.5294	for | particular
0.327586	-11.5295	a | strategy
0.382353	-11.53	c 2011 | computational
0.461538	-11.5302	a | impact on
0.37931	-11.5305	table 2. | of
0.521739	-11.5307	grammar | rules
0.257294	-11.5313	is | through
0.294118	-11.5318	in addition to | and
0.833333	-11.533	it | impossible
0.7	-11.5333	as | see,
0.303704	-11.5334	a | word,
0.331858	-11.5334	the | form.
0.152416	-11.5337	we | three
0.28115	-11.5346	we present | for
0.679245	-11.5346	collection | documents
0.344498	-11.5347	sentence | its
0.391304	-11.5354	the | setting.
0.25876	-11.5356	in | languages.
0.39375	-11.5375	the | verbs.
0.359375	-11.5378	often | or
0.49505	-11.5397	and | (2009).
0.247525	-11.5399	similar to | of
0.583333	-11.5402	automatic | manual
0.5	-11.5402	w1 | w2
0.679245	-11.5404	unsupervised | induction
0.146957	-11.5409	are | using
0.331839	-11.5411	et | 2011).
0.794872	-11.5412	introduce | novel
0.217143	-11.5415	module | the
0.168981	-11.5417	in | previous
0.723404	-11.5433	bigram | trigram
0.485437	-11.545	some | (e.g.,
0.594203	-11.5453	best | achieved
0.331839	-11.5458	a | application
0.440945	-11.5459	used | experiments.
0.708333	-11.5459	in | format.
0.270833	-11.5462	as | function
0.241135	-11.5462	in | text.
0.660714	-11.5466	evaluate | quality
0.708333	-11.5472	arabic | english
0.0210343	-11.5473	the | of | and
0.223124	-11.5475	et | 2007).
0.522222	-11.5479	α | α
0.644068	-11.5486	amount | available
0.35567	-11.5487	to reduce | of
0.267442	-11.5493	the | eq.
0.528736	-11.5503	set | parameters
0.301115	-11.551	key | is
0.212569	-11.5511	the | rules.
0.156784	-11.5513	et | 2009).
0.193009	-11.5537	provides | of
0.606061	-11.5543	bleu | meteor
0.316872	-11.5563	comparing | to
0.371429	-11.5564	it | interesting
0.456897	-11.557	adding | features
0.456897	-11.5581	move | to
0.535714	-11.5617	running | on
0.456897	-11.5617	in | experiment.
0.0772448	-11.5625	we | this
0.516484	-11.5629	sum | all
0.828571	-11.5633	acl | workshop
0.183844	-11.5635	that | both | and
0.303846	-11.564	the | constraints.
0.733333	-11.5641	200 | 600
0.452991	-11.5642	number | clusters
0.135878	-11.5647	for | given
0.49	-11.5652	best | obtained
0.414286	-11.5657	agreement | agreement
0.183079	-11.5657	np | np
0.494845	-11.5658	corpora | corpora
0.848485	-11.5664	force | contract
0.558442	-11.5665	extending | to
0.567568	-11.5674	original | original
0.348485	-11.5679	given | its
0.134384	-11.5682	following | and
0.464286	-11.5684	is | extension
0.558442	-11.5685	instances | class
0.693878	-11.5686	connection between | and
0.242054	-11.5688	5.3 | the
0.325991	-11.5691	that | occur
0.775	-11.5692	children | children
0.928571	-11.5696	xylitol | preventing
0.294545	-11.5712	give | to
0.416058	-11.5717	broad | of
0.372093	-11.5718	while | other
0.480769	-11.5719	would | like
0.633333	-11.572	for | generation.
0.343137	-11.5734	system | baseline
0.31405	-11.5741	enough | to
0.310484	-11.5744	et | 2006),
0.805556	-11.575	depth | density
1	-11.5756	12: | 13:
0.489796	-11.5757	over | baseline.
0.165708	-11.5758	the | systems.
0.597015	-11.5766	candidate | candidates
0.673077	-11.5769	it | straightforward to
0.925926	-11.577	shed | on
0.567568	-11.5771	(from | to
0.421053	-11.5772	and | (2004)
0.925926	-11.5782	make | assumption
0.693878	-11.5783	easier | than
0.149254	-11.5787	association for | linguistics
0.424242	-11.5789	for | evaluation.
0.5	-11.5796	lexicon | words
0.96	-11.5797	(marcus | 1993)
0.380368	-11.5801	t | h
0.247423	-11.5802	the | becomes
0.303502	-11.581	a | if it
0.637931	-11.5813	some | (e.g.
0.295203	-11.5821	a | resource
0.239709	-11.5822	can | any
0.187779	-11.5828	a | classifier
0.377246	-11.5829	using | features.
0.769231	-11.5844	(p | 0.05)
0.727273	-11.5847	parent | child
0.270186	-11.5853	number | per
0.358696	-11.5865	major | in
0.517241	-11.5871	tree | trees
0.654545	-11.5876	event extraction | event
0.463636	-11.5879	compared | other
0.273016	-11.588	we describe | and
0.244898	-11.5886	2003; | al.,
0.660377	-11.5888	three | four
0.96	-11.5897	with | mind,
0.552632	-11.5907	relies | on
0.569444	-11.5915	otherwise, | is
0.0933679	-11.5926	for | for | for
0.275974	-11.5932	use | use
0.230769	-11.5933	that | specific
0.32287	-11.5937	the | alignment.
0.282313	-11.5941	two | their
0.213178	-11.595	the | method.
0.435484	-11.5956	propose | method
0.407143	-11.5971	the | community.
0.925926	-11.5975	mcdonald | nivre
0.505495	-11.5977	prior | prior
0.289286	-11.5977	the | lowest
0.363636	-11.598	that | actually
0.441667	-11.5981	2008; | 2009;
0.660377	-11.5981	on | held-out
0.370588	-11.5982	data | available
0.484848	-11.5983	table | summarizes
0.12716	-11.5985	the number of | in
0.660377	-11.6	restricting | to
0.214712	-11.6008	in | model,
0.571429	-11.6008	using | 2003).
0.411765	-11.6013	data, | data
0.38125	-11.6017	include | or
0.186012	-11.602	statistics | the
0.298851	-11.6022	again, | the
0.257143	-11.6024	check | the
0.603175	-11.6027	model, | model,
0.157113	-11.6035	the first | is
0.627119	-11.6037	subject | subject
0.96	-11.604	fifth | statistical
0.642857	-11.6042	main | main
0.823529	-11.6044	preceding | following
0.203571	-11.6046	in | sense
0.896552	-11.6047	rieser | lemon
0.590909	-11.6049	on | level,
0.289855	-11.6064	we propose | for
0.738095	-11.6072	shown | fig.
0.511364	-11.6077	appears | be
0.892857	-11.609	at | stage,
0.235012	-11.6092	entry | the
0.165485	-11.6094	we used | and
0.75	-11.6095	historical | historical
0.192616	-11.6104	on | data.
0.127465	-11.6107	number of | is
0.603175	-11.6113	false | false
0.167679	-11.6115	the | i.e.,
0.148467	-11.612	on | test
0.96	-11.6128	hundreds | thousands
0.255682	-11.6133	the | function.
0.616667	-11.6134	thanks to | for
0.188854	-11.614	the | it.
0.769231	-11.6142	impose | on
0.15	-11.6144	= | +
0.385621	-11.6147	from | corpora.
0.75	-11.6162	attribute | attribute
0.517647	-11.6163	features, | feature
1	-11.6181	(darpa) | laboratory
0.300395	-11.6188	and | relations.
0.355556	-11.6192	the | vector.
0.31982	-11.6195	amount | data
0.958333	-11.6204	defense advanced | program
0.56338	-11.6204	reason for | that
0.84375	-11.621	over | course
0.892857	-11.621	as | show,
0.200351	-11.6213	1 introduction | a
0.479592	-11.6227	automatic | human
0.711111	-11.623	achieved | performance
0.818182	-11.6236	5 conclusion | presented
0.410448	-11.6236	verbs, | and
0.573529	-11.6245	not all | are
0.147935	-11.6246	the | dataset
0.606557	-11.6249	contains | contains
0.695652	-11.6249	lda | topics
0.470588	-11.6252	summationtext | summationtext
0.648148	-11.6253	seed | seed
0.0237244	-11.6255	the | of | in
0.418605	-11.6257	a | verb.
0.8	-11.6258	while | latter
0.711111	-11.6272	al | al
0.494624	-11.6276	show that | approach
0.84375	-11.6277	(munteanu | marcu,
0.349462	-11.6281	from | corpus,
0.0707516	-11.6281	et al., | et
0.219149	-11.6282	address | of
0.341969	-11.6282	a | choice
0.335	-11.6288	a | one.
0.214141	-11.6297	word | is | word
0.923077	-11.6299	conditional | (crf)
0.474747	-11.6303	in | space,
0.31441	-11.6306	as: | =
0.8	-11.6306	men | women
0.291667	-11.6307	examine | of
0.238693	-11.6308	any | or
0.505747	-11.6309	− | −
0.59375	-11.6331	our | competitive
0.892857	-11.6331	unique | unique
0.222467	-11.6339	therefore | to
0.479592	-11.6346	words | occur
0.818182	-11.6348	conducted | experiments
0.407407	-11.6348	use | features.
0.254335	-11.6352	was | were
0.272131	-11.6355	a | system,
0.308017	-11.6355	model | baseline
0.234719	-11.6363	the | annotation.
0.171466	-11.6372	from | corpus
0.360465	-11.6377	figures | and
0.32243	-11.6378	we propose | that
0.168996	-11.6386	our | system
0.12474	-11.639	to be | in
0.261398	-11.6394	the | topics.
0.212982	-11.64	used | used
0.252874	-11.6408	the | policy
0.290566	-11.6427	selects | the
0.641509	-11.6434	1000 | 1000
0.267742	-11.6443	if | a | then
0.143514	-11.6443	of | words.
0.323944	-11.6449	frequency of | in
0.2	-11.6456	on | task
0.62069	-11.6463	table 1: | results
0.923077	-11.6464	during | process,
0.958333	-11.6478	||
0.259939	-11.6488	for | tasks
0.335025	-11.6494	and | 2005),
0.47	-11.6499	e.g., | or
0.0567951	-11.6499	= | =
0.351955	-11.6527	a | indicator
0.281588	-11.653	the | conditions
0.763158	-11.6533	reading | reading
0.763158	-11.6541	type | type.
0.083229	-11.6546	it is | to
0.818182	-11.6548	full | partial
0.363095	-11.6553	is | same as
0.680851	-11.6559	group | into
0.479167	-11.6574	and | 1997).
0.17689	-11.6577	there is | in
0.160093	-11.6577	the | compared to
0.862069	-11.6579	function | maps
0.253644	-11.6579	since | can
0.274914	-11.6581	is | difficult
0.445455	-11.6583	for | domains
0.610169	-11.6586	table | compares
0.153029	-11.6593	for instance, | the
0.201852	-11.6595	we | find
0.777778	-11.6599	x | y,
0.511905	-11.6609	form | form
0.6	-11.6617	current | current
0.438596	-11.6619	of | nouns.
0.263492	-11.6621	complete | of
0.338542	-11.6622	describe | that
1	-11.6626	nucleus | satellite
0.302521	-11.6628	results | evaluation
0.83871	-11.6629	solving | problem
0.267101	-11.6633	we evaluate | and
0.74359	-11.6641	said | be
0.777778	-11.6641	at | bottom
0.688889	-11.6651	on | dataset,
0.74359	-11.6652	learning | strategies
0.244505	-11.6653	we consider | of
0.549296	-11.6653	rely | on
0.567164	-11.6657	and | (2004),
0.304721	-11.6661	may be | for
0.304721	-11.6663	may | may
0.5	-11.6663	contain | information
0.8125	-11.6671	problem | solved
0.956522	-11.6673	2nd | june,
0.460784	-11.6673	cases | which
0.209677	-11.6679	is | 1.
0.956522	-11.6693	24 | usa
0.2125	-11.6696	and | 2010),
0.241287	-11.6697	more than | of
0.614035	-11.6702	compares | with
0.316279	-11.6707	the | standard.
0.266447	-11.6708	range | to
0.333333	-11.6708	present | which
0.371795	-11.6709	and | 2001).
0.301255	-11.6713	in | document.
0.6	-11.6715	conclusions | those
0.888889	-11.6718	tweets | tweet
0.361446	-11.6724	mostly | and
0.558824	-11.6732	(wang | al.,
0.92	-11.6737	semantic | (srl)
0.578125	-11.6745	for | iterations.
0.74359	-11.6747	for | details.
0.375	-11.6747	improvement | over
0.578125	-11.676	responses | responses
0.558824	-11.676	improving | by
0.725	-11.6763	from | table,
0.421488	-11.6764	due | fact
0.28	-11.6771	in particular, | and
0.614035	-11.6772	aspect | aspect
0.483516	-11.6777	other | like
0.181957	-11.6785	shows that | the
0.533333	-11.6786	languages | english
0.380952	-11.6789	= | (6)
0.590164	-11.6792	relate | to
0.272414	-11.681	what | is
0.203883	-11.6817	1 introduction | in
0.74359	-11.6818	labels | labels.
0.312217	-11.6819	the | 2009
0.794118	-11.6825	during | process.
0.697674	-11.6827	tokens | token
0.833333	-11.6836	location | location
0.4375	-11.6842	similar | other
0.297521	-11.6848	useful | for
0.162762	-11.6848	some | are
0.242424	-11.685	the | labels.
0.414634	-11.6862	is | indicator
0.409449	-11.6862	the | hypotheses.
0.697674	-11.6871	training | testing.
0.147208	-11.6871	for example, | is
0.62963	-11.6875	confidence | confidence
0.444444	-11.6881	require | be
0.62963	-11.6882	(for | (for
0.35503	-11.6885	each | assigned
0.697674	-11.6893	by | annotator.
0.361963	-11.6894	the | string.
0.289683	-11.6895	a | network
0.203488	-11.6907	investigate | of
0.409449	-11.691	expected | be
0.653061	-11.6913	we denote | as
0.494253	-11.6918	french, | and
0.348571	-11.692	protein | protein
0.329897	-11.6934	average | per
0.888889	-11.6934	(banerjee | lavie,
0.297071	-11.6935	= | i=1
0.956522	-11.694	vertices | edges
0.558824	-11.6943	compound | compound
0.189831	-11.6945	process | a
0.216336	-11.6949	the | output.
0.423729	-11.6953	(chen | al.,
0.433628	-11.6955	leads to | in
0.290837	-11.6958	1 introduction | is | of
0.560606	-11.6966	0.1 | 0.3 | 0.5
0.956522	-11.697	over | years,
0.367742	-11.6978	generate | from
0.269896	-11.6978	(2008) | a
0.30303	-11.6984	context | which
0.884615	-11.6986	modular | unification
0.149003	-11.6986	the | language model
0.174603	-11.6989	once | the
0.384615	-11.7	of | x.
0.220183	-11.7008	the | accuracy.
0.373333	-11.7009	the | conditions.
0.127008	-11.7012	of | types
0.226277	-11.7021	is | good
0.264026	-11.7025	the | grammar.
0.5	-11.7026	but | do not
0.315789	-11.7028	workshop | oregon,
0.18225	-11.7031	that | such as
0.0552815	-11.7037	is | by
0.884615	-11.704	presence | absence
0.5125	-11.7042	table 5: | for
0.459184	-11.7047	he | he
0.30531	-11.7058	to obtain | of
0.219907	-11.706	the | baseline.
0.342697	-11.7064	in | ways
0.697674	-11.7067	amounts | data.
0.697674	-11.7071	types | types.
0.240331	-11.7072	we | obtain
0.857143	-11.7074	in | proc.
0.281369	-11.7078	overall | for
0.5	-11.7081	features, | features,
0.182692	-11.709	strategy | the
0.163265	-11.709	there | no
0.187291	-11.7097	the | dataset.
0.240997	-11.7098	1 introduction | is
0.583333	-11.7098	due | lack
0.151815	-11.7105	the | component
0.681818	-11.7106	whose | was
0.736842	-11.7109	semantics | semantics
0.59322	-11.7109	university | pa
0.368421	-11.7111	captures | of
0.697674	-11.7113	tagged | tagged
0.954545	-11.7117	sixth | applications,
0.176119	-11.7124	authors | the
0.181388	-11.7124	it | possible | the
0.736842	-11.7125	weight | weights
0.483146	-11.7126	for | learning.
0.37415	-11.7128	perhaps | of
0.59322	-11.7145	making | more
0.0371326	-11.7164	this | the | of
0.787879	-11.7164	it | desirable
0.622642	-11.7169	about | million
0.488372	-11.717	for | evaluation,
0.164276	-11.7178	component | the
0.382979	-11.7179	combines | and
0.290984	-11.7183	finding | in
0.396947	-11.7186	the | parse.
0.542857	-11.7196	we expect | be
0.272401	-11.7214	the | side.
0.326425	-11.7224	table 2 shows | the
0.431193	-11.7241	better | better
0.305936	-11.7243	average | all
0.0804525	-11.726	the | for each
0.415254	-11.7266	for | analysis.
0.259016	-11.7273	consider | as
0.17594	-11.7277	and | 2007).
0.242775	-11.7278	the | (which
0.527027	-11.7283	tasks: | and
0.225806	-11.7295	and | 2009),
0.333333	-11.7296	we | introduce
0.571429	-11.7296	types | errors
0.328042	-11.73	once | is
0.168994	-11.7304	a | representation
0.454545	-11.7313	in | framework,
0.369128	-11.7316	we | two different
0.167813	-11.7322	of | and | respectively.
0.194085	-11.7327	the | project
0.916667	-11.7331	despite | still
0.354037	-11.7337	of | importance
0.717949	-11.7337	asr | asr
0.553846	-11.7338	annotations | annotation
0.583333	-11.7338	agreement | kappa
0.884615	-11.7338	random | (lafferty
0.139713	-11.7339	our | model
0.296943	-11.7342	solution | to
0.445545	-11.7346	the | literature.
0.187608	-11.7349	final | of
0.214932	-11.7361	our | was
0.280156	-11.7367	a | factor
0.28125	-11.7367	no | no
0.527778	-11.7368	in | contexts,
0.150334	-11.7369	allows | to
0.153667	-11.737	a | pair
0.178964	-11.7371	full | of
0.247734	-11.7371	again | the
0.690476	-11.7375	number | gender
0.375	-11.7389	split | into
0.4	-11.7393	in addition | also
0.6	-11.7394	systems | systems.
0.5	-11.7395	reaches | of
0.233244	-11.7398	and | them.
0.385185	-11.7407	select | from
0.806452	-11.7408	speed | per
0.205882	-11.7412	chosen | the
0.249231	-11.7412	and | combined
0.1331	-11.7412	according to | the
0.57377	-11.7415	mention | mentions
0.351852	-11.7421	the | use.
0.564516	-11.7423	arguments | argument
0.232	-11.7434	each | different
0.717949	-11.7438	text | text,
0.154762	-11.7445	the | performance.
0.331522	-11.7456	list | words
0.666667	-11.7467	phrases | phrases.
0.47191	-11.7467	possible | possible
0.198058	-11.7469	the | mentioned
0.319797	-11.747	create | for
0.236769	-11.7479	the | model | model.
0.615385	-11.7481	uses | features.
0.611111	-11.7485	we | trying
0.615385	-11.7488	clusters | cluster
0.615385	-11.7491	result | result
0.358974	-11.7492	account for | of
0.433962	-11.7495	which | (e.g.
0.536232	-11.7505	recall | f-score
0.88	-11.7507	email | email
0.666667	-11.7508	perplexity | model
0.553846	-11.751	initial | initial
0.16188	-11.751	description | the
0.441176	-11.7517	and | 2000)
0.302752	-11.7519	trained | using
0.419643	-11.7524	adapt | to
0.187063	-11.7524	estimated | the
0.7	-11.7525	yields | performance
0.851852	-11.7525	at | iteration,
0.378571	-11.7526	a | pair.
0.72973	-11.7527	it | assumed that
0.378571	-11.7537	have | high
0.364238	-11.7537	considers | of
0.408333	-11.7539	systems | their
0.545455	-11.754	significantly | baseline
0.352201	-11.7553	practical | of
0.453608	-11.7553	that | successfully
0.235457	-11.7554	and | (e.g.
0.954545	-11.7556	hartung | frank
0.586207	-11.7561	then | resulting
0.163978	-11.7562	the | could be
0.2225	-11.7567	those | are
0.31068	-11.7569	a | understanding
0.20332	-11.7571	solution | the
0.211236	-11.7574	on | development
0.528571	-11.7575	was | who
0.441176	-11.7575	smt | translation
0.0757925	-11.7578	the | where | is
1	-11.7579	johansson | nugues
1	-11.7579	law | 23-24
1	-11.7579	(zollmann | venugopal,
0.493827	-11.7583	when | user
0.851852	-11.7585	latent | indexing
0.228947	-11.7587	from | corpus.
0.396825	-11.7588	linguistics:shortpapers, | oregon,
0.576271	-11.7589	other | include
0.255738	-11.7589	a | representation of
0.370629	-11.7611	intuition | that
0.652174	-11.7616	in | iteration,
0.952381	-11.7617	street | (wsj)
0.44898	-11.7621	numbers | are
0.373239	-11.7623	the | stage.
0.462366	-11.7631	corpus, | corpus
0.528571	-11.764	in | direction.
0.404959	-11.7644	given | input
0.452632	-11.765	summarization | summarization
0.183849	-11.7652	3 | 4
0.316327	-11.7664	of | reviews
0.19685	-11.7672	here | a
0.386364	-11.7677	content | content
0.393701	-11.7678	difference | between
0.290598	-11.7683	the | case.
0.415929	-11.7683	2005; | 2007;
0.952381	-11.7691	until | reached.
0.30622	-11.7698	performance | when
0.341317	-11.7698	we discuss | in
0.379562	-11.7705	the | considered.
0.324324	-11.7711	given | given
0.142122	-11.7721	such | are
0.78125	-11.7733	incorrect | correct
0.682927	-11.7736	building | portland, | 24
0.415929	-11.7737	them | their
0.216152	-11.7739	there are | for
0.233983	-11.7739	were | was
0.546875	-11.774	used | training.
0.290043	-11.7742	the | value.
0.444444	-11.7742	listed | table
0.387597	-11.7743	as | final
0.143763	-11.7748	for | task
0.310345	-11.7749	a | solution
0.466667	-11.7755	a | manner
0.28692	-11.7759	the | producing
0.521127	-11.7763	after | after
0.383459	-11.7763	for | classification.
0.27668	-11.7765	of | greater
0.757576	-11.7765	letter | letter
0.444444	-11.7766	that | valid
0.377778	-11.777	terms | their
0.0628399	-11.7776	two | and
0.444444	-11.7777	show | improvement
0.659091	-11.7778	tends | be
0.659091	-11.7781	lower | higher
0.757576	-11.7781	introduce | into
0.682927	-11.7788	spanish | spanish
0.476744	-11.779	of | network.
0.513514	-11.7791	for example, | like
0.171429	-11.7792	of | have been
0.317708	-11.7794	an example | is
0.521127	-11.7798	in | applications,
0.432692	-11.7799	the | properties.
0.14256	-11.7808	problem | a
0.354839	-11.7811	an | system.
0.181973	-11.7813	results | results
0.447917	-11.7816	and | 2001)
0.521127	-11.7817	is | entirely
0.452632	-11.7832	modify | to
0.232044	-11.7835	there is a | of
0.481928	-11.7836	dependency | head
0.250814	-11.7836	mainly | the
0.36	-11.7837	a | function.
0.521127	-11.784	true | false
0.435644	-11.7845	our | approach.
0.319149	-11.7851	supported | grant
0.363014	-11.7851	it | necessary
0.742857	-11.7864	by | margin
0.240964	-11.7864	all | same
0.447917	-11.7873	entities | entity
0.493671	-11.7875	calculated | each
0.219451	-11.7876	and | 2003).
0.578947	-11.7878	many | them
0.348101	-11.7882	in | domains
0.16129	-11.7884	contains | and
0.319149	-11.7887	early | of
0.197183	-11.7887	2004; | al.,
0.471264	-11.7892	1 2 | 6
0.846154	-11.7897	bilingual lexicon | comparable
0.428571	-11.7897	underlying | is
0.365517	-11.7897	prime | prime
0.277108	-11.79	based | based
0.846154	-11.7903	outperforms | baselines
0.607843	-11.791	however, | did not
0.846154	-11.7913	comparable | state-of-the-art
0.264493	-11.7917	development | test
0.4875	-11.7919	and | (2007).
0.546875	-11.7928	a | concept.
0.710526	-11.7936	generally | but
0.122656	-11.7942	we will | the
0.410714	-11.7944	looking at | of
0.682927	-11.7944	makes | difficult
0.350649	-11.7944	are: | and
0.692308	-11.7947	automatically | manually
0.263538	-11.7958	of | 2011 | 2011.
0.201271	-11.7977	of | resulting
0.249191	-11.7984	in contrast, | the
0.913043	-11.7988	(see | details).
0.172897	-11.7991	relative | of
0.521739	-11.8002	gene | protein
0.0484425	-11.8008	in | and | in
0.123596	-11.801	the | among
0.0655002	-11.801	et al., | al.,
0.451613	-11.8011	2005; | 2006;
0.100423	-11.8017	a | over
0.427184	-11.8021	the | resource.
0.424528	-11.8035	have | effect
0.521739	-11.8035	information | available
0.625	-11.8036	report | accuracy
0.303922	-11.8037	both | same
0.821429	-11.8037	map | into
0.460674	-11.8041	known | be
0.29148	-11.8047	model, | model
0.581818	-11.8049	may be | other
0.568966	-11.8051	in addition, | features
0.875	-11.8054	minimum | maximum
0.10231	-11.8056	the | correct
0.269231	-11.8057	table 5 | the
0.513889	-11.8064	vector | vectors
0.568966	-11.8065	who | who
0.952381	-11.8067	pros | cons
0.333333	-11.807	20 | 40
0.607843	-11.8077	using | 2002).
0.333333	-11.8079	fact that | are
0.266917	-11.808	and | together
0.227397	-11.8081	one | two
0.137966	-11.8081	a | function
0.47561	-11.8084	even | still
0.2175	-11.8093	the | boundary
0.446809	-11.8093	models, | model
0.264925	-11.8095	precision | recall | and
0.235294	-11.8098	changes | the
0.487179	-11.8101	and | 1998)
0.3	-11.8103	first | first
0.821429	-11.8105	count | counts
0.95	-11.8107	weka | (hall
0.223097	-11.811	present | that
0.95	-11.8112	(9) | (10)
0.224	-11.8114	model | models
0.363636	-11.8116	for example, | two
0.434343	-11.8117	under | no.
0.692308	-11.8132	paraphrase | paraphrases
0.112075	-11.8132	and | 2010).
0.506849	-11.8137	for | queries.
0.47561	-11.8146	used | create
0.258993	-11.8149	good | for
0.735294	-11.8152	using | metrics.
0.43	-11.8153	the | scenario.
0.465116	-11.8154	these | include
0.16289	-11.8154	thus | of
0.612245	-11.8155	as | baseline,
0.538462	-11.8161	4: | 5:
0.350993	-11.8167	performs | on
0.506849	-11.8169	compared | baseline
0.548387	-11.8178	term | terms
0.308081	-11.8189	assume that | is
0.506849	-11.8192	kernel | kernel
0.735294	-11.8194	agreement | disagreement
0.276423	-11.8195	next, we | the
0.568966	-11.8196	system | evaluated
0.245161	-11.8197	the | step.
0.142241	-11.8199	this | is | by
0.361702	-11.82	placed | the
0.213075	-11.8211	compare | to
0.434343	-11.8213	in | directions
0.487179	-11.8217	latent | latent
0.814815	-11.8235	taken | account.
0.909091	-11.824	far | know,
0.251701	-11.8244	corpus | used
0.445652	-11.8245	french | french
0.257143	-11.8245	the former | the
0.075357	-11.8245	the | both | and
0.261029	-11.825	that | relevant
0.5	-11.8251	a | sense.
0.507042	-11.8253	evaluation | evaluated
0.186567	-11.8253	an | system
0.373134	-11.8253	a | place
0.774194	-11.8255	after | iterations
0.702703	-11.8256	students | their
0.419048	-11.826	higher | more
0.548387	-11.826	a | strategy.
0.909091	-11.8266	session, | portland,
0.272	-11.8268	the | parse tree
0.47561	-11.8268	60 | 90
0.167421	-11.8268	two | are | and
0.875	-11.8269	we refer | reader
0.361702	-11.8273	in | condition
0.0764873	-11.8274	the | of | sentence
0.793103	-11.8274	unknown | unknown
0.198312	-11.8275	to | translations
0.291667	-11.8277	are | across
0.442105	-11.8279	and | (2007),
0.323864	-11.8279	interestingly, | the
0.559322	-11.8284	argument | arguments
0.909091	-11.8285	university | southern
0.254355	-11.8286	so | can
0.95	-11.8291	solve | problems
0.270916	-11.8295	and | scores.
0.280851	-11.8296	the | structures.
0.309278	-11.83	in | gold standard
0.154044	-11.8301	the | match
0.333333	-11.8312	are more | than
0.95	-11.8313	rows | show
0.280851	-11.8315	review | and
0.875	-11.8317	utilize | information
0.215539	-11.832	if | has
0.364964	-11.8323	which | correct
0.0431937	-11.8328	and | are
0.164223	-11.8328	we | the following
0.55	-11.8328	a | scale.
0.202673	-11.8331	a | step
0.814815	-11.8331	comparing | against
0.714286	-11.8334	state | state.
0.584906	-11.834	achieved | accuracy
0.493333	-11.8344	larger | more
0.267442	-11.8354	the | resources.
0.230548	-11.8354	that | useful
0.612245	-11.8355	overall | scores
0.735294	-11.8357	increase | accuracy
0.793103	-11.8361	it | unclear
0.162824	-11.837	pages | portland,
0.651163	-11.8375	turn | turn
0.411215	-11.8376	for | domains.
0.651163	-11.8381	a | manner,
0.176271	-11.8384	complexity | the
0.714286	-11.8385	assigns | score
0.584906	-11.8389	an | area
0.714286	-11.839	intuition behind | is
0.714286	-11.839	make | decision
0.385246	-11.839	this | especially
0.306122	-11.8394	as | 1,
0.27459	-11.8396	the | event.
0.507042	-11.84	contexts | which
0.208531	-11.8402	role | in
0.242038	-11.8403	data | data.
0.702703	-11.8403	differences | differences
0.43299	-11.8403	used | build
0.793103	-11.8404	token | considered
0.295238	-11.841	the | corpus | corpus.
0.275	-11.8412	fact that | a
0.209639	-11.8419	our | method
0.486842	-11.8434	with | addition
0.270161	-11.8434	a | it.
0.454545	-11.8435	of | components.
0.84	-11.8443	cluster | cluster.
0.152956	-11.8444	matrix | the
0.909091	-11.8446	based on | observation
0.131256	-11.8448	performed | the
0.522388	-11.8448	treebank | treebank
0.236923	-11.8453	characteristics | the
0.651163	-11.8454	table 3: | different
0.382114	-11.8454	m. | and
0.256318	-11.8457	a | tree.
0.522388	-11.8465	these | will be
0.481013	-11.8466	2011, | 2011.
0.323699	-11.8466	parameters | are
0.166667	-11.8476	within | of
0.38843	-11.8482	task | text
0.104819	-11.8486	also | for
0.375	-11.8491	s. | and
0.17479	-11.8501	change | the
0.385246	-11.8501	set | experiments
0.287671	-11.8503	combinations | and
0.814815	-11.8515	shares | with
0.359712	-11.8532	examined | of
0.458824	-11.8532	mentioned | section
0.327381	-11.8532	to | domains
0.379032	-11.8532	when | new
0.227666	-11.8534	as | possible
0.12457	-11.8537	of | type
0.258303	-11.8538	the | space,
0.636364	-11.8541	some | contain
0.0648837	-11.8544	that | not
0.173045	-11.8545	the | ones
0.413462	-11.8549	john | and
0.401786	-11.8553	analyzed | and
0.469136	-11.8559	there | much
0.333333	-11.8559	select | that
0.523077	-11.8563	log | log
0.18232	-11.8564	the | approach.
0.766667	-11.8568	|||
0.407407	-11.8569	50 | 50
0.176471	-11.857	a | parser
0.43617	-11.8575	remaining | are
0.153247	-11.8583	provide | of
0.95	-11.8593	vectors | vectors.
0.369231	-11.8594	rule | rules
0.325444	-11.8599	also | some
0.140351	-11.8599	in | order
0.240385	-11.86	let | =
0.114599	-11.8606	result | the
0.251773	-11.8607	fact that | in
0.305699	-11.8608	on | web
0.161337	-11.8608	of | statistical
0.480519	-11.8611	some | them
0.869565	-11.8614	named | (ner)
0.321839	-11.8614	ratio | to
0.187135	-11.8614	of | classes
0.325444	-11.8616	because | only
0.346667	-11.8616	the | expression.
0.274262	-11.863	extracted | from
1	-11.8633	(lin | hovy,
0.279476	-11.8634	in | time.
0.947368	-11.8638	views | u.s.
0.5	-11.8638	conditional | fields
0.174576	-11.8639	the | ranked
0.283784	-11.8646	the | layer
0.727273	-11.8649	fragments | fragments
0.766667	-11.8649	2 | 6 8
0.362963	-11.8651	similar | et al.
0.149254	-11.8653	there are | and
0.357143	-11.8655	where | ∈
0.43956	-11.866	in | 2000).
0.20936	-11.8664	the main | is
0.311475	-11.8665	et | 2004)
0.342105	-11.8666	slightly | than
0.391304	-11.8668	this | requires
0.785714	-11.8672	demonstrate | effectiveness
0.111188	-11.8675	step | the
0.0887809	-11.8679	this | is | in
0.191753	-11.8681	the | learner
0.617021	-11.8682	emnlp | pages
0.234568	-11.8682	the | program
0.152941	-11.8685	calculated | the
0.551724	-11.8685	from | web.
0.225071	-11.869	similarly, | the
0.0986914	-11.8693	the | values
0.274262	-11.8693	that | improves
0.333333	-11.8695	the | collection.
0.2891	-11.8696	identify | that
0.288372	-11.8698	extended | to
0.463415	-11.8698	which | appear
0.316384	-11.8702	verb | its
0.694444	-11.871	tweets | twitter
0.342105	-11.8715	three | were
0.444444	-11.8717	each | binary
0.33125	-11.8726	3 | .
0.5	-11.8726	our | method.
0.327273	-11.8732	for | identification
0.434783	-11.8737	are | others
0.325301	-11.8738	in | dataset.
0.376	-11.8741	obtaining | of
0.480519	-11.8741	collected | from
0.233846	-11.8742	are shown | table
0.199552	-11.8749	the | below
0.84	-11.8756	associating | with
0.155525	-11.8756	were | from
0.807692	-11.8763	hindi | hindi
0.52381	-11.8766	by | speakers
0.257576	-11.8766	the | node.
0.387931	-11.8768	based on | information
0.766667	-11.877	improves | quality
0.532258	-11.8772	explained | section
0.84	-11.8775	run | iterations
0.486486	-11.8776	use | techniques
0.309783	-11.8778	representations | in
0.474359	-11.8778	bleu | nist
0.245734	-11.8781	instead of | a
0.727273	-11.8782	boundaries | boundaries
0.391304	-11.8788	observation | that
0.588235	-11.8791	find | optimal
0.259542	-11.8795	using | same
0.166405	-11.8798	common | and
0.4	-11.8807	accuracy | when
0.532258	-11.8811	language | (nlp)
0.869565	-11.8815	transition-based | graph-based
0.553571	-11.8818	with | precision.
0.507463	-11.8826	average | scores
0.283105	-11.8827	a | structure.
0.329193	-11.8829	into | single
0.270833	-11.8837	performance | system
0.302083	-11.8837	that | tend to
0.380165	-11.8841	as | above,
0.474359	-11.8844	up | up
0.0176331	-11.8852	the | is
0.666667	-11.8853	support | (svm)
0.212435	-11.8854	combined | and
0.0971551	-11.886	distribution | the
0.370079	-11.886	mark | and
0.785714	-11.8861	strategy | strategy
0.507463	-11.8862	of | conversation.
0.219178	-11.8863	identifying | and
0.137192	-11.8867	the | annotations
0.27234	-11.8871	area | the
0.247387	-11.8873	key | in
0.727273	-11.8878	annotating | with
0.627907	-11.8881	features | classifier.
0.240132	-11.8884	the | down
0.335484	-11.8888	that | contexts
0.392857	-11.889	nearly | of
0.566038	-11.8896	three | methods
0.486486	-11.8899	other | (e.g.
0.221289	-11.89	the | phrase.
0.675676	-11.8902	ratio | total
0.492958	-11.8903	all | mentions
0.415842	-11.8912	linking | to
0.329193	-11.8912	analyze | and
0.118211	-11.8912	thus | the
0.741935	-11.8913	documents, | documents
0.65	-11.8914	states | state
0.176157	-11.8916	since | not
0.486111	-11.8917	system, | system
0.347222	-11.8918	the | argument.
0.515625	-11.8926	a | environment
0.188912	-11.8927	one | each
0.833333	-11.8933	is | surprising,
0.52381	-11.8953	1 introduction | systems
0.0944559	-11.896	then | to
0.227545	-11.8962	computing | of
0.352518	-11.8963	the | test.
0.15427	-11.897	on | corpus
0.758621	-11.897	french | english.
0.285714	-11.8973	our | does not
0.281106	-11.8973	the | correct.
0.904762	-11.8974	only | handful
0.705882	-11.8988	effective at | tooth
0.627907	-11.8989	measures | measure
0.492958	-11.899	provide | about
0.807692	-11.8992	objective | subjective
0.16	-11.8994	to | original
0.0356776	-11.8997	a | a | a
0.904762	-11.8997	even if | still
0.244755	-11.8997	of | relations.
0.473684	-11.9	e.g., | e.g.,
0.259843	-11.9005	some | have
0.173684	-11.9008	coverage | the
0.306011	-11.9016	of | patterns.
0.0808768	-11.9018	was | to
0.214477	-11.902	potential | of
0.301587	-11.9022	the | gets
0.35	-11.9022	our | comparable
0.273913	-11.9022	the | candidates.
0.4625	-11.9023	of | utterances,
0.947368	-11.9025	attributes | attributes
0.65	-11.9025	particular | particular
0.193059	-11.9026	thus | a
0.195556	-11.9028	the | increases
0.221902	-11.903	is | common
0.241497	-11.9032	representing | of
0.373984	-11.9034	the | better.
0.675676	-11.9036	and | 2009.
0.863636	-11.9048	formal | nonregular
0.705882	-11.9048	state-of-the-art | systems
0.289216	-11.9051	for | prediction
0.705882	-11.9052	plan | explore
0.479452	-11.9053	the | application.
0.355556	-11.9054	outperforms | on
0.074943	-11.9055	use | to | the
0.385965	-11.9057	each | four
0.208122	-11.9058	results | different
0.165079	-11.906	built | the
0.212598	-11.9064	some | or
0.5	-11.9064	latent | analysis
0.049481	-11.9068	this | is
0.180952	-11.9071	between | between
0.35	-11.9077	classifier | features
0.298429	-11.9085	based on | words
0.281106	-11.909	we have | our
0.904762	-11.9093	 | 
0.171527	-11.9093	through | and
0.414141	-11.9096	using | simple
0.403846	-11.9103	(including | and
0.54386	-11.9104	however, | difficult
0.307263	-11.9106	in | evaluation.
0.863636	-11.9108	sequence | characters
0.308989	-11.911	is | whereas
0.479452	-11.9113	building | from
0.414141	-11.9118	and | (2008),
0.059455	-11.9122	then | the
0.685714	-11.9131	floor | floor
0.438202	-11.9137	used | compute
0.634146	-11.9137	arg0 | arg1
0.317365	-11.9147	show | results
0.329032	-11.9148	with | features,
0.0978078	-11.9152	obtained | the
0.595745	-11.9154	acknowledgments | thank | helpful
0.675676	-11.9156	question | questions
0.263374	-11.9157	many | have
0.28169	-11.9158	the | measure.
0.807692	-11.9159	larger | smaller
0.58	-11.9159	argument | predicate
0.280374	-11.916	a | space.
0.277273	-11.9161	technique | in
0.151556	-11.9175	if | or
0.0989074	-11.9185	2. | the
0.447059	-11.9192	a | metric.
0.253788	-11.9194	towards | of
0.249084	-11.9194	automatically | from
0.185714	-11.9198	a | task.
0.9	-11.9199	|bi|
0.863636	-11.9199	hand | hand
0.54386	-11.92	if | then | will
0.566038	-11.9201	sampling | sampling
0.331169	-11.9201	for | models,
0.396226	-11.9202	the | papers.
0.461538	-11.9202	data, | data.
0.368	-11.9203	the | lm.
1	-11.9204	munteanu | marcu
1	-11.9204	heritage, | sciences,
0.685714	-11.9206	symbol | symbol
0.265823	-11.9206	of | speakers
0.21727	-11.9207	then | each
0.944444	-11.9214	kim | hovy
0.263374	-11.9214	our system | a
0.144266	-11.9214	another | of
0.190578	-11.922	may | or
0.833333	-11.922	what | call
0.52459	-11.9222	labeled | label
0.205514	-11.9222	the | parser.
0.313953	-11.9223	that is, | a
0.229102	-11.9225	recent | et al.,
0.131823	-11.9227	are | but
0.507692	-11.9227	and | (2006).
0.758621	-11.923	frame | framenet
0.310345	-11.9234	performs | than
0.421053	-11.9234	thank | comments.
0.507692	-11.924	contexts | word
0.224551	-11.9242	direction | the
0.568627	-11.9244	uses | rules
0.461538	-11.9248	2009; | 2009).
0.568627	-11.9251	2006; | 2006).
0.233766	-11.9257	a | variable
0.0799848	-11.9259	the | model.
0.433333	-11.926	we | optimize
0.485714	-11.9261	every | has
0.177239	-11.9267	(2010) | a
0.657895	-11.927	a | basis.
0.396226	-11.9271	and | communication
0.396226	-11.9271	events, | and
0.436782	-11.9288	λ | λ
0.758621	-11.9292	sum | probabilities
0.288557	-11.9292	b | c
0.212938	-11.9298	we | generate
0.863636	-11.9298	6: | 11:
0.18595	-11.9299	in | current
0.485714	-11.9299	led | to
0.634146	-11.93	predicted | gold
0.196759	-11.9303	the | problem.
0.164251	-11.9306	run | the
0.277778	-11.9309	a variety of | and
0.236667	-11.9316	et | 2005).
0.0905088	-11.9318	the | value
0.371901	-11.9321	no | was
0.235099	-11.9321	empirical | of
0.249071	-11.9324	matched | the
0.71875	-11.9325	evaluate | effectiveness
0.287129	-11.9326	the | themselves
0.758621	-11.9327	figures | show
0.685714	-11.9329	algorithms | algorithm
0.207161	-11.9332	explore | of
0.310345	-11.9339	involves | and
0.340278	-11.9344	presented | section
0.223565	-11.9348	the | classes.
0.194954	-11.935	our approach | to
0.365854	-11.9356	corpora | corpus
0.8	-11.9357	analyses | analyses
0.272321	-11.9362	to compute | of
0.758621	-11.9365	tools | tools
0.298913	-11.9365	explored | and
0.263598	-11.9365	in | group
0.583333	-11.937	assigned | assigned
0.242105	-11.9372	for each | all
0.71875	-11.9374	and | 2004.
0.613636	-11.9378	by | points.
0.777778	-11.938	five | five
0.264706	-11.9381	we | add
0.436782	-11.9384	percentage | words
0.229299	-11.9385	but | more
0.365854	-11.9386	systems | different
0.380531	-11.9389	a | c,
0.231511	-11.9393	for instance, | is
0.833333	-11.9394	2011 | 23-24,
0.863636	-11.9399	measure | similarity.
0.455696	-11.9401	at | top
0.461538	-11.9402	although | may
0.833333	-11.9402	(church | hanks,
0.362205	-11.9402	in | contexts.
0.863636	-11.9405	show | effectiveness of
0.1632	-11.941	the | scores.
0.290816	-11.9415	in | approach,
0.319018	-11.9415	the | 5).
0.534483	-11.9418	twitter | twitter
0.666667	-11.9427	existing | systems
0.269912	-11.9432	parsed | the
0.6	-11.9438	these | properties
0.272727	-11.9441	two | both
0.160686	-11.9441	we have | for
0.634146	-11.9443	2004; | 2008).
0.685714	-11.9443	each | xi
0.485714	-11.9445	we perform | on
0.733333	-11.9446	percentage | tokens
0.408163	-11.9446	there | few
0.0678343	-11.9447	to | et al.,
0.279621	-11.9449	a | property
0.353846	-11.9454	identified | as
0.245421	-11.9455	procedure | in
0.777778	-11.9468	transition | emission
0.309942	-11.9473	first, | are
0.109183	-11.9474	quality | the
0.777778	-11.9476	students | learning
0.25498	-11.9478	the | behind
0.369748	-11.9489	the | scheme.
0.175701	-11.949	in | space
0.568627	-11.9497	generates | from
0.291667	-11.9499	positive | positive
0.440476	-11.9506	makes | more
0.280193	-11.9513	most | most
0.507937	-11.9517	a | role.
0.545455	-11.9518	evaluation | machine translation
0.246269	-11.9523	and | tasks.
0.343066	-11.9526	was performed | the
0.239437	-11.9527	generate | for
0.75	-11.953	replacing | by
0.290155	-11.9532	finally, | are
0.333333	-11.9538	without | or
0.267544	-11.9546	the | employed
0.40404	-11.9547	1 | 4 5
0.132337	-11.9548	work | et al.,
0.45	-11.9549	usually | or
0.777778	-11.9552	emails | emails
0.0541898	-11.9559	the | from | to
0.583333	-11.9559	gender | number
0.311377	-11.9563	however, | also
0.0941049	-11.9563	our | are
0.233333	-11.9563	ambiguity | the
0.777778	-11.9563	adopt | approach
0.30814	-11.9565	and | (2006)
0.507937	-11.9574	these | appear
0.373913	-11.9584	the | subject.
0.9	-11.9587	some | others,
0.426966	-11.9588	an alternative | to
0.193995	-11.9592	the | exactly
0.120214	-11.9624	errors | the
0.290155	-11.9627	of | combinations
0.357143	-11.9628	in | preceding
0.259414	-11.9632	of | systems,
0.274419	-11.9633	refer to | and
0.709677	-11.9634	1the | is
0.155689	-11.9635	a | problem
0.6	-11.9635	each | represented by
0.444444	-11.9635	as | increases
0.171846	-11.9636	comparison | and
0.6	-11.9638	divided | total
0.390476	-11.9644	two | methods
0.586957	-11.9645	specific | general
0.187364	-11.9646	the | types.
0.34058	-11.965	can | multiple
0.381818	-11.965	language | languages
0.167832	-11.9654	described | section
0.526316	-11.9656	data | available.
0.167832	-11.9656	of | texts
0.625	-11.9657	categories | category
0.185501	-11.966	entries | the
0.557692	-11.9661	even when | is
0.248092	-11.9662	positions | the
0.444444	-11.9662	under | under
0.262931	-11.9663	and | rules.
0.857143	-11.9664	number | grows
0.9	-11.9664	automated | scoring
0.430233	-11.9666	syntactic | lexical
0.230263	-11.9673	in | sentence,
0.228758	-11.9675	find | that
0.226115	-11.9675	the | errors.
0.535714	-11.9681	because | them
0.318471	-11.9682	the | list,
0.9	-11.9687	fifth | v),
0.305882	-11.9688	the | distribution,
0.324503	-11.969	decoder | a
0.127382	-11.9693	weights | the
0.314815	-11.9695	a | gold standard
0.113691	-11.9695	the | corpus,
0.253012	-11.9699	modeling | in
0.193925	-11.9702	for instance, | and
0.435294	-11.9704	rules | extracted
0.8	-11.9705	enrich | with
0.184713	-11.9707	root | the
0.289474	-11.9708	might be | to
0.410526	-11.9708	resources, | and
0.516667	-11.9709	uses | information
0.56	-11.9711	does not | well
0.124046	-11.9722	the | presented
0.269406	-11.9723	we used | from
0.448718	-11.9724	3.2 | our
0.857143	-11.9726	groups | groups
0.769231	-11.9729	achieved | f-score
0.228013	-11.9736	the | (table
0.769231	-11.974	subjective | subjective
0.54717	-11.974	asking | a
0.168739	-11.9746	we | several
0.181818	-11.9747	every | in
0.0802568	-11.9748	and | their
0.239286	-11.9753	approach | using
0.238434	-11.9754	of | text,
0.293478	-11.9756	predicting | of
0.189616	-11.9759	capture | of
0.535714	-11.9762	statistical | translation (smt)
0.69697	-11.9768	numbers | numbers
0.641026	-11.9777	alone | not
0.312883	-11.9778	great | of
0.413043	-11.9779	evaluation | evaluate
0.516667	-11.9785	comparison | different
0.894737	-11.9785	sri | modeling
0.331034	-11.9791	the | selected.
0.351562	-11.9794	in contrast, | a
0.648649	-11.9797	parsers | parsers
0.269406	-11.9797	trained | training
0.119217	-11.9797	our | our
0.260684	-11.9798	the | label.
0.255144	-11.9799	combine | and
0.367521	-11.98	scores | score
0.574468	-11.9801	factors | factors
0.257261	-11.9804	a | having
1	-11.981	5th | humanities,
1	-11.981	(gao | vogel,
1	-11.981	dragged | food
1	-11.981	additive | multiplicative
0.257261	-11.9812	the | table.
0.75	-11.9815	integration | into
0.471429	-11.9817	it | also be
0.101828	-11.9819	further | the
0.287958	-11.9823	used | evaluation
0.265487	-11.9824	the | state.
0.857143	-11.9825	as | whole,
0.291892	-11.9827	the | query.
0.17433	-11.9829	average | for
0.135889	-11.983	other | are
0.202046	-11.9834	s | t
0.331034	-11.9836	appropriate | for
0.465753	-11.9837	phrase-based | (koehn
0.197044	-11.9838	the | classifier.
0.586957	-11.9843	say | if
0.213256	-11.9845	those | that
0.112253	-11.9848	the | sentences.
0.676471	-11.9851	thank | suggestions.
0.253061	-11.9851	nn | nn
0.492308	-11.9852	graph | edge
0.244361	-11.9852	seed | and
0.604651	-11.9853	dependencies | dependencies
0.75	-11.9854	meaning | meanings
0.338129	-11.9856	and | distinct
0.941176	-11.9857	verbs | prepositions
0.894737	-11.9858	parser | marneffe et
0.492308	-11.9859	replaced | with
0.338129	-11.9859	single | all
0.40625	-11.986	for example, | about
0.255144	-11.9861	our | model.
0.227723	-11.9862	the | dimensions
0.178862	-11.9862	finally, | a
0.465753	-11.9864	(e.g., | or | (e.g.,
0.508197	-11.9869	models: | and
0.268182	-11.9869	baseline | model
0.625	-11.987	to | extent.
0.40625	-11.9872	sentences | length
0.255144	-11.9874	word | phrase
0.56	-11.9882	for | reasons
0.18125	-11.9884	proposed | for
0.769231	-11.9885	increasing | size
0.492308	-11.9888	linguistic | linguistic
0.0660611	-11.9891	we | using
0.397959	-11.9892	instead, we | a
0.826087	-11.9893	abstract | investigate
0.724138	-11.9898	fifteenth | june
0.941176	-11.9905	taking | consideration
0.517241	-11.9905	with | minimal
0.314465	-11.9906	how | how
0.0101095	-11.9908	the | of
0.155864	-11.9912	sentence | the | sentence
0.190805	-11.9912	the | class.
0.196078	-11.9912	english | with
0.305389	-11.9914	end | end
0.75	-11.9914	our focus | on
0.0799031	-11.9914	the | which is
0.26087	-11.9915	finally | the
0.604651	-11.9917	that | it,
0.508197	-11.9922	in | sense,
0.625	-11.9922	roles | roles
0.443038	-11.9923	discriminative | model
0.826087	-11.9931	under | model,
0.100577	-11.9931	the | system.
0.75	-11.9932	nouns, | adjectives
0.235088	-11.9939	a | (or
0.453333	-11.9939	processed | and
0.527273	-11.9939	unfortunately, | not
0.75	-11.9943	as | proxy
0.676471	-11.9946	entries | entries
0.5	-11.9946	three | 1.
0.302326	-11.9948	type | or
0.631579	-11.9953	in | 3.1,
0.492308	-11.9954	term | used
0.443038	-11.9955	with | tags.
0.224359	-11.9955	only | one
0.857143	-11.9957	exact | approximate
0.28877	-11.9958	the | utterance.
0.354839	-11.9962	entity | name
0.204244	-11.9964	an | set of
0.354839	-11.9971	a | deal
0.791667	-11.9975	useful | applications
0.609756	-11.9976	we ran | with
0.477612	-11.9977	balance | and
0.274038	-11.9977	results | accuracy
0.857143	-11.9982	statistics | statistics
0.397959	-11.9987	a | label.
0.318182	-11.999	the | (line
0.609756	-11.9991	assigns | probability
0.471429	-12.0001	and | null.
0.165493	-12.0002	a | domain
0.574468	-12.0009	because | cannot
0.464789	-12.0013	dialogue act | dialogue
0.127066	-12.0014	in | document
0.376147	-12.0015	the | high.
0.404255	-12.0015	taking | as
0.104022	-12.0016	the | do not
0.676471	-12.002	has been | project
0.305389	-12.0023	building | for
0.352	-12.0023	a | preference
0.247059	-12.0026	this | several
0.338235	-12.0027	the | times.
0.238971	-12.0027	a | description
0.12475	-12.0033	1 | 0
0.252033	-12.0035	counts | in
0.609756	-12.0044	parsing | parsers
0.459459	-12.0045	1 | 2.
0.12987	-12.005	this paper | a
0.172745	-12.0055	elements | the
0.648649	-12.0057	by | authors.
0.676471	-12.0058	across | languages
0.404255	-12.0058	four | two
0.32	-12.006	abstract | language
0.37963	-12.0062	many | but
0.194175	-12.0064	but | it
0.471429	-12.0068	we performed | on
0.538462	-12.0068	japanese | english
0.184615	-12.007	each | was
0.292818	-12.007	and | 2004),
0.229452	-12.0071	hypothesis | is
0.173077	-12.0072	dataset | and
0.352	-12.0072	in | tasks,
0.177551	-12.0077	likelihood | the
0.894737	-12.0078	makes | easier
0.380952	-12.0084	popular | in
0.12959	-12.0089	constraints | the
0.577778	-12.009	t | t.
0.508475	-12.009	system | developed
0.185268	-12.0093	to | text.
0.740741	-12.0094	</s>] | vp@2]
0.4375	-12.01	a | phrase,
0.384615	-12.0117	a | category.
0.301775	-12.0122	submitted | the
0.724138	-12.0123	size | corpus,
0.150665	-12.0123	decision | the
0.333333	-12.0127	provides | with
0.631579	-12.0128	defined | equation
0.740741	-12.0129	topics | lda
0.335766	-12.0132	with | maximum
0.349206	-12.0135	of | european
0.193627	-12.0136	% | %
0.288043	-12.0136	models | trained
0.791667	-12.0137	reward | reward
0.508475	-12.0138	algorithm | find
0.818182	-12.0143	task, | june
0.340909	-12.0147	the | object.
0.151111	-12.0147	allows | of
0.131609	-12.0148	the | parameter
0.4	-12.015	due | its
0.857143	-12.0151	two | a)
0.411111	-12.0155	finally, we | our
0.186364	-12.0159	the | structure.
0.208451	-12.0162	the | (figure
0.286486	-12.0164	a | baseline.
0.657143	-12.0165	name | names
0.111831	-12.0168	show that | the
0.330935	-12.0173	of | texts,
0.248	-12.018	occur | in
0.4375	-12.0182	checking | of
0.254237	-12.0183	let | be | in
0.101478	-12.0184	= | and | =
0.147518	-12.0186	four | and
0.941176	-12.0188	asked | judge
0.54902	-12.0189	these | help
0.0604762	-12.019	a | model
0.326389	-12.0194	considerably | the
0.317881	-12.0195	determining | a
0.595238	-12.0201	comparative | comparative
0.240458	-12.0202	a simple | to
0.54902	-12.0204	a | procedure.
0.340909	-12.0204	and | tags.
0.328571	-12.0206	and | errors.
0.312102	-12.0206	a | graph.
0.106956	-12.0207	length | the
0.0935539	-12.0212	is | its
0.321918	-12.0216	a | method.
0.423529	-12.0218	typically | or
0.380952	-12.0221	the | similar.
0.724138	-12.0223	drawback | that
0.163732	-12.0226	is | case
0.447368	-12.0228	only | top
0.227119	-12.023	in | system,
0.34375	-12.0239	∑ | ∑
0.222222	-12.0246	= | .
0.538462	-12.0246	a | forest.
0.34375	-12.026	the | transcripts
0.791667	-12.0263	from | clothing
0.215152	-12.0264	the | values.
0.107881	-12.0265	the | pages | 2011.
0.894737	-12.0267	create | create
0.477612	-12.0268	from | database
0.209913	-12.0269	the | (with
0.296512	-12.0274	the | submitted
0.894737	-12.0277	mcdonald | pereira
0.341085	-12.0278	in this way, | the
0.282723	-12.0279	the | one,
0.235294	-12.0279	abstract | is
0.615385	-12.0283	chose | because
0.311688	-12.0296	the | neighboring
0.225589	-12.0298	in | compared to
0.270531	-12.0299	of | but not
0.740741	-12.0299	connectives | connective
0.246964	-12.0301	while | have
0.6875	-12.0303	train | classifiers
0.0936416	-12.0304	size | the
0.317568	-12.0305	in | systems,
0.55102	-12.0313	lda | topic
0.236162	-12.0323	and | 2006),
0.517857	-12.0324	score | computed
0.391753	-12.0327	translations | source
0.28125	-12.0329	present | results
0.666667	-12.0331	spelling | spelling
0.432099	-12.0331	improved | by
0.418605	-12.0334	in | applications.
0.406593	-12.0337	for | research.
0.615385	-12.0344	(ohta | al.,
0.638889	-12.0345	9 | .
0.261468	-12.0349	is | appropriate
0.096201	-12.0351	we use | of
0.296512	-12.0352	the | senses.
0.290503	-12.0355	defines | of
0.227586	-12.0359	department | university
0.226804	-12.036	a | summary
0.333333	-12.0361	is | different from
0.326241	-12.0361	a | cannot be
0.225589	-12.0361	the | representation.
0.85	-12.0371	(barzilay | 2003;
0.158333	-12.0379	approach, | the
0.352941	-12.038	analysis | analysis
0.346774	-12.038	in | time,
0.179487	-12.0383	2009; | al.,
0.160684	-12.0383	the | domain.
0.888889	-12.0384	imported | exported
0.302469	-12.0386	a | rich
0.7	-12.0387	different | strategies
0.140604	-12.0387	consider | and
0.373832	-12.0389	we chose | and
0.25	-12.039	the | constituents
0.0904139	-12.0394	we | these
0.255411	-12.0394	while | has
0.231317	-12.0395	result | that
0.441558	-12.0396	depends | on
0.321678	-12.04	to | back
0.638889	-12.04	there | enough
0.180435	-12.0406	the | to produce
0.338462	-12.0406	the | base.
0.409091	-12.0407	in | study.
0.212766	-12.0411	in | experiments.
0.581395	-12.0413	tree | root
0.941176	-12.0416	mix | o-en
0.818182	-12.042	precision, | recall
0.284153	-12.042	(as | in
0.191646	-12.0422	the | not.
0.323944	-12.0424	user | system
0.888889	-12.0425	smoothing | goodman,
0.362832	-12.0428	are | via
0.491803	-12.0431	transfer | from
0.349593	-12.0433	recently | et
0.165746	-12.0434	then, | the
0.409091	-12.0434	model | inference
0.85	-12.0435	tables 1 | 2
0.528302	-12.0439	findings, | expressed
0.85	-12.0443	initial | update
0.528302	-12.0444	believe | will
0.421687	-12.0444	method | large
0.302469	-12.0446	the | dictionary,
0.888889	-12.0447	first | second.
0.491803	-12.0448	in | domains,
0.235075	-12.045	the | pipeline
0.666667	-12.0453	is | efficient.
0.149701	-12.0453	we | three | and
0.315436	-12.0459	challenge | to
0.565217	-12.046	some | e.g.,
0.5	-12.046	for | annotation,
0.281915	-12.0461	a | constructed
0.266667	-12.0464	value of | is
0.226804	-12.0465	results | both
0.518519	-12.0465	using | method.
0.85	-12.0465	page | page
0.85	-12.0467	discriminative | generative
0.76	-12.0467	actions | actions
0.469697	-12.0472	6.1 | we
0.818182	-12.0472	applicable | any
0.173387	-12.0472	set | features
0.278947	-12.0475	corresponding to | and
0.0876289	-12.0476	the | the | sentence.
0.9375	-12.0478	(chiang, | chiang,
0.245902	-12.0479	tasks, | and
0.6	-12.048	replace | by
0.277487	-12.0482	of | classes.
0.565217	-12.0487	automatically | corpus.
0.714286	-12.0488	and | (2002).
0.5	-12.0491	for | readers
0.483871	-12.0492	originally | in
0.277487	-12.0497	sentences | their
0.491803	-12.0498	type | other
0.9375	-12.0499	texas | austin
0.29878	-12.0503	for | applications
0.116728	-12.0509	translations | the
0.445946	-12.0513	a | transducer
0.173824	-12.0513	the | module
0.621622	-12.0513	graph | graph.
0.740741	-12.0514	plaintext | cipher
0.323944	-12.0515	as | etc.
0.85	-12.0516	further | needed
0.231884	-12.0517	values | are
0.0707786	-12.0523	the | algorithm
0.7	-12.0524	segmentation | pos-tagging
0.553191	-12.0532	shows | improvement
0.553191	-12.0534	are | rare
0.818182	-12.0536	variants | same
0.21148	-12.054	in | example,
0.375	-12.0543	and | (2009),
0.7	-12.0552	improve | quality.
0.243902	-12.0552	for | generation
0.212963	-12.0556	comparing | and
0.508772	-12.0558	used | estimate
0.311258	-12.0561	“the | of
0.335878	-12.0565	we | most frequent
0.5	-12.0565	constrain | of
0.740741	-12.0568	use | hedges
0.260465	-12.0569	the | probability.
0.378641	-12.057	a | medical
0.269608	-12.057	the | dataset,
0.323741	-12.0571	we expect | to
0.9375	-12.0571	part | speech.
0.169261	-12.0572	the | ratio
0.251064	-12.0572	if | would
0.185185	-12.0573	in | section, we
0.235075	-12.0576	the | metrics.
0.168605	-12.0578	standard | for
0.483871	-12.0578	begins | a
0.638889	-12.0579	had | been
0.245902	-12.058	application | to
0.259091	-12.0581	our | uses
0.29697	-12.0582	propose | which
0.206997	-12.0583	we | 10
0.335878	-12.0586	in | virtual
0.179825	-12.0588	a | (e.g.
0.210843	-12.059	we | dataset
0.254386	-12.0592	(e.g., | or
0.333333	-12.0593	from | text.
0.166352	-12.0594	a | level
0.127636	-12.0594	evaluated | the
0.284916	-12.0596	attempts to | the
0.621622	-12.0598	mst | mst
0.529412	-12.06	approaches | approaches
0.421687	-12.0601	our | includes
0.0872483	-12.0603	the | being
0.220395	-12.0607	of | textual
0.194805	-12.0608	this | very
0.389474	-12.0615	a | parser,
0.133495	-12.0615	that, | the
0.508772	-12.0617	notion | is
0.568182	-12.0617	co-occurrence | words
0.76	-12.0617	perform | task.
0.317241	-12.0619	our | state-of-the-art
0.647059	-12.0622	and | 7).
0.581395	-12.0624	users | about
0.239216	-12.0629	phrases | are
0.491525	-12.0629	our | outperform
0.416667	-12.063	a | trend
0.126224	-12.063	times | the
0.677419	-12.0631	since | usually
0.128523	-12.0632	there are | of
0.16996	-12.0633	a | improvement
0.27957	-12.0636	is | root
0.356522	-12.0639	and | learning.
0.283333	-12.0639	possibility | the
0.85	-12.0643	correcting | errors
0.315068	-12.0647	the | iteration.
0.457143	-12.0649	0.0 | 1.0
0.38	-12.065	using | features,
0.303797	-12.0664	a | distribution.
0.76	-12.0671	edges | vertices
0.76	-12.0671	sentences, | words.
0.782609	-12.0674	parameters | on | development
0.581395	-12.0674	judged | be
0.344262	-12.0675	of | queries.
0.185714	-12.0678	the | involved
0.0831365	-12.0681	a | system
0.317241	-12.0691	but | some
0.888889	-12.0693	wi | wj
0.647059	-12.0694	english | italian
0.228571	-12.0696	a | vocabulary
0.568182	-12.0696	2010; | 2010),
0.621622	-12.0698	words | dictionary.
0.677419	-12.07	for | adaptation.
0.333333	-12.07	in | 2007),
0.347107	-12.0701	crucial | of
0.216129	-12.0702	resulting | a
0.9375	-12.0702	corpus (brants | franz,
0.367925	-12.0705	the | correctly.
0.434211	-12.0707	measures | similarity
0.134663	-12.0713	the | object
0.341463	-12.0714	reflected | the
0.169291	-12.0714	common | in
0.730769	-12.0718	grouping | into
0.730769	-12.0718	there | considerable
0.29697	-12.072	2011 | pages | edinburgh,
0.541667	-12.0721	it seems | that
0.621622	-12.0722	achieves | results
0.677419	-12.0723	relates | to
0.141864	-12.0728	a | value
0.103627	-12.0729	systems | and
0.395604	-12.0731	perform | using
0.509091	-12.0733	classes | verbs
0.256881	-12.0733	is | i.e.,
0.319149	-12.0738	however, | some
0.553191	-12.074	there | room
0.376238	-12.074	however, | many
0.123679	-12.0745	automatic | of
0.809524	-12.0747	under | curve
0.350427	-12.0752	setting | for
0.149231	-12.0753	a | grammar
0.518519	-12.0754	algorithm | algorithm.
0.9375	-12.0754	how | varies
0.553191	-12.0762	google | google
0.444444	-12.0765	and | locations
0.186158	-12.0766	the | outputs
0.275132	-12.0769	participants | a
0.33871	-12.0769	an initial | of
0.290698	-12.0771	n/a | n/a
0.131422	-12.0773	next | the
0.809524	-12.0775	no | exists
0.76	-12.0775	grammar | grammars
0.553191	-12.0779	a | decoder.
0.323529	-12.0781	reflects | of
0.677419	-12.0783	address | problem,
0.585366	-12.0787	each | step.
0.419753	-12.0795	with | goal
0.289017	-12.0796	system | systems
0.395604	-12.0799	word | words,
0.483333	-12.08	approaches | they
0.585366	-12.0807	recently, | et al.
0.335938	-12.0809	finding | that
0.677419	-12.081	combination | individual
0.112174	-12.0814	and | 2009).
0.289017	-12.0814	number | for each
0.207831	-12.0815	of | contexts
0.209091	-12.082	the | lexicon.
0.26601	-12.082	= | (4)
0.124597	-12.082	the | hypothesis
0.419753	-12.0822	row | table
0.85	-12.0825	supporting | supporting
0.391304	-12.0827	in | graph,
0.85	-12.0829	(law | 2011.
0.254545	-12.083	previous work | et
0.395604	-12.0832	way | which
0.509091	-12.0833	dept. | university
0.357143	-12.0841	never | in
0.0551447	-12.0844	using | and
0.179372	-12.0848	we | evaluate
0.153974	-12.0851	table 3 | the
0.188119	-12.0851	morphological | and
0.438356	-12.0852	a | phenomenon
0.605263	-12.0853	primary | primary
0.367925	-12.0854	introduces | of
0.65625	-12.0855	survival | survival
0.529412	-12.0856	features | predict
0.325926	-12.0859	we will | our
0.255708	-12.086	of | type.
0.491525	-12.0861	approach | supervised
0.628571	-12.0862	technologies | pittsburgh,
0.328244	-12.0865	a | interface
0.444444	-12.0872	compute | probability
0.164773	-12.0874	the | item
0.529412	-12.0874	the | adjectives.
0.323529	-12.0875	used | experiments
0.183962	-12.0875	three | are
0.519231	-12.0878	that | to.
0.214058	-12.0879	summationdisplay | summationdisplay | summationdisplay
0.555556	-12.0882	1 | 2,
0.65625	-12.0886	then | remaining
0.291667	-12.0886	= | otherwise
0.628571	-12.0888	almost | same
0.138441	-12.0897	before | the
0.166344	-12.09	this | over
0.677419	-12.0902	all | paths
0.14497	-12.0904	in | original
0.414634	-12.0907	sequences | are
0.235294	-12.0907	given | sentence
0.677419	-12.0909	(zhu | al.,
0.330769	-12.0909	compare | results
0.323529	-12.0911	are | together
0.126824	-12.0912	1 introduction | of
0.325758	-12.0913	for | items
0.888889	-12.0914	metrics | metrics.
0.677419	-12.0916	because | requires
0.344538	-12.0918	an overview | the
0.568182	-12.0918	incorporate | features
0.361111	-12.0921	supported | part
0.402299	-12.0923	sizes | and
0.277174	-12.0924	using | method
0.5	-12.0926	because | often
0.166994	-12.0927	is | correct
0.585366	-12.0935	process | until
0.080994	-12.0937	one | is
0.647059	-12.0937	candidates | candidate
0.162921	-12.0938	only | but
0.244726	-12.0939	written | and
0.270833	-12.094	determining | of
0.318841	-12.0941	extension | to
0.207317	-12.0941	the | relation.
0.519231	-12.0944	(1999) | a
0.278689	-12.095	the | algorithms.
0.318841	-12.0957	of | meaning.
0.242678	-12.0959	is | i.e.
0.387097	-12.0961	identifies | and
0.182243	-12.0962	representation | in
0.381443	-12.0965	the | separately.
0.438356	-12.0965	for | selection.
0.530612	-12.0966	nist | evaluation
0.23622	-12.0968	it | better
0.423077	-12.0968	formulate | a
0.304636	-12.0968	5.1 | we
0.677419	-12.0969	is | say,
0.202312	-12.0976	system | it
0.372549	-12.0978	in | 3)
0.202312	-12.0979	the | overlap
0.253394	-12.098	depends on | the
0.628571	-12.0984	(2010) | similar
0.888889	-12.0986	avoid | problem
0.299363	-12.0986	provide | on
0.223404	-12.0987	the | algorithm,
0.888889	-12.0988	synchronous | (scfg)
0.432432	-12.0993	in | situation
0.238866	-12.0994	we | here
0.0974462	-12.0995	we have | of
0.156089	-12.0997	the | documents.
0.251121	-12.1002	into | into
0.328244	-12.1005	cross-validation | the
0.333333	-12.1005	linguistics:shortpapers, | 19-24,
0.282486	-12.1005	determines | the
0.284884	-12.1012	a | score.
0.428571	-12.1012	giving | to
0.782609	-12.1012	sentences, | sentences,
0.221453	-12.1012	exact | of
0.263682	-12.1017	the | antecedent
0.0962672	-12.1019	for example, | of
0.274194	-12.1025	however, | has
0.265	-12.1028	performed | on
0.239837	-12.1029	we | predict
0.318841	-12.103	are | them.
0.519231	-12.1032	order | avoid
0.166667	-12.1033	using | data
0.302632	-12.1039	for example, | has
0.184466	-12.1042	normalized | the
0.330709	-12.1043	for | construction
0.292683	-12.1044	and | 2002)
0.397727	-12.1049	used | same | as
0.200573	-12.1053	the | score,
0.16	-12.1056	5.2 | the
0.842105	-12.1057	body | research
0.75	-12.1057	take | form
0.21843	-12.1058	5.1 | of
0.75	-12.1059	(referred | as
0.344828	-12.1059	in | dictionary.
0.354545	-12.106	reviewers | comments.
0.314286	-12.1062	in | direction
0.509434	-12.1063	of | condition.
0.571429	-12.1064	with | judgment
0.5	-12.1065	labels | label
0.123243	-12.1067	abstract | of
0.530612	-12.1067	category | categories
0.5	-12.1068	mark | as
0.272727	-12.1072	the | sequence.
0.365385	-12.1072	integrating | and
0.157801	-12.1073	finally, we | the
0.336066	-12.1075	returns | of
0.397727	-12.1076	factored | and
0.409639	-12.1077	after | was
0.185644	-12.1077	of | utterances
0.605263	-12.1078	computing | using
0.292683	-12.1079	word | words.
0.119192	-12.1082	the | still
0.316547	-12.1083	since | may
0.402299	-12.1087	ask | to
0.280899	-12.1087	gives | for
0.809524	-12.1088	utterances | utterances.
0.24359	-12.1093	for example, | word
0.238683	-12.1094	extended | and
0.882353	-12.1096	mstparser | maltparser
0.357798	-12.1096	generic | in
0.252294	-12.1098	the | summary.
0.530612	-12.1099	as | earlier,
0.555556	-12.1101	giza++ | and | 2003)
0.882353	-12.1102	bleu | ter.
0.155979	-12.1107	so | is
0.200573	-12.1112	in | final
0.272727	-12.1119	idea | that
0.314286	-12.112	a | might be
0.281609	-12.112	with | frequency
0.115789	-12.1121	provided | the
0.248889	-12.1122	(the | is
0.300654	-12.1124	from | list
0.347826	-12.1126	defines | and
0.393258	-12.1126	this | quite
0.152066	-12.1126	in | way | the
0.461538	-12.1127	extend | with
0.11326	-12.1133	the | individual
0.842105	-12.1137	italian | english
0.131841	-12.1137	learning | for
0.361905	-12.1139	machine learning | for
1	-12.1146	(denkowski | lavie,
1	-12.1146	goldberg | elhadad
1	-12.1146	(griffiths | steyvers,
0.138696	-12.1147	the | indicates
0.636364	-12.1147	from | sources,
0.933333	-12.1148	ssst-5, | usa,
0.842105	-12.1151	no | found.
0.284884	-12.1152	the | obtaining
0.47541	-12.1154	for | classification,
0.289157	-12.1156	the | recognizer
0.605263	-12.1156	all | yield
0.378947	-12.1158	even though | in
0.809524	-12.1158	evaluate | ability
0.809524	-12.1163	mwes | mwes
0.145234	-12.1164	et al., | 2010).
0.75	-12.1166	edinburgh, | 30,
0.703704	-12.1168	occurs | times
0.338843	-12.1169	the | parsers.
0.325581	-12.1169	using | methods
0.491228	-12.1169	acknowledgements | thank
0.145929	-12.1171	able to | of
0.148265	-12.1172	the | whole
0.225926	-12.1172	the | configuration
0.258373	-12.1173	measuring | of
0.194521	-12.1173	moreover, | of
0.12118	-12.1177	is | most
0.4125	-12.1179	compare | against
0.316547	-12.118	the | procedure.
0.259804	-12.1182	compared | with
0.314286	-12.1187	especially | for
0.442857	-12.1187	selecting | from
0.689655	-12.1189	applications | text
0.0739284	-12.1189	analysis | the
0.5	-12.1189	and | unified
0.358491	-12.119	consider | where
0.409639	-12.1191	as | become
0.161959	-12.1194	how | is
0.105431	-12.1197	the | provided
0.730769	-12.1199	nodes | connected
0.225926	-12.1199	and | 2003)
0.150082	-12.1202	results | using
0.354545	-12.1205	data | data,
0.382979	-12.1207	case | one
0.351351	-12.1208	a | fit
0.153584	-12.121	in | sentence.
0.666667	-12.1211	for | insightful
0.155323	-12.1216	utterances | the
0.636364	-12.1219	of | (2004).
0.703704	-12.1219	probability | generating
0.404762	-12.122	where | represents
0.365385	-12.1221	after | were
0.933333	-12.1231	(buchholz | 2006).
0.703704	-12.1231	additional | added
0.135099	-12.1236	and | scores
0.243478	-12.1237	the | plus
0.636364	-12.124	outperforms | methods
0.882353	-12.1241	each | separately,
0.147619	-12.1242	edges | the
0.101873	-12.1243	task | a
0.373737	-12.1251	but | them
0.772727	-12.1253	adequacy | fluency
0.4	-12.1256	using | metric
0.182927	-12.1262	4 | our
0.150082	-12.1262	as | can be
0.543478	-12.127	measuring | similarity
0.195531	-12.1276	the | steps
0.482759	-12.1277	each | separate
0.133766	-12.128	input | a
0.395349	-12.1285	treated | as
0.223443	-12.1291	a | increase
0.232283	-12.1292	then | using
0.25	-12.1299	this | through
0.244541	-12.13	where | where
0.330645	-12.13	incrementally | the
0.171306	-12.1303	an | sentence
0.306122	-12.1304	in | analysis.
0.0764331	-12.1304	into | and
0.460317	-12.1308	to | beliefs
0.18408	-12.1309	the | type.
0.361905	-12.131	a | cover
0.361905	-12.1315	requires | be
0.160985	-12.1319	the | parts
0.426667	-12.1322	expressed | reflect
0.0955882	-12.1325	some | in
0.296774	-12.1326	results | those
0.351351	-12.1327	implicit | in
0.611111	-12.1328	figure | plots
0.454545	-12.1332	a | permutation
0.378947	-12.1333	several | different
0.102212	-12.1337	the | representation
0.474576	-12.1337	since | need
0.0865753	-12.1342	is | using
0.27027	-12.1344	baseline | system
0.772727	-12.1345	constraints | constraints.
0.318519	-12.1351	for every | in
0.280702	-12.1352	(2005) | and
0.636364	-12.1353	best | selected
0.509804	-12.1353	supervised | labeled
0.351351	-12.1359	were | based on
0.320611	-12.136	for | dataset.
0.211921	-12.1365	is | true
0.325581	-12.1368	data | train
0.246637	-12.1372	a | task,
0.348214	-12.1374	for example, | only
0.490909	-12.1375	vectors | vector
0.474576	-12.1375	which | typically
0.509804	-12.1376	informatics | of
0.531915	-12.1379	with | users.
0.0831643	-12.1379	used | for
0.232	-12.1379	a | domain.
0.333333	-12.138	and | s.
0.265625	-12.1382	the | decoding.
0.72	-12.1382	czech | czech
0.390805	-12.1384	a | class,
0.454545	-12.1386	unsupervised | parsing
0.288344	-12.1386	given | target
0.292994	-12.1387	to predict | a
0.291139	-12.1389	proportion | in
0.179669	-12.1392	gold standard | the
0.72	-12.1393	prefix | prefix
0.313869	-12.1397	features | model.
0.156194	-12.14	7 | and
0.244444	-12.1401	each | type
0.703704	-12.1401	clean | raw
0.75	-12.1408	as | consequence
0.543478	-12.141	offers | for
0.509804	-12.141	performance | improved
0.178404	-12.1412	we | a new
0.402439	-12.1416	features | useful
0.611111	-12.1418	study | human
0.233871	-12.1419	language | english
0.3	-12.1423	since | different
0.153846	-12.1429	process | and
0.933333	-12.143	family | metrics
0.272222	-12.143	the | measures.
0.311594	-12.1431	defining | of
0.239316	-12.1436	history | the
0.636364	-12.1436	there has | been
0.179426	-12.1437	proceedings | conference on
0.72	-12.1442	had | highest
0.205607	-12.1443	the | again
0.395349	-12.1443	= | k=1
0.268817	-12.1445	estimates | the
0.284848	-12.1446	2008; | 2010).
0.772727	-12.145	reduction | size
0.16701	-12.1451	model | features
0.207668	-12.1455	and | algorithms
0.241379	-12.146	in general, | the
0.291139	-12.1461	the | period
0.26	-12.1461	directed | the
0.617647	-12.1461	distributed | distributed
0.421053	-12.1462	for | relation.
0.286585	-12.1463	varying | of
0.236515	-12.1469	the | read
0.882353	-12.1473	generating | summary
0.223048	-12.1473	individual | in
0.169528	-12.1478	span | the
0.482759	-12.1478	not | nor
0.375	-12.1485	the | is.
0.33913	-12.1487	the | reordered
0.882353	-12.1488	55 | 65
0.482759	-12.1488	central | is
0.229249	-12.1492	the | experiment.
0.842105	-12.1494	positive | negative,
0.195467	-12.1494	the | stage
0.246575	-12.1494	a | textual
0.243363	-12.1495	we | discuss
0.257426	-12.1495	datasets | and
0.367347	-12.1498	2007; | 2009;
0.147967	-12.1499	factor | the
0.342105	-12.1501	a | relation,
0.376344	-12.1502	assigned | each
0.72	-12.1502	were | filtered
0.103668	-12.1503	representation | the
0.367347	-12.1503	2007; | 2009).
0.531915	-12.1507	add | feature
0.678571	-12.1509	automatic | generation
0.325397	-12.151	allows | be
0.311594	-12.1514	the | in fig.
0.72	-12.1518	small | small
0.157407	-12.1518	with | with | with
0.141369	-12.152	is | very
0.407407	-12.1526	each | five
0.305556	-12.1527	of | research.
0.39759	-12.1527	counts | number
0.842105	-12.1529	textual | (rte)
0.5	-12.1531	translation, | translation
0.156364	-12.1532	have | have
0.453125	-12.1532	or | e.g.
0.254902	-12.1533	the | clusters.
0.268817	-12.1533	precision | precision
0.256158	-12.1534	can | without
0.190217	-12.1534	these | they
0.161165	-12.1534	under | and
0.490909	-12.1537	scores | those
0.474576	-12.1539	texts | text
0.367347	-12.1542	the | comprises
0.509804	-12.1546	extends | to
0.135431	-12.1547	performance | on
0.226923	-12.1548	novel | for
0.117769	-12.155	the | classes
0.594595	-12.1553	limiting | to
0.192201	-12.1558	a | relevant
0.15608	-12.1559	the | chosen
0.415584	-12.1561	for | entities.
0.15879	-12.1562	the | setting
0.115811	-12.1563	we | pairs
0.8	-12.1563	team | their
0.8	-12.1563	turning | into
0.8	-12.1563	scheme | similar
0.273743	-12.1565	this | process
0.560976	-12.1567	latter | more
0.490909	-12.1568	for | texts,
0.185567	-12.1569	first | which
0.131105	-12.1572	the | information.
0.0935864	-12.1572	examples | the
0.116923	-12.1572	the | (the
0.5	-12.1572	nouns | noun
0.147541	-12.1573	proceedings | on | pages
0.102686	-12.1574	an | which
0.678571	-12.1576	item | items
0.363636	-12.1577	that | missing
0.545455	-12.1578	daniel | and
0.20122	-12.1579	all | their
0.178571	-12.1581	interaction | the
0.251185	-12.1582	this | further
0.772727	-12.1582	r | r.
0.2875	-12.1583	is | finding
0.225191	-12.1585	in | context.
0.330579	-12.1585	and | categories.
0.37234	-12.1585	assignment | to
0.645161	-12.1587	increase | size
0.233607	-12.1588	note that | as
0.188679	-12.159	these | using
0.430556	-12.1591	concerning | of
0.168443	-12.1592	and | sentences.
0.247706	-12.1593	gives | to
0.14093	-12.1595	questions | the
0.2875	-12.1595	in | text,
0.37234	-12.1598	all | within
0.117954	-12.1601	was | with
0.289308	-12.1602	the | updates
0.263158	-12.1602	instructions | the
0.164969	-12.1608	this | based on
0.545455	-12.1608	score | calculated
0.386364	-12.1609	learning | patterns
0.133511	-12.1611	the | them.
0.531915	-12.1611	words | express
0.207792	-12.1611	thus, | is
0.575	-12.1612	function: | =
0.126812	-12.1614	the | string
0.203125	-12.1616	49th | the
0.181592	-12.162	interpretation | the
0.342105	-12.162	our | performs
0.8	-12.1621	2010 | 2010).
0.5	-12.1621	performs | well
0.560976	-12.1622	in | condition,
0.176056	-12.1623	(the | of
0.410256	-12.1624	is | equal to
0.560976	-12.1625	directions | future
0.5	-12.1626	corpora | languages
0.195402	-12.1631	all | only
0.156538	-12.1632	the | actually
0.441176	-12.1632	which | provide
0.274286	-12.1635	is | smaller
0.39759	-12.1636	in | dataset,
0.645161	-12.1637	2001; | 2004;
0.17226	-12.164	in | next
0.5	-12.1644	several | systems
0.933333	-12.1649	|c|
0.617647	-12.1649	models | across
0.392857	-12.165	other | (e.g.,
0.235294	-12.1652	will | more
0.12875	-12.1654	are | into
0.405063	-12.1655	and | at | level.
0.23913	-12.1655	difficulty | the
0.453125	-12.1657	the | database,
0.189189	-12.1658	to | them.
0.578947	-12.1658	contain | contain
0.933333	-12.1666	character | character.
0.617647	-12.1666	differ | they
0.313433	-12.1668	this | called
0.386364	-12.167	for each | corresponding
0.263158	-12.1676	indeed, | the
0.533333	-12.1676	by | factor of
0.127295	-12.1677	the | dictionary
0.203762	-12.1678	design | of
0.617647	-12.1678	a | comprising
0.466667	-12.1679	people | their
0.278107	-12.168	finds | the
0.37234	-12.1682	= | parenrightbig
0.73913	-12.1693	graph | graph,
0.418919	-12.1699	p. | and
0.617647	-12.1702	simplification | simplification
0.196481	-12.1704	a | matrix
0.304965	-12.1708	the | inference.
0.882353	-12.1708	bunescu | mooney,
0.882353	-12.1708	“x | y”
0.21978	-12.1719	of | models,
0.228	-12.1723	however, | more
0.424658	-12.1728	(2009) | using
0.482143	-12.1734	the | community,
0.520833	-12.1734	typically, | a
0.386364	-12.1739	which | contain
0.18136	-12.174	these | used
0.382022	-12.174	produces | for
0.6	-12.1741	vocabulary, | a
0.72	-12.1743	in | scenarios,
0.302817	-12.1749	= | (7)
0.230769	-12.1749	the | methods | language
0.205128	-12.1751	the | induced
0.246512	-12.1752	construction | a
0.352381	-12.1755	and | j.
0.4	-12.176	all | including
0.0551053	-12.1762	our | is
0.207921	-12.1762	similarity | between
0.25	-12.1763	each | single
0.73913	-12.1764	supposed | be
0.308824	-12.1767	synonyms | the
0.428571	-12.1767	a | contiguous
0.264865	-12.1769	some | but
0.678571	-12.177	train | model,
0.545455	-12.177	due | nature
0.205882	-12.1771	(2009) | a
0.645161	-12.1772	formula: | =
0.257576	-12.1774	but | different
0.154265	-12.1775	for | extraction
0.418919	-12.1778	for | extraction,
0.692308	-12.1779	are | related.
0.482143	-12.1781	in | 7,
0.174014	-12.1783	components | the
0.547619	-12.1785	amount | data,
0.4	-12.1785	annotated | corpus
0.352381	-12.1785	7 | 8
0.283951	-12.1792	approximately | of
0.108992	-12.1796	the | of | original
0.933333	-12.1798	competitive | state-of-the-art
0.678571	-12.1799	for | french-english
0.0702417	-12.1799	both | and | the
0.302817	-12.1799	5 | 15
0.617647	-12.1799	clusters | clustering
0.8	-12.1799	given | pair,
0.875	-12.1804	knn | crf
0.315385	-12.1812	2007; | 2008;
0.288462	-12.1812	a | underlying
0.290323	-12.1813	regarding | of
0.325203	-12.1813	probability of | given
0.349057	-12.1814	identify | or
0.330508	-12.1814	the | groups.
0.645161	-12.1815	a | point | view.
0.103001	-12.1815	each | is | a
0.692308	-12.1815	mwes | mwe
0.345794	-12.1818	that | never
0.547619	-12.1819	institute | usa
0.194203	-12.182	considering | of
0.188525	-12.182	all | one
0.428571	-12.1821	many | many
0.418919	-12.1825	figure 1 shows | and
0.8	-12.1825	joint | pipeline
0.16886	-12.1826	during | of
0.257732	-12.1828	use | algorithm
0.327731	-12.1829	a | lexicon.
0.311111	-12.1832	choose | for
0.0842391	-12.1836	when | a
0.189944	-12.1836	and | final
0.164583	-12.1837	because | not
0.875	-12.184	interpolation | eq.
0.266304	-12.1842	a | parser.
0.134831	-12.1842	and | corresponding
0.692308	-12.1845	consumer | on
0.0589176	-12.1846	performance | the
0.339286	-12.1846	in | pp.
0.304348	-12.1849	nouns, | and
0.0679957	-12.185	and | are | in
0.22619	-12.1851	each | corresponding
0.8	-12.1852	sms | twitter
0.418919	-12.1854	does | contain
0.327731	-12.1856	every | an
0.342342	-12.186	institute | university
0.115306	-12.1863	the | category
0.255	-12.1865	consisting of | of
0.139013	-12.1869	of | sentences.
0.655172	-12.187	observe | improvement
0.280488	-12.1872	for | partial
0.215054	-12.1875	this | do not
0.274854	-12.1875	of | types,
0.171233	-12.1882	3 | our
0.6	-12.1885	cannot | because
0.271676	-12.1886	using | standard
1	-12.1887	(hu | liu,
1	-12.1887	(noun | sentence)
1	-12.1887	epigenetics | post-translational
0.459016	-12.1889	the | decides
0.928571	-12.1891	causal | causal
0.252475	-12.1892	is | enough
0.377778	-12.1892	but | many
0.560976	-12.1893	in-domain | data
0.115702	-12.1896	this | has
0.73913	-12.1897	nodes, | nodes
0.339286	-12.1898	of | choices
0.422535	-12.19	dictionary | words
0.248804	-12.19	expected | in
0.617647	-12.19	an | lexicon.
0.16849	-12.1901	indicates | of
0.271676	-12.1901	format | the
0.17866	-12.1903	a | whose
0.165957	-12.1904	for both | and
0.183246	-12.1905	in | thus
0.183246	-12.1905	estimation | the
0.465517	-12.1908	conclude | section
0.128535	-12.1912	include | of
0.5	-12.1912	comparing | different
0.547619	-12.1913	abstract | problem
0.31746	-12.1914	develop | for
0.266667	-12.1919	automated | of
0.147458	-12.1919	parser | a
0.928571	-12.1922	issues | addressed
0.533333	-12.1924	n-best | n-best
0.0802993	-12.1929	the | structure
0.833333	-12.1932	reasonable | assume
0.73913	-12.1935	methods | been proposed
0.428571	-12.1941	tuned | on
0.655172	-12.1942	his | he
0.127364	-12.1942	( | (
0.125152	-12.1945	general | of
0.271676	-12.1945	authors | a
0.245283	-12.1946	evaluation | system
0.73913	-12.1949	combined | any
0.106608	-12.195	the | contain
0.217712	-12.1952	in | experiment
0.428571	-12.1958	classified | or
0.692308	-12.1958	senses | words.
0.465517	-12.196	100 | 400
0.625	-12.1961	english | english.
0.32	-12.1962	a | representation.
0.388235	-12.1968	2011, | oregon,
0.459016	-12.1969	intuitive | to
0.195783	-12.197	of | documents.
0.0842047	-12.1973	the | 1.
0.578947	-12.1973	predicates | predicates
0.0683388	-12.1975	the | the | set.
0.209622	-12.1975	a | sample
0.287582	-12.1979	optimal | for
0.928571	-12.1982	subjects | objects
0.219697	-12.1983	the | framework.
0.5	-12.1986	are | here.
0.439394	-12.1987	extract | rules
0.210345	-12.1987	different | each
0.521739	-12.199	only | whose
0.212766	-12.1993	the | distance between
0.183246	-12.1995	this, | the
0.73913	-12.1996	refer | entity.
0.175904	-12.2	is computed | the
0.833333	-12.2003	indicator | argument
0.416667	-12.2004	the | tests.
0.390244	-12.2006	model | predict
0.625	-12.2009	when | available.
0.655172	-12.201	active | passive
0.190883	-12.201	words | have
0.278788	-12.2011	the | itself,
0.193548	-12.2012	set | sentences
0.451613	-12.2013	each | pair,
0.413333	-12.2014	5 experiments | and
0.373626	-12.2017	is: | =
0.225806	-12.2019	for | translation.
0.521739	-12.2019	mentions | entities
0.155303	-12.2021	test | with
0.439394	-12.2024	effects | on
0.451613	-12.2025	out | out
0.692308	-12.2032	statistical | systems.
0.534884	-12.2034	and | 1997)
0.127389	-12.2034	note that | of
0.308271	-12.2041	present | system
0.432836	-12.2043	general | can be
0.402597	-12.2043	although | only
0.2249	-12.2044	the | foreign
0.186813	-12.2045	2.1 | of
0.156371	-12.2048	framework | and
0.217228	-12.2058	of | occurrences
0.325	-12.2058	in | area
0.390244	-12.2059	purpose | is
0.457627	-12.2059	we | write
0.373626	-12.206	the | up.
0.928571	-12.2062	theoretical | empirical
0.875	-12.2063	meanings | meanings
0.290541	-12.2073	1 introduction | machine
0.451613	-12.2073	4.3 | our
0.357143	-12.2074	use | instead
0.655172	-12.2075	before | first
0.239819	-12.2079	set | rules
0.349515	-12.2079	taggers | the
0.360825	-12.2082	reordering | source
0.37931	-12.2082	related | related
0.439394	-12.2083	contribute | to
0.270115	-12.2086	decided to | the
0.416667	-12.2088	extend | by
0.416667	-12.2089	in | steps.
0.73913	-12.2089	becomes | when
0.373626	-12.2091	use | linguistic
0.247573	-12.2092	the | gold-standard
0.277778	-12.2097	these | types
0.390244	-12.21	d. | and
0.606061	-12.2104	pcfg | pcfg
0.444444	-12.2104	web | search
0.18254	-12.2105	aspect | the
0.127907	-12.2108	is | such as
0.244131	-12.2108	introduce | for
0.136905	-12.2111	column | the
0.0873494	-12.2112	the | generated
0.0745287	-12.2113	was | in
0.481481	-12.2114	size of | size
0.24186	-12.2116	this | useful
0.279503	-12.2117	second | which
0.510638	-12.2119	into | classes
0.346154	-12.212	all | values
0.37931	-12.212	a | verb,
0.708333	-12.212	after | final
0.547619	-12.2121	propose | framework
0.236607	-12.2123	is | sum
0.761905	-12.2123	rhetorical | theory
0.172986	-12.2123	the | evaluation.
0.155172	-12.2125	represent | and
0.708333	-12.2125	oracle | oracle
0.186301	-12.2128	that | often
0.692308	-12.2128	outputs | outputs
0.583333	-12.2129	trees | trees.
0.606061	-12.2131	dependency | (de
0.451613	-12.2132	in | graphs.
0.26087	-12.2136	trained | model
0.625	-12.2143	decoding | hypotheses
0.23176	-12.2153	1) | a
0.0927694	-12.2153	finally, | the
0.21147	-12.2153	a | i.e.
0.3125	-12.2157	all | systems
0.181579	-12.2161	a single | to
0.18663	-12.2161	not | because
0.875	-12.2162	computing | computing
0.297872	-12.2164	a | decoder
0.283871	-12.2166	in | framework.
0.0902597	-12.2171	the | space
0.547619	-12.2174	for | comparison,
0.131761	-12.2177	through | of
0.606061	-12.2182	contribution | different
0.73913	-12.2186	types, | types
0.692308	-12.2186	languages, | languages
0.142626	-12.2186	results | results | the
0.213768	-12.2187	only | all
0.0662483	-12.2187	can be | of
0.432836	-12.2188	these | represent
0.208333	-12.2188	the | increased
0.204698	-12.2188	performance | using
0.16895	-12.2189	include | in
0.789474	-12.2189	sequence | actions
0.135371	-12.2189	articles | the
0.322314	-12.2192	the | frame.
0.14841	-12.2194	choice | the
0.324786	-12.2195	are | e.g.
0.12844	-12.2197	uses | and
0.875	-12.2197	de-en | es-en
0.185596	-12.2199	for | query
0.928571	-12.2199	users | themselves
0.187151	-12.2201	provides | to
0.481481	-12.2202	method | automatically
0.357143	-12.2203	senses | word
0.397436	-12.2203	some | specific
0.490196	-12.2204	our | handle
0.218391	-12.2207	columns | the
0.928571	-12.2213	using | regression.
0.175309	-12.2217	the | database
0.085168	-12.2221	the | under
0.564103	-12.2223	a | scheme.
0.0996835	-12.2226	instances | the
0.0918919	-12.2226	that | used
0.510638	-12.2227	has | highest
0.162105	-12.2227	maximum | of
0.111888	-12.2229	0 | 1
0.317073	-12.223	rules | used
0.310078	-12.223	a | phrase.
0.365591	-12.223	the | marker
0.123472	-12.223	the | of | under
0.276074	-12.2232	four | are
0.55	-12.2232	substantially | than
0.390244	-12.2234	for | corpora,
0.276074	-12.2237	a | unit
0.235556	-12.2237	the | connective
0.224	-12.2237	operation | the
0.35	-12.2239	text | its
0.295775	-12.2241	et | 2003),
0.27439	-12.2243	do | end
0.224	-12.2244	choice of | the
0.583333	-12.2247	recognition | recognition
0.297872	-12.2252	costs | the
0.14527	-12.2253	no | is
0.175743	-12.2253	is | sequence
0.928571	-12.226	qb c | .
0.280255	-12.2266	our work | in
0.0644518	-12.2266	these | in
0.261111	-12.227	investigated | of
0.426471	-12.227	its | form
0.45	-12.2271	at | position
0.375	-12.2272	for | sets,
0.42029	-12.2276	by | authors
0.222222	-12.2277	5 | 5
0.390244	-12.2279	of | n,
0.55	-12.2279	different | types.
0.55	-12.2285	yields | results
0.286667	-12.2288	such as | have
0.0353876	-12.2293	as | in
0.185083	-12.2294	e.g., | and
0.241706	-12.2294	hypothesis | that
0.265537	-12.2295	in | tasks.
0.633333	-12.2296	vocabulary | words.
0.4375	-12.23	the | θ.
0.353535	-12.2301	for example, | verb
0.606061	-12.2301	compare | performance | with
0.172662	-12.2302	in | more | than
0.212454	-12.2302	the | fragment
0.875	-12.2304	existing | tools
0.291667	-12.2304	the | ontonotes
0.480769	-12.2309	that | closer
0.339623	-12.2314	a | simpler
0.312	-12.2316	the | applied.
0.666667	-12.2317	development | testing
0.583333	-12.2317	achieve | improvement
0.157058	-12.2317	in | languages
0.231441	-12.2319	traditional | and
0.928571	-12.2322	names | names.
0.330357	-12.2324	the | decisions.
0.307692	-12.2325	training | classifier
0.45	-12.2325	were | during
0.148746	-12.2326	note that | in
0.181333	-12.2328	we | identify
0.154739	-12.2329	the | 2005).
0.875	-12.2329	7 5 | .
0.108262	-12.2331	what | the
0.55	-12.2332	if | occurs
0.0884397	-12.2335	was | for
0.0621487	-12.2337	system | and
0.490196	-12.2337	and | (2005),
0.165919	-12.234	the | arabic
0.196262	-12.234	the | co-occurrence
0.45	-12.2342	this | reduces
0.666667	-12.2345	follow | same
0.208481	-12.2345	for | full
0.192771	-12.2346	and | information.
0.761905	-12.235	is | general.
0.606061	-12.2355	a | traversal
0.265896	-12.2357	proceedings | 27–31,
0.137195	-12.2359	that | similar
0.789474	-12.2359	ideally, | would
0.120187	-12.2359	from | data
0.198718	-12.2362	a | dictionary
0.291667	-12.2365	greatly | the
0.185393	-12.2366	however, | this
0.4	-12.2366	which | occur
0.5	-12.2368	for | processing.
0.174564	-12.2369	in | language.
0.55	-12.237	that | harder
0.158537	-12.2372	summationdisplay | summationdisplay
0.0675526	-12.2373	we | two
0.147687	-12.2376	note that | and
0.161017	-12.2377	has | it
0.375	-12.2378	type | event
0.181333	-12.238	to | performance.
0.761905	-12.238	very | especially
0.207018	-12.2386	and, | the
0.199357	-12.2387	example | with
0.510638	-12.2387	summarization | summary
0.366667	-12.2387	statistical significance | the
0.14711	-12.239	the | document.
0.727273	-12.2391	horizontal | vertical
0.255319	-12.2391	and | methods.
0.160417	-12.2394	the | summaries
0.522727	-12.2395	in | ranking.
0.567568	-12.2395	from | (a
0.234234	-12.2395	research | research
0.366667	-12.2396	these | can | be
0.522727	-12.2398	many | processing
0.221774	-12.2399	word | context
0.823529	-12.24	histograms | histograms
0.139461	-12.24	match | the
0.202703	-12.2401	we | (section
0.385542	-12.2402	motivation | is
0.154599	-12.2403	correctly | the
0.197452	-12.2405	condition | the
0.324786	-12.2405	the | tagger.
0.238318	-12.2407	they | different
0.231441	-12.2408	has | its
0.149909	-12.241	the | sequences
0.727273	-12.2411	innovative | 2011.
0.727273	-12.2413	regardless | whether
0.823529	-12.2413	in-domain | in-domain
0.0972973	-12.2413	methods | and
0.875	-12.2417	factors | affect
0.472727	-12.2419	connecting | to
0.102215	-12.242	also | as
0.322034	-12.2421	the | varies
0.124524	-12.2425	the | group
0.633333	-12.2428	13 | 14
0.823529	-12.2431	petrov | klein
0.442623	-12.2431	rate | rate
0.45614	-12.2433	used | represent
0.392405	-12.2433	name | which
0.322034	-12.2433	a | b.
0.139461	-12.2434	a | labeled
0.606061	-12.2436	is | large,
0.236364	-12.2436	the | used,
0.583333	-12.2437	at | start
0.235294	-12.2437	the | weights.
0.195584	-12.2439	a | context.
0.272727	-12.2445	of | versus
0.4375	-12.2448	mentions | entity
0.380952	-12.2449	x | z
0.309524	-12.2453	include | which
0.161359	-12.2455	basic | of
0.633333	-12.2458	core | core
0.238318	-12.2461	explicit | of
0.322034	-12.2463	the | same.
0.267857	-12.2464	are | under
0.234234	-12.2467	model | best
0.157464	-12.2469	for | classification
0.324561	-12.2472	question, | the
0.316667	-12.2473	the | factors.
0.380952	-12.2476	among | or
0.336449	-12.2476	present | method
0.362637	-12.2481	summarized | table
0.0885183	-12.2482	use | as
0.137134	-12.2483	a | model,
0.708333	-12.2485	take | take
0.252632	-12.2488	set | possible
0.261364	-12.2492	are | relevant
0.330357	-12.2494	directly | from
0.928571	-12.2495	modification | modification
0.159329	-12.2498	significant | in
0.666667	-12.2499	let | represent
0.376471	-12.2503	existence | a
0.217899	-12.2508	already | in
0.536585	-12.251	classifiers | classifier
0.727273	-12.2511	compare | approaches.
0.144082	-12.2511	in | 1,
0.18732	-12.2517	of | queries
0.312	-12.2519	our | achieves
0.408451	-12.2522	individual | each
0.108317	-12.2522	the | (e.g.
0.471698	-12.2522	tags | tag
0.612903	-12.2522	in | direction,
0.118255	-12.2523	our approach | the
0.0785606	-12.2524	the | distribution
0.275	-12.2524	all | them
0.42029	-12.2526	only | available
0.42029	-12.2527	require | or
0.124051	-12.2528	of | available
0.175573	-12.2528	response | the
0.29078	-12.2529	the | semantics.
0.346535	-12.253	j. | and
0.789474	-12.2533	models, | models,
0.165909	-12.2533	as | source
0.588235	-12.2538	sentences, | sentences.
0.612903	-12.2539	in | column.
0.158333	-12.254	which | have
0.333333	-12.2542	set | sets
0.376471	-12.2543	in fact, | is
0.789474	-12.2544	converted | into
0.0977918	-12.2546	the | of | model.
0.366667	-12.2548	tree | nodes
0.273292	-12.2548	that | too
0.238318	-12.255	the same | used
0.154455	-12.2551	figure 2 | the
0.266272	-12.2554	to | noisy
0.408451	-12.2556	(see | 2).
0.0859232	-12.2556	using | for
0.372093	-12.2557	parser | parsing
0.489796	-12.2559	are | helpful
0.358696	-12.2559	toward | of
0.394737	-12.2559	the | effect.
0.0700405	-12.2561	the | of | target
0.168618	-12.2561	means that | the
0.394737	-12.2562	then | them
0.55	-12.2563	in | conversation,
0.435484	-12.2564	theory | that
0.30303	-12.2564	use | standard
0.642857	-12.2564	single | per
0.346535	-12.2568	to improve | by
0.823529	-12.257	while | remains
0.68	-12.257	descriptions | descriptions
0.114224	-12.2571	of | within
0.163355	-12.2571	the | intended
0.376471	-12.2573	definition | which
0.236111	-12.2574	the | old
0.31405	-12.2575	to | if they
0.287671	-12.2575	that is, | is
0.430769	-12.2579	in | immediate
0.511111	-12.2582	spanish | english
0.21374	-12.2584	example | figure
0.324561	-12.2586	for | pair.
0.511111	-12.2589	still | but
0.866667	-12.2591	university | boulder
0.41791	-12.2593	we obtain | for
0.666667	-12.2594	predicting | predicting
0.4	-12.2595	title | a
0.823529	-12.2598	articles | prepositions
0.464286	-12.2599	tree | grammar
0.25	-12.2599	to | paraphrases
0.522727	-12.2599	natural | processing.
0.0944649	-12.26	these | as
0.480769	-12.2601	(2003) | for
0.222222	-12.2601	the | accuracies
0.311475	-12.2601	in | affect
0.0850809	-12.2602	the | parser
0.0781329	-12.2603	part | the
0.164414	-12.2604	allow | of
0.68	-12.2605	readability | readability
0.350515	-12.2608	of | variations
0.424242	-12.2611	maintain | of
0.480769	-12.2612	that | captured
0.187135	-12.2614	the | location
0.150094	-12.2618	the | components
0.245	-12.2621	analyze | of
0.302326	-12.2623	the | improvement.
0.823529	-12.2627	university | massachusetts
0.256983	-12.2627	different | their
0.928571	-12.2628	(13) | (14)
0.612903	-12.2628	using | berkeley
0.318966	-12.2629	which | output
0.633333	-12.263	smt | smt
0.382716	-12.2635	20 | 100
0.147541	-12.2635	techniques | and
0.666667	-12.2636	initialize | with
0.789474	-12.2637	evaluate | performance.
0.158898	-12.2638	meaning | a
0.362637	-12.2639	another | is that
0.22314	-12.264	extraction | from
0.448276	-12.2642	more | see
0.489796	-12.2644	column | column
0.184659	-12.2644	of | entire
0.378049	-12.2646	phrase-based | model
0.823529	-12.2652	developed | specific
0.708333	-12.2652	appears | times
0.45614	-12.2653	most | approaches
0.233945	-12.2654	results | three
0.288732	-12.2655	(2010), | and
0.129346	-12.2655	summary | the
0.281879	-12.2655	of | american
0.343137	-12.2655	in | positions
0.633333	-12.2656	some | e.g.
0.195513	-12.2663	translation | source
0.350515	-12.2665	word | a | word.
0.283784	-12.2665	quickly | the
0.115861	-12.2666	example | a
0.120292	-12.2669	is | because
0.042621	-12.2673	the | number of
0.256983	-12.2673	english, | and
0.150094	-12.2674	way | to
0.38961	-12.2674	offer | to
0.642857	-12.2674	cause | cause
0.612903	-12.2674	submit | to
0.156701	-12.2677	rest | the
0.242574	-12.2679	error analysis | the
0.187135	-12.2681	for | words.
1	-12.2687	2nd | cognitive
1	-12.2687	(strapparava | valitutti,
1	-12.2687	monolingual | generation,
1	-12.2687	marina | ca
0.5	-12.2687	research | nsf
0.552632	-12.2688	but | improvements
0.232877	-12.2689	a | language,
0.923077	-12.2692	pointwise mutual | (pmi)
0.923077	-12.2692	ip@null | [vp@2null2
0.124837	-12.2692	study | of
0.159236	-12.2697	the | row
0.196141	-12.2697	instead, | the
0.186589	-12.2697	weighted | of
0.244898	-12.2698	of | units
0.254054	-12.2699	however, | may
0.75	-12.27	task | classifying
0.727273	-12.2702	document | mixture
0.511111	-12.2702	incorporate | information
0.259887	-12.2702	the | coherence
0.435484	-12.2703	machine | (smt)
0.22807	-12.2703	y | x
0.133433	-12.2712	thus, | of
0.727273	-12.2712	requires | time
0.442623	-12.2714	= | (9)
0.435484	-12.2718	say that | is
0.298507	-12.2721	on | task,
0.382716	-12.2722	exploring | of
0.612903	-12.2722	counting | number
0.190769	-12.2724	of | whole
0.75	-12.2727	noun, | noun
0.223629	-12.2728	of | entries
0.321739	-12.2728	adopted | and
0.321739	-12.2728	exploit | in
0.255556	-12.2729	that | highest
0.571429	-12.2732	for | mining.
0.38961	-12.2733	individual | or
0.284722	-12.2733	the | reached
0.567568	-12.2734	building | corpus
0.31405	-12.2734	is | valid
0.588235	-12.2736	languages: | and
0.552632	-12.2737	and | speakers.
0.188623	-12.2737	idea | a
0.172932	-12.2744	a | range
0.75	-12.2744	term | terms.
0.552632	-12.2748	words | words)
0.536585	-12.2748	which | quite
0.536585	-12.2751	which | available.
0.588235	-12.2754	introduced | new
0.402778	-12.2756	the | change.
0.278146	-12.2757	where | only
0.45614	-12.2758	into | sequence
0.511111	-12.2761	clearly, | is
0.102023	-12.2761	a | tree
0.358696	-12.2763	i | you
0.177898	-12.2763	model | only
0.552632	-12.2765	if | word,
0.306452	-12.2768	morphology | and
0.261628	-12.2768	of | entities.
0.214844	-12.277	a | could be
0.0994983	-12.2773	here | the
0.123711	-12.2773	we | our | on
0.408451	-12.2775	the | decoder,
0.408451	-12.2775	modeling | model
0.789474	-12.2775	we used | default
0.172589	-12.2776	of | sentences,
0.144621	-12.2781	for | systems
0.318966	-12.2784	the | in.
0.209738	-12.279	primary | of
0.536585	-12.2793	a | solution.
0.440678	-12.2794	intended | be
0.282759	-12.2796	language | text
0.0771948	-12.2799	after | the
0.205776	-12.2799	a | match
0.330189	-12.2799	and | d.
0.177419	-12.2799	a | point
0.137987	-12.28	the | required
0.318966	-12.2802	one | possible
0.823529	-12.2804	experimental results | significantly
0.358696	-12.2804	subject | verb
0.193038	-12.281	the | graph.
0.536585	-12.281	called | or
0.75	-12.2811	compared | traditional
0.823529	-12.2813	after | step,
0.167464	-12.2813	verbs | and
0.194805	-12.2818	and | represent
0.351064	-12.2818	pmi | and
0.211538	-12.2819	baseline, | the
0.789474	-12.282	languages | italian
0.190769	-12.2824	noise | the
0.923077	-12.2824	estimating | parameters
0.448276	-12.2824	used | nlp
0.29771	-12.2825	our | framework
0.306452	-12.2826	a | output.
0.0882159	-12.2826	note that | the
0.182336	-12.2828	also, | the
0.289855	-12.2829	consequently, | the
0.180055	-12.2832	decoder | the
0.428571	-12.2838	in | lexicon,
0.185294	-12.2838	the | ontology
0.359551	-12.2839	therefore, | not
0.198653	-12.284	we show that | the
0.378049	-12.284	for | english,
0.224138	-12.2841	probability | given
0.135303	-12.2841	to | correct
0.52381	-12.2842	300 | 500
0.179558	-12.2843	a | similar to
0.435484	-12.2843	differences | between
0.923077	-12.2844	resources | wordnet,
0.5	-12.2845	(koehn | 2003).
0.68	-12.2846	malt | malt
0.268293	-12.2846	baseline | all
0.428571	-12.2847	y | z
0.219008	-12.2855	added | to
0.642857	-12.2856	gains | when
0.29771	-12.2856	that | predicted
0.5	-12.2856	learning | inference
0.27451	-12.2857	similarly, | and
0.207407	-12.2858	though | is
0.27451	-12.2859	type | which
0.411765	-12.2859	end | return
0.321429	-12.2861	binary | with
0.612903	-12.2861	paraphrase | monolingual
0.428571	-12.2863	grateful | for
0.278146	-12.2867	then, | is
0.308943	-12.2871	parse | sentence
0.239024	-12.2874	judges | the
0.866667	-12.2874	classifiers | classifiers.
0.3	-12.2876	2006; | 2008;
0.448276	-12.2876	it, | a
0.204301	-12.2877	for | detection
0.489796	-12.2877	consider | possible
0.252747	-12.2879	of | verbal
0.0687575	-12.2882	we have | the
0.642857	-12.2882	an | operator
0.26875	-12.2884	training | development
0.48	-12.2888	expect | will
0.351064	-12.2889	each | token
0.2287	-12.289	for | words,
0.15	-12.2891	the | run
0.428571	-12.2901	we | seven
0.48	-12.2902	this | produces
0.15812	-12.2902	natural language | and
0.369048	-12.2903	consists | two
0.363636	-12.2905	tokens, | and
0.276316	-12.2906	and | texts.
0.252747	-12.2909	et | 2011)
0.333333	-12.2915	languages | language
0.131462	-12.2916	sample | the
0.428571	-12.2918	all | candidates
0.074856	-12.292	this | which
0.823529	-12.2922	types | transformations
0.354839	-12.2922	of | onto
0.823529	-12.2925	parsing | dependencies
0.369048	-12.2926	matrices | and
0.0383938	-12.2928	results | the
0.0974817	-12.2929	one | that
0.241206	-12.293	in | space.
0.470588	-12.2932	with | parameters.
0.421875	-12.2932	after | first
0.330189	-12.2934	a | memory
0.179063	-12.2936	we | need
0.187879	-12.2938	that | able to
0.384615	-12.2938	using | web
0.642857	-12.2938	haitian | translation
0.289855	-12.294	ranks | the
0.866667	-12.294	1 introduction | past
0.355556	-12.2941	our | collection
0.137318	-12.2943	to | problem
0.351064	-12.2944	in | style
0.182336	-12.2945	experiment | a
0.205036	-12.2946	that | content
0.727273	-12.2948	this | harder
0.285714	-12.2948	of | paths
0.511628	-12.295	the | paradigm.
0.823529	-12.2951	spatial | rhetorical
0.318584	-12.2955	demonstrates | of
0.402778	-12.2955	our | predict
0.15528	-12.2955	to determine | the
0.104186	-12.2957	we have | a
0.160356	-12.2958	the | space.
0.218107	-12.2961	highly | and
0.133641	-12.2961	the | path
0.204301	-12.2964	we | define
0.272727	-12.2974	the | techniques.
0.223176	-12.2974	may | different
0.62069	-12.2974	europarl | (koehn,
0.359551	-12.2975	are | problems
0.26875	-12.2976	participant | the
0.171285	-12.2976	containing | and
0.727273	-12.2977	kneser-ney | (chen
0.52381	-12.2979	in | dialogues.
0.369048	-12.298	(lin | al.,
0.555556	-12.2983	brings | to
0.18314	-12.2986	we | consider
0.39726	-12.2991	" | "
0.0898955	-12.2991	in | first
0.231481	-12.2996	the | here.
0.555556	-12.2998	richard | and
0.280822	-12.3002	model | standard
0.823529	-12.3005	mean | rank
0.866667	-12.3005	precision | muc
0.215447	-12.3006	order | which
0.152305	-12.301	that | human
0.405797	-12.3011	an interesting | of
0.214286	-12.3012	classification | as
0.571429	-12.3013	evaluation | evaluation.
0.462963	-12.3014	user | users
0.122108	-12.3017	from | using
0.202847	-12.3018	5.2 | of
0.62069	-12.3018	estimate | distribution
0.277027	-12.3018	16 | of
0.327103	-12.3024	but | so
0.16	-12.3024	when | or
0.0813288	-12.3028	the | candidate
0.118437	-12.3029	following | in
0.347368	-12.303	that | gain
0.454545	-12.3031	influenced by | of
0.124	-12.3031	token | the
0.777778	-12.3031	problem | addressed
0.489362	-12.3034	et | 2010a).
0.373494	-12.3035	100 | each
0.415385	-12.3036	because | very
0.218107	-12.3036	used | other
0.192926	-12.3036	no | are
0.248649	-12.304	additionally, | of
0.146568	-12.3045	aspects | the
0.147448	-12.3045	model | it
0.866667	-12.3045	hybrid | statistical
0.355556	-12.3046	automatic | using
0.39726	-12.3049	analysis | errors
0.327103	-12.305	systems | use
0.149225	-12.3053	of | systems.
0.258824	-12.3053	co-occurrence | and
0.318584	-12.3054	(2004) | a
0.923077	-12.3058	decoder | produce
0.327103	-12.3065	null, | null
0.219409	-12.3065	s, | the
0.155136	-12.3066	therefore | of
0.285714	-12.3066	second, | that
0.216327	-12.3066	lists | and
0.336634	-12.3067	first, | each
0.22467	-12.3068	each | three
0.191693	-12.3069	we | a novel
0.181034	-12.3069	general | for
0.283688	-12.3075	evaluation | different
0.21371	-12.3079	occurrence | the
0.0553486	-12.308	has | and
0.149805	-12.308	set of | set
0.251397	-12.3082	database | and
0.0916422	-12.3087	words | are
0.288889	-12.3087	7 | .
0.48	-12.309	corpus | german
0.232227	-12.309	a | scheme
0.695652	-12.3095	at | point.
0.923077	-12.3095	(as | (as
0.310924	-12.31	only | no
0.262195	-12.3101	the | instance.
0.575758	-12.3102	in | dimensions.
0.59375	-12.3103	using | approach,
0.315789	-12.3108	and | values.
0.538462	-12.3108	is | maximal
0.270968	-12.3111	= | (5)
0.327103	-12.3111	submission | the
0.194079	-12.3115	influence | the
0.525	-12.3117	rule | derivation
0.226244	-12.312	our | outperforms
0.461538	-12.3121	the | reviewers | comments
0.575758	-12.3122	yields | best
0.75	-12.3122	tuning | weights
0.923077	-12.3122	w | w.
0.75	-12.3128	model; | model
0.151822	-12.3131	the | written
0.299213	-12.3136	simple | or
0.235294	-12.3139	in | 5,
0.653846	-12.314	i | ii
0.347826	-12.3143	and | knowledge.
0.14902	-12.3143	(2010) | and
0.146617	-12.3143	c©2011 association | university
0.211765	-12.3145	obtained | with
0.233333	-12.3147	results | show
0.5	-12.3147	and | darpa,
0.25	-12.3152	groups | and
0.695652	-12.3152	probability | probability.
0.5	-12.3152	basic | basic
0.162791	-12.3153	only | or
0.571429	-12.3153	60 | 120
0.347826	-12.3157	this | includes
0.230047	-12.316	report | for
0.210117	-12.3161	the | arguments.
0.310345	-12.3163	the resulting | are
0.653846	-12.3163	we calculate | using
0.360465	-12.3164	we | maximum entropy
0.172775	-12.3167	results, | the
0.470588	-12.3167	proceedings | june,
0.5	-12.3168	involve | or
0.59375	-12.3169	conversely, | is
0.59375	-12.3171	make | available
0.223684	-12.3171	symbols | the
0.114583	-12.3173	human | and
0.226244	-12.3174	that | had
0.218487	-12.3175	technique | to
0.235294	-12.3177	contributions | the
0.135922	-12.3178	amount of | the
0.25731	-12.3178	in | review
0.75	-12.3178	use | decoding.
0.138983	-12.3178	initial | of
0.0415655	-12.3181	the | system
0.454545	-12.3185	just | but
0.489362	-12.3186	chart | with
0.0569801	-12.3186	with | with
0.27972	-12.3188	10 | 30 | 50
0.381579	-12.3188	create | each
0.126761	-12.319	in | 2010).
0.212	-12.3191	predicted | and
0.351648	-12.3194	all | tokens
0.538462	-12.3194	dependency | (mcdonald
0.923077	-12.3194	many | 1998;
0.182891	-12.3195	is | what
0.147228	-12.3195	in | example
0.1875	-12.3195	weights | to
0.23301	-12.32	structural | of
0.391892	-12.3201	different | sets.
0.260606	-12.3204	medical | and
0.241026	-12.3207	a | degree
0.866667	-12.3208	memory | memory
0.139415	-12.321	a | list
0.0802048	-12.321	the | label
0.300813	-12.3212	contain | or
0.479167	-12.3212	hope | this
0.183735	-12.3214	score, | the
0.195286	-12.3214	the | texts.
0.190323	-12.3217	a | mt
0.575758	-12.3221	a | post.
0.162736	-12.3222	focus on | the
0.5	-12.3223	where, | and
0.347826	-12.3224	en | en
0.8125	-12.3224	hours | hours
0.461538	-12.3225	did not | our
0.288889	-12.3225	although | have
0.525	-12.323	system | learn
0.283688	-12.3231	the latter | a
0.15121	-12.3232	representation | a
0.235294	-12.3234	and | properties
0.375	-12.3234	rank | by
0.37037	-12.3235	our method | by
0.6	-12.3237	constraint | relation
0.8125	-12.3238	love | hate
0.313043	-12.324	our | existing
0.142086	-12.3242	their | on
0.866667	-12.3246	semantic relatedness | semantic
0.340206	-12.3247	extensive | and
0.292308	-12.3247	from | general
0.75	-12.3249	aspects | reviews
0.415385	-12.325	and | values,
0.172414	-12.325	clearly | the
0.8125	-12.325	composition | composition
0.190323	-12.3251	evaluation, | the
0.28169	-12.3252	true | true
0.433333	-12.3252	it is | difficult
0.75	-12.3255	interface | interface
0.212851	-12.3257	table 1: | in
0.330097	-12.3257	the | hierarchy.
0.152893	-12.3257	when | from
0.218884	-12.3258	occurring | the
0.4	-12.326	and | representations.
0.197232	-12.3261	the | game
0.446429	-12.3264	implemented | which
0.62069	-12.3265	statistically significant | over
0.176796	-12.3266	define | as
0.923077	-12.3267	 | 0
0.511628	-12.3269	al., | blunsom
0.433333	-12.3269	is | 2000).
0.101928	-12.3271	a | task
0.203704	-12.3271	h | a
0.145833	-12.3272	often | in
0.653846	-12.3274	transfer | transfer
0.386667	-12.3274	for | i.
0.525	-12.3275	baseline | random
0.438596	-12.3276	this | why
0.303279	-12.3276	patient | the
0.248619	-12.3276	put | the
0.255814	-12.3277	while | at
0.145833	-12.328	the | variable
0.163507	-12.328	we | use | as
0.32381	-12.3281	model | probabilities
0.340206	-12.3282	because | may
0.130699	-12.3288	the | equation
0.264151	-12.3288	the | steps.
0.540541	-12.3289	proper | proper
0.164649	-12.3289	et | 2006).
0.415385	-12.3291	from | external
0.262195	-12.3291	the | questions.
0.259036	-12.3294	what | are
0.461538	-12.3294	let us | in
0.438596	-12.3297	column | table
0.135621	-12.3297	of | text.
0.245902	-12.3297	probabilistic | for
0.248619	-12.3297	to | translations.
0.148438	-12.33	of | word.
0.340206	-12.33	strong | for
0.117216	-12.3304	results | our
0.315315	-12.3305	6 conclusion | of
0.313043	-12.3305	intuitively, | to
0.360465	-12.3307	these | often
0.333333	-12.3307	however, | very
0.386667	-12.3309	character | word
0.227907	-12.3309	n | 1
0.540541	-12.331	task | identification
0.121854	-12.3315	the | predicted
0.167082	-12.3316	features | model
0.438596	-12.3317	closest | to
0.211155	-12.3318	observations | the
0.235	-12.332	the | criterion
0.6	-12.332	advantage | over
0.415385	-12.3321	summationdisplay | (1)
0.555556	-12.3324	of | derivations.
0.184615	-12.3325	the | heuristic
0.267516	-12.3326	theory | and
0.180758	-12.3326	hence, | the
0.0412621	-12.3327	the | of | word
0.525	-12.3329	applied | tasks
0.300813	-12.3331	describing | of
0.10865	-12.3332	the | name
0.225806	-12.3333	performance, | the
0.0993843	-12.3335	we have | and
0.242105	-12.3335	that | already
0.277778	-12.3336	4.3 | we
0.257485	-12.3337	background | and
0.62069	-12.3337	parsers | parsing
0.206897	-12.3337	the | cover
0.777778	-12.334	evaluating | against
0.218884	-12.3344	of | underlying
0.0733461	-12.3348	given | and
0.273973	-12.3349	topics, | the
0.273333	-12.3352	where | were
0.32381	-12.3353	apply | on
0.488889	-12.3355	official | for
0.469388	-12.3356	train | model.
0.207843	-12.3358	the | engine
0.240838	-12.3358	a | response
0.178161	-12.336	therefore | a
0.540541	-12.3361	node | parent
0.0846154	-12.3362	output | the
0.365854	-12.3362	zero | in
0.33	-12.3363	we | behavior
0.8125	-12.3363	parameter | controls
0.224771	-12.3364	of | fragments
0.164649	-12.3366	features | used
0.653846	-12.3369	mine | from
0.257485	-12.3371	is | otherwise
0.286765	-12.3373	more | better
0.286765	-12.3375	there | significant
0.160839	-12.3377	used | training
0.290076	-12.3377	with | information.
0.461538	-12.3379	a | specifying
0.212245	-12.3384	a | monolingual
0.8125	-12.3387	acknowledgements | authors
0.222222	-12.3389	features | their
0.344086	-12.339	that | explicitly
0.479167	-12.3392	generally | or
0.217021	-12.3394	the | posts
0.318182	-12.3394	in | 6,
0.10956	-12.3397	was | as
0.0595352	-12.3401	also | to
0.653846	-12.3402	learn | parameters
0.147059	-12.3404	therefore, | of
0.0903747	-12.3406	the | nodes
0.575758	-12.3408	pos tags | word
0.239583	-12.341	two | can be
0.438596	-12.3411	et | 2000)
0.302521	-12.3412	to | ones.
0.187302	-12.3414	variation | the
0.32381	-12.3415	the | outlined
0.296	-12.3417	4.4 | of
0.140071	-12.342	as | such as
0.126619	-12.3422	the | variables
0.653846	-12.3424	number | segments
0.220264	-12.3426	corpus | english
0.376623	-12.3427	have been | using
0.137931	-12.3429	is | many
0.857143	-12.3431	obama | president
0.488889	-12.3431	tuning | bleu
0.134206	-12.3433	structure | a
0.211382	-12.3435	factors | and
0.244565	-12.3436	efficiently | the
0.512195	-12.3438	documents | contain
0.120104	-12.3441	learned | the
0.365854	-12.3441	word | context.
0.857143	-12.3442	chart | cells
0.166247	-12.3444	from | text
0.857143	-12.3449	incoming | outgoing
0.16307	-12.3449	of | above
0.0706571	-12.3449	the | relation
0.361446	-12.345	in | treebank.
0.857143	-12.3454	j. | l.
0.348315	-12.3455	a | feature,
0.33	-12.3459	is | regular
0.0926791	-12.3467	knowledge | the
0.525	-12.347	differs | from
0.264151	-12.347	models | their
0.104024	-12.347	weight | the
0.135906	-12.3473	performance | with
0.0627451	-12.3475	- | -
0.285714	-12.3477	table | lists
0.525	-12.3478	models | built
0.190164	-12.3478	arabic | and
0.257485	-12.3478	sentences | same
0.160465	-12.3479	on | such as
0.388889	-12.348	this | considering
0.558824	-12.348	experiments | state-of-the-art
0.444444	-12.3483	with | increasing
0.479167	-12.3485	tries | a
0.239583	-12.3486	in | models,
0.923077	-12.3487	bigram | htb
0.45283	-12.3491	is | strict
0.388889	-12.3492	the | observations.
0.163415	-12.3495	the | link
0.857143	-12.3496	distributional | (harris,
0.0606061	-12.3497	is | which | the
0.298387	-12.3498	the | treebank,
0.0909774	-12.3498	process | the
0.123098	-12.3499	estimate | the
0.10382	-12.3499	total | of
0.857143	-12.3501	risk | loss
0.273973	-12.3501	of | tokens.
0.412698	-12.3501	an | solution
0.32381	-12.3502	relational | of
0.512195	-12.3502	english | arabic
0.205426	-12.3503	4.3 | of
0.8125	-12.3504	scotland, | 31,
0.525	-12.3505	2006; | 2009),
0.123269	-12.3505	annotators | the
0.469388	-12.3505	is | literal
0.438596	-12.3511	introduce | new
0.736842	-12.3512	title | title
0.348315	-12.3512	the | (eq.
0.714286	-12.3514	5the | is
0.8125	-12.3514	at | levels,
0.736842	-12.3516	gave | best
0.196491	-12.3516	the | domain,
0.282609	-12.3517	tuples | the
0.348315	-12.3517	words | aligned
0.208835	-12.3518	is | which | an
0.170667	-12.3519	performance | our
0.0932806	-12.352	for each | we | the
0.431034	-12.3522	improvements | over
0.208835	-12.3522	experiments | data
0.285714	-12.3523	define | set
0.200743	-12.3525	our | three
0.857143	-12.3527	terminal | non-terminal
0.213389	-12.3528	still | a
0.438596	-12.3531	implement | for
0.352273	-12.3535	bilingual | parallel
0.0965284	-12.3537	this | an | of
0.0863358	-12.3543	lexical | and
0.106076	-12.3543	is | any
0.423729	-12.3546	with | fewer
0.857143	-12.3547	ei | used
0.714286	-12.3547	sum | 1.
0.257669	-12.3547	and | training.
0.580645	-12.3547	complex | into
0.352273	-12.3547	the | interface.
0.197842	-12.3548	our goal | to
0.469388	-12.3548	chinese | segmentation
0.175141	-12.355	is | general
0.239362	-12.3551	the | vary
0.309735	-12.3551	those | were
0.388889	-12.3553	this paper | organized
0.361446	-12.3554	on | text,
0.228571	-12.3556	overall, | the
1	-12.3557	(finkel | 2005)
1	-12.3557	0.85 | 0.95
1	-12.3557	intransitive | transitive
1	-12.3557	(lsm | 23
1	-12.3557	(witten | frank,
1	-12.3557	rate | informality
1	-12.3557	wong | dras
0.162621	-12.3558	of | (i.e.,
0.27027	-12.3559	sentiment analysis | and
0.431034	-12.356	p, | and
0.916667	-12.356	hu | liu
0.320755	-12.3564	shown | performance
0.154348	-12.3564	developed | and
0.314815	-12.3566	from | initial
0.143667	-12.3567	(4) | the
0.206226	-12.3567	a | partial
0.469388	-12.3568	average | time
0.24581	-12.357	assumption | is
0.412698	-12.357	abstract | novel
0.540541	-12.3571	opinion | opinions
0.777778	-12.3571	we compared | against
0.6	-12.3575	simply | its
0.144213	-12.3581	research | in
0.323529	-12.3582	behind | that
0.916667	-12.3582	insertion | deletion
0.916667	-12.3582	ng | cardie
0.251462	-12.3583	of | categories.
0.122421	-12.3583	of | phrases
0.219298	-12.3584	with | high
0.25	-12.3585	is | expected
0.469388	-12.3586	9 | 9
0.208835	-12.3589	rather than | a
0.488889	-12.3594	bionlp | shared
0.208835	-12.3595	use | their
0.229268	-12.3596	short | and
0.46	-12.3596	obtain | results
0.431034	-12.3598	would | interesting
0.376623	-12.3598	it | reasonable
0.388889	-12.3601	captures | in
0.203846	-12.3604	use | same
0.736842	-12.3606	framenet | frame
0.187702	-12.3606	and | sets.
0.857143	-12.3607	aspect | consumer
0.326733	-12.3609	to | question.
0.307018	-12.361	is | hence
0.736842	-12.361	percentages | emotion
0.8125	-12.361	si | si
0.239362	-12.3612	of | classification.
0.916667	-12.3613	reordering | reordering.
0.247191	-12.3614	merged | the
0.478261	-12.3614	university | texas
0.199262	-12.3614	our | algorithm
0.127628	-12.3614	have | or
0.46	-12.3616	place | on
0.125541	-12.3619	line | the
0.209677	-12.362	new | which
0.412698	-12.362	has | (e.g.
0.208835	-12.3621	included | and
0.172131	-12.3622	we also | with
0.333333	-12.3622	features | n-gram
0.371795	-12.3623	then, | are
0.916667	-12.3624	within | framework,
0.371795	-12.3625	highlight | of
0.558824	-12.3626	creates | for
0.15	-12.3627	it | that | not
0.397059	-12.3629	an | known
0.777778	-12.363	in | introduction,
0.218341	-12.3631	approach, | a
0.215517	-12.3631	is | containing
0.211618	-12.3633	strong | of
0.777778	-12.3633	choosing | best
0.199262	-12.3635	aligned | in
0.226415	-12.3637	the | pcfg
0.311927	-12.3638	uses | features
0.431034	-12.3639	the | set).
0.777778	-12.364	it | evident
0.469388	-12.3645	the | chunks.
0.323529	-12.3647	that | yield
0.234694	-12.3648	a | chosen
0.540541	-12.3648	score | above
0.136054	-12.3648	output | a
0.46	-12.3648	step | at
0.857143	-12.3649	summaries | pyramid
0.244444	-12.365	can | improve
0.25	-12.365	methods | 27–31,
0.254545	-12.3652	the | template
0.247191	-12.3652	the | expressions.
0.344444	-12.3652	by | contract
0.231527	-12.3653	performance | performance
0.276596	-12.3656	in addition, | that
0.213675	-12.3658	and | (2008)
0.478261	-12.3659	compute | scores
0.243094	-12.3659	(2006) | and
0.444444	-12.366	each | exactly
0.326733	-12.3661	release | the
0.222222	-12.3662	actual | of
0.431034	-12.3664	model | likelihood
0.336842	-12.3666	the | directions.
0.19244	-12.3667	and | pairs.
0.123077	-12.3667	the | whereas
0.314815	-12.3668	a | rule-based
0.62963	-12.3671	0.25 | 0.35
0.478261	-12.3674	for | comparison.
0.357143	-12.3679	and | directions
0.136442	-12.368	sum | the
0.159624	-12.368	manual | of
0.122734	-12.3683	a | structure
0.239362	-12.3686	that | produce
0.607143	-12.3686	because | too
0.916667	-12.3686	raises | question
0.857143	-12.3689	distortion | reordering
0.478261	-12.3689	we | ran
0.357143	-12.369	of | utility
0.540541	-12.3691	this | essentially
0.151709	-12.3691	the | links
0.314815	-12.3692	the | response.
0.279412	-12.3697	multilingual | and
0.336842	-12.3697	in | format
0.112019	-12.3698	3.1 | the
0.108791	-12.3701	make | of
0.193103	-12.3703	component | a
0.182663	-12.3705	the | clearly
0.101825	-12.3709	the | have been
0.281481	-12.371	the | before | after
0.145669	-12.3712	idea | to | the
0.0776536	-12.3717	value | the
0.8125	-12.3717	amounts | monolingual
0.367089	-12.3719	and | meaning.
0.323529	-12.3721	examples of | are
0.478261	-12.3721	recognizing | in
0.184127	-12.3721	the | instances.
0.317757	-12.3723	by | resulting
0.40625	-12.3723	best | combination
0.184127	-12.3724	of | via
0.27027	-12.3725	finding | for
0.155756	-12.3726	method | by
0.681818	-12.3726	15th | 23-24
0.279412	-12.3728	definitions | and
0.916667	-12.3728	achieve | goal,
0.274648	-12.3729	two | sets
0.6	-12.3729	system, | system.
0.714286	-12.373	because | give
0.357143	-12.3731	in | international conference
0.141527	-12.3733	2005; | al.,
0.326733	-12.3733	of | training data,
0.348315	-12.3734	only | words.
0.172222	-12.3735	probability | the | probability
0.0816832	-12.3735	4 | and
0.159145	-12.3735	all | have
0.488889	-12.3738	using | rules.
0.542857	-12.374	relating | to
0.45098	-12.374	a | token.
0.416667	-12.3745	the | observed.
0.3625	-12.3745	tasks, | for
0.681818	-12.3747	pairs, | pairs.
0.558824	-12.3752	combining | two
0.3625	-12.3752	da | da
0.222222	-12.3753	are | without
0.127273	-12.3754	the | expected
0.4	-12.3756	to | wikipedia.
0.916667	-12.3756	center | computational
0.311927	-12.3759	and | rest
0.284615	-12.3762	consists of | with
0.216814	-12.3763	simply | in
0.168449	-12.3766	are | so
0.320388	-12.3768	is | hard
0.205534	-12.3768	workshop on | and
0.192982	-12.3769	test | training
0.397059	-12.3773	conll | shared
0.301724	-12.3774	mappings | and
0.344444	-12.3776	used | extract
0.412698	-12.3776	characteristic | of
0.238095	-12.3779	composed | and
0.156322	-12.3781	implemented | the
0.468085	-12.3782	abstract | describe
0.188119	-12.3784	the | distribution.
0.436364	-12.3787	is | difficulty
0.8125	-12.3787	people | express
0.309091	-12.3789	identifying | that
0.3625	-12.3793	model | hierarchical
0.423729	-12.3795	φ | φ
0.204724	-12.3797	the | nodes.
0.289062	-12.3797	for | remaining
0.512821	-12.3798	a | perceived
0.110604	-12.3799	these | by
0.336957	-12.38	outperforms | by
0.289062	-12.3802	sample | from
0.488372	-12.3806	we present | evaluation
0.736842	-12.3807	disadvantage | that
0.580645	-12.3807	we | (joachims,
0.62963	-12.3807	while | quite
0.526316	-12.3808	our | published
0.478261	-12.3808	learning | parameters
0.209877	-12.3808	is | complex
0.284615	-12.381	plot | the
0.142586	-12.3812	contribution | the
0.329897	-12.3813	there | still
0.681818	-12.3815	learnability | language
0.301724	-12.3815	the | prediction.
0.5	-12.3815	each | remaining
0.916667	-12.3816	association | colour.
0.145709	-12.3818	two | are | in
0.0963542	-12.3823	thus, | the
0.416667	-12.3824	relative | over
0.3625	-12.3825	an | would be
0.8125	-12.3828	subject | constraints
0.416667	-12.3829	to | with.
0.352941	-12.3836	the | small.
0.117117	-12.3836	the | pairs.
0.736842	-12.3837	0.01 | 0.05
0.109131	-12.3837	only | are
0.8125	-12.3838	prepositions | preposition
0.236842	-12.3839	first | into
0.259494	-12.3839	negative | negative
0.40625	-12.384	5.2 | our
0.143689	-12.384	current | and
0.309091	-12.3841	an edge | the
0.428571	-12.3843	independent | each
0.235602	-12.3843	et | 2003)
0.4	-12.3848	query | search
0.228856	-12.3849	sentences | words
0.916667	-12.3853	portion | gigaword
0.558824	-12.3853	roughly | from
0.135652	-12.3854	third | the
0.416667	-12.3854	the | from.
0.234375	-12.3855	formulation | the
0.267123	-12.3855	each | class
0.161369	-12.3859	from | data.
0.115578	-12.3861	model | which
0.254658	-12.3861	other | were
0.148148	-12.3862	if | can
0.358025	-12.3863	the | standard,
0.304348	-12.3863	a | variant
0.191638	-12.3863	the | assignment
0.121212	-12.3864	we can | that
0.857143	-12.3866	by | ?
0.323232	-12.3866	create | by
0.195652	-12.3867	to | left
0.340659	-12.3868	many | were
0.329897	-12.3868	in | subsequent
0.916667	-12.3868	stochastic | grammars
0.478261	-12.3869	were | manual
0.15942	-12.3872	the | constraint
0.176471	-12.3874	the | index
0.736842	-12.3876	apply | techniques
0.176471	-12.3876	generated by | the
0.136767	-12.3877	right | the
0.277372	-12.3877	4) | the
0.478261	-12.3878	grammar | rule
0.180428	-12.3878	furthermore, | to
0.40625	-12.388	choose | from
0.127106	-12.3889	to | words.
0.8125	-12.3891	social | or,
0.29661	-12.3894	the | distributions.
0.416667	-12.3895	method | uses
0.111241	-12.3896	a | vector
0.168449	-12.3896	showed | of
0.333333	-12.3897	length | each
0.314286	-12.3898	produce | with
0.0760394	-12.3898	to be | of
0.22381	-12.3898	each | vector
0.428571	-12.3898	classifiers | features
0.380282	-12.3899	and | 2010b).
0.336957	-12.3901	a | algorithm,
0.558824	-12.3907	reduces | number
0.326531	-12.3911	of | respectively
0.580645	-12.3912	singular | decomposition
0.320388	-12.3912	a | b,
0.8125	-12.3912	expect | see
0.0994371	-12.3914	that | has
0.151515	-12.3915	system | which
0.64	-12.3917	weighted | weight
0.148536	-12.3919	derivation | the
0.131579	-12.3919	current | of
0.106109	-12.392	6 | and
0.442308	-12.3921	is | member
0.123547	-12.3922	figure 1 | the
0.083388	-12.3924	the | domain
0.124814	-12.3924	in | second
0.352941	-12.3925	sentences | documents
0.263158	-12.3927	conjunction | the
0.0666667	-12.3931	context | the
0.282443	-12.3931	training data | from
0.478261	-12.3931	due to | large
0.157277	-12.3933	link | the
0.764706	-12.3936	most | systems,
0.138434	-12.3943	content | and
0.314286	-12.3945	the | in,
0.290323	-12.3945	because | has
0.428571	-12.3945	for | methods,
0.23913	-12.3945	so | it
0.348837	-12.3945	the | target,
0.393939	-12.3946	while | et al.
0.109714	-12.3946	the | joint
0.45098	-12.3948	by | majority
0.267123	-12.3949	when | into
0.353659	-12.3951	assign | each
0.258065	-12.3953	for | sentences,
0.526316	-12.3955	output | systems
0.277372	-12.3956	to extend | to
0.151648	-12.3958	the | lists
0.344828	-12.3958	baseline | performance
0.40625	-12.3959	when | small
0.130016	-12.396	the | line
0.294118	-12.3961	speaker | to
0.137011	-12.3961	one | which
0.607143	-12.3963	three | 1)
0.607143	-12.3963	all | into | single
0.301724	-12.3963	indices | the
0.235294	-12.3964	(a | of
0.137931	-12.3965	and | values
0.428571	-12.3972	encoding | for
0.393939	-12.3973	systems | same
0.542857	-12.3975	our | framework,
0.190972	-12.3975	portland, | 19-24,
0.0557029	-12.3976	we | each
0.314286	-12.3977	learned | from
0.0815424	-12.3977	the | after
0.380282	-12.398	from | perspective,
0.373333	-12.3982	the | message.
0.527778	-12.3983	dependency | arcs
0.253086	-12.3986	links | to
0.323232	-12.3989	in general, | a
0.607143	-12.399	a | criterion.
0.280303	-12.3991	the | 1999).
0.171831	-12.3991	our approach | in
0.131579	-12.3991	the | includes
0.288	-12.3994	it | good
0.125186	-12.3995	path | the
0.261438	-12.3995	the | grammar,
0.306306	-12.3995	a | google
0.857143	-12.3996	solving | problems
0.607143	-12.3996	there | substantial
0.736842	-12.3997	bad | bad
0.203187	-12.3999	a | potential
0.857143	-12.4001	noun | noun.
0.190972	-12.4001	2002; | al.,
0.213043	-12.4003	the | size.
0.468085	-12.4005	wordnet | sense
0.115044	-12.4006	contains | of
0.124814	-12.4007	they | not
0.681818	-12.401	representation | representation.
0.303571	-12.401	more | higher
0.271429	-12.4013	σ | of
0.236559	-12.4016	and | tools
0.681818	-12.4016	briefly | some
0.421053	-12.4016	improvements | bleu
0.254658	-12.4017	know | and
0.336957	-12.4019	also | 2010).
0.421053	-12.402	of | complexity.
0.336957	-12.402	further | into
0.269504	-12.4025	we extract | and
0.285714	-12.4028	find | which
0.150655	-12.4029	have been | al.,
0.45098	-12.4029	a | matrix.
0.311321	-12.403	computed | each
0.380282	-12.403	we | separately
0.458333	-12.4031	dictionary | used
0.764706	-12.4036	because | generally
0.363636	-12.4036	could | could
0.607143	-12.4038	half | all
0.512821	-12.4038	search | possible
0.458333	-12.4039	(pyysalo | al.,
0.0957821	-12.404	parameters | the
0.160494	-12.404	considered | and
0.130719	-12.4041	various | of
0.736842	-12.4044	belief | belief
0.477273	-12.4046	some | certain
0.586207	-12.4046	alternatively, | can
0.681818	-12.4046	and | (1991)
0.428571	-12.4047	that | equally
0.190311	-12.4048	purpose | the
0.202381	-12.4048	our | about
0.237838	-12.405	because | have
0.607143	-12.4052	represent | meaning
0.764706	-12.4055	information | communications
0.542857	-12.4055	does not | so
0.271429	-12.4057	not | case
0.084478	-12.4058	two | are
0.4	-12.4058	is | perfect
0.32	-12.4058	the | choice.
0.147368	-12.4059	thus, | to
0.393939	-12.4061	combination | two
0.7	-12.4063	taxonomy | ontology
0.615385	-12.4063	words | contexts.
0.280303	-12.4066	another | which
0.2103	-12.4068	and | forms
0.138632	-12.4069	the | scheme
0.468085	-12.4071	total | per
0.112183	-12.4072	when | with
0.12	-12.4073	annotation | a
0.201581	-12.4073	of | components
0.349398	-12.4074	conversation | a
0.102851	-12.4077	example | and
0.25641	-12.4077	some | their
0.363636	-12.4077	since | very
0.098777	-12.4078	the | word.
0.2	-12.4078	overall | is
0.527778	-12.4079	on | only.
0.857143	-12.408	averaged | algorithm
0.192171	-12.408	therefore, | in
0.25641	-12.408	the | forms.
0.237569	-12.408	(2006) | a
0.216216	-12.4082	a | action
0.135957	-12.4083	of | considered
0.0773672	-12.4083	the | rule
0.229592	-12.4084	when | was
0.145194	-12.4086	for | annotation
0.512821	-12.4088	add | new
0.380282	-12.4088	used | learn
0.385714	-12.4091	we | removing
0.25641	-12.4091	the | speech.
0.5	-12.4091	these | simply
0.512821	-12.4092	used | annotate
0.2	-12.4092	proposed | et al.,
0.271429	-12.4093	the | software
0.141199	-12.4093	is | standard
0.349398	-12.4094	metrics | human
0.764706	-12.4097	it | feasible
0.245614	-12.4098	chunks | the
0.176119	-12.41	we | perform
0.212121	-12.41	the | voice
0.7	-12.4101	factorization | are
0.135957	-12.4101	the | gives
0.329787	-12.4103	is | directed
0.211207	-12.4103	that | effective
0.242775	-12.4103	and | parsing.
0.488372	-12.4104	unigrams, | and
0.254777	-12.4106	is | limited
0.652174	-12.4107	after | iterations.
0.5625	-12.4109	a | syntax.
0.527778	-12.4109	context | context.
0.144289	-12.4109	this | some
0.916667	-12.4111	|v|
0.184211	-12.4113	last | in
0.421053	-12.4114	techniques, | to
0.468085	-12.4116	textual | which
0.652174	-12.4116	system | fully
0.916667	-12.4117	uniformly | random
0.542857	-12.4119	and | discussion.
0.163636	-12.412	some | on
0.615385	-12.4122	train | classifier.
0.192727	-12.4123	the | entities.
0.527778	-12.4124	given | positive | negative
0.421053	-12.4125	because | are not
0.45098	-12.4127	that | conventional
0.358974	-12.4128	relations | other
0.291667	-12.4128	method | and | method
0.64	-12.4129	building | models
0.247059	-12.4129	from | extracted
0.280303	-12.4132	in | connected
0.182109	-12.4133	requires | to
0.179245	-12.4133	results | performance
0.222222	-12.4134	resulting | is
0.326316	-12.4135	separately | and
0.165333	-12.4139	the | 2002).
0.468085	-12.414	and | over | baseline.
0.0523948	-12.4142	sentences | the
0.184818	-12.4151	for | model,
0.764706	-12.4153	images | their
0.2	-12.4153	a | consisting
0.294118	-12.4154	annotation | was
0.344828	-12.4154	some | most
0.8	-12.4154	elements | elements.
0.242775	-12.4155	intersection | the
0.615385	-12.4156	(papineni | 2002).
0.157385	-12.4157	the | translations.
0.586207	-12.4157	natural | china
0.271429	-12.4158	we | choose
0.652174	-12.416	but | longer
0.265734	-12.4161	correspond to | of
0.607143	-12.4161	using | biomedical
0.442308	-12.4162	that | exhibit
0.458333	-12.4166	thus, | only
0.298246	-12.4166	continue | the
0.586207	-12.417	both | languages.
0.126168	-12.417	the | utterance
0.0811502	-12.4177	and | both
0.152273	-12.4181	these | our
0.167123	-12.4181	steps | the
0.195489	-12.4182	i.e., | and
0.271429	-12.4182	the | tag.
0.234973	-12.4182	are | through
0.174041	-12.4183	a | latent
0.615385	-12.4183	represents | value
0.7	-12.4184	have | meanings.
0.736842	-12.4185	collecting | data
0.311321	-12.4185	the | name.
0.230366	-12.4188	the | referent
0.0569361	-12.4189	an | for
0.232804	-12.419	in | span
0.0925771	-12.4191	the | of | corresponding
0.369863	-12.4191	2005; | 2009).
0.5625	-12.4192	how | help
0.38806	-12.4192	given | w
0.512821	-12.4194	of | constructions,
0.442308	-12.4196	potential | which
0.542857	-12.4197	but | them.
0.295652	-12.4198	enables | of
0.308411	-12.4198	contains | words
0.652174	-12.4198	part | larger
0.132302	-12.4199	the | estimated
0.208511	-12.4199	long | of
0.308411	-12.4199	and | patterns.
0.313725	-12.4204	of | categories,
0.0204181	-12.4205	for | and
0.116866	-12.4205	and | 2006;
0.26	-12.4206	this | every
0.64	-12.4207	exploit | information
0.24	-12.4207	the | database.
0.615385	-12.4209	1, | 4,
0.329787	-12.4211	using | bilingual
0.44898	-12.4214	up | length
0.25	-12.4214	our | against
0.0842541	-12.4217	computed | the
0.380282	-12.422	and | 2010.
0.393939	-12.4223	for each | average
0.269504	-12.4223	3.3 | we
0.64	-12.4224	using | link
0.163212	-12.4224	in | process
0.128664	-12.4224	the | counts
0.512821	-12.4225	summarization | different
0.413793	-12.4226	the | axis
0.132759	-12.4226	that | should
0.8	-12.4226	ambiguity | resolved
0.215596	-12.4226	of | matches
0.166667	-12.4228	issue | the
0.349398	-12.4228	problem | learning
0.857143	-12.4229	leacock | chodorow
0.857143	-12.4229	(teufel | moens,
0.413793	-12.4229	among | most
0.214612	-12.4229	increases | of
0.358974	-12.4231	the | α.
0.5	-12.4234	aligning | in
0.102775	-12.4235	the | position
0.681818	-12.4241	two | baselines.
0.220096	-12.4243	a | alternative
0.234973	-12.4248	duration | the
0.114396	-12.4248	first | that
0.369863	-12.4248	the | converge
0.513514	-12.4249	identified | their
0.615385	-12.4249	implementation | algorithm
0.32	-12.4249	performance | method
0.5	-12.4252	perform | analysis
0.263889	-12.4253	notation | the
0.149123	-12.4258	and | models.
0.409836	-12.4258	why | not
0.764706	-12.4259	definite | indefinite
0.0726804	-12.4259	of | text
0.193431	-12.4259	document, | the
0.337079	-12.4259	simplicity, | the
0.230366	-12.4262	letters | the
0.8	-12.4262	within | scope
0.274074	-12.4263	words | those
0.270073	-12.4263	of | article.
0.227979	-12.4263	performance | all
0.166667	-12.4264	translation | using
0.130217	-12.4266	the | span
0.204082	-12.4266	the | scores,
0.5625	-12.4267	knowing | is
0.225	-12.4268	of | speaker
0.382353	-12.427	opinion | polarity
0.5	-12.427	an | gain
0.305556	-12.4273	abstract | task
0.128247	-12.4274	tasks | and
0.8	-12.4275	tune | feature
0.433962	-12.4276	agreement | annotators
0.586207	-12.4278	length | frequency
0.308411	-12.4278	on top | the
0.477273	-12.4282	comparison | systems
0.375	-12.4282	such | might
0.291667	-12.4283	the | dialogue.
0.340909	-12.4284	to | this problem,
0.615385	-12.4285	so | must
0.222772	-12.4286	group | in
0.316832	-12.4287	size | corpus
0.586207	-12.4292	and | forth.
0.286885	-12.4295	option | the
0.0687413	-12.4296	is | all
0.326316	-12.4296	define | which
0.298246	-12.4297	graph | between
0.369863	-12.4298	we will use | to
0.425926	-12.4299	no | available
0.345238	-12.4299	annotators | each
0.8	-12.4302	in | (ed.),
0.293103	-12.4304	of | errors,
0.229167	-12.4309	accurate | of
0.254777	-12.431	the | rows
0.466667	-12.4315	a | matrix,
0.180064	-12.4316	studies | of
0.7	-12.4318	which | past
0.527778	-12.4318	based on | current
0.375	-12.4318	us | us
0.237288	-12.4319	university | usa
0.144628	-12.4319	the | gain
0.111931	-12.4322	the | (i.e.,
0.223881	-12.4323	• | –
0.073235	-12.4324	in | sentence
0.233696	-12.4324	the | incoming
0.268116	-12.4325	nodes, | the
0.466667	-12.4326	distributional | meaning
0.5625	-12.4327	conditional | field
0.289256	-12.4327	average | over
0.916667	-12.4328	michigan | arbor,
0.8	-12.4328	analysis | underlying
0.5625	-12.4329	and | n-grams.
0.175227	-12.433	is | measure
0.3	-12.4334	the | company
0.458333	-12.4336	(2000) | a
0.35443	-12.4336	the | directly.
0.375	-12.4339	features | computed
0.369863	-12.434	(as | by
0.566667	-12.4342	used | express
0.413793	-12.4342	number | required
0.11855	-12.4344	simple | and
0.545455	-12.4344	obtain | best
0.337079	-12.4346	since | many
0.916667	-12.4349	outperforms | baselines.
0.21267	-12.4349	models | features
0.298246	-12.4349	of | entities,
0.433962	-12.4353	another | same
0.333333	-12.4353	the | definition.
0.28125	-12.4355	remove | and
0.333333	-12.4356	using | documents
0.178914	-12.4359	method | which
0.326316	-12.4363	revision | the
0.396825	-12.4366	dramatically | the
0.846154	-12.4367	agent | patient
0.615385	-12.4368	senses | wordnet
0.125589	-12.4368	model | using
0.652174	-12.4368	machine | techniques.
0.44898	-12.4372	software | for
0.477273	-12.4373	1 introduction | research
0.345238	-12.4374	and | p.
0.5625	-12.4375	some | still
0.35443	-12.4375	mentions | same
0.134409	-12.4376	of | various
0.153132	-12.4377	a | word.
0.235955	-12.4377	for | system,
0.262069	-12.4377	mainly | and
0.7	-12.4378	srl | argument
0.0879878	-12.4379	for | sentence
0.376812	-12.4379	these | through
0.345238	-12.4382	model | learn
0.290598	-12.4383	besides | of
0.214612	-12.4383	properties | a
0.191882	-12.4385	samples | the
0.158025	-12.4385	thus, | a
0.305556	-12.439	an event | a
0.290598	-12.4391	illustrates | of
0.319588	-12.4391	potential | that
0.764706	-12.4392	predicate | predicate.
0.40678	-12.4392	note that | because
0.764706	-12.4394	reach | performance
0.421053	-12.4395	that | others.
0.110706	-12.4398	3.2 | the
0.358974	-12.4398	and | feedback.
0.5	-12.4401	about | topic
0.764706	-12.4402	into | subsets
0.316327	-12.4402	bilingual | pairs
0.307692	-12.4404	successful | of
0.40678	-12.4404	weights | features
0.272059	-12.4405	final | by
0.268116	-12.4409	that, | are
0.586207	-12.4411	university, | china
0.545455	-12.4411	shown | tables
0.7	-12.4412	the | decade
0.144958	-12.4413	of | annotations
0.722222	-12.4416	nonterminal | rules
0.32967	-12.4417	utterances | dialogue
0.168571	-12.4421	that | might
0.722222	-12.4421	when | available,
0.232432	-12.4422	native | of
0.221675	-12.4422	the | similarity.
0.527778	-12.4423	of | simplification.
0.246914	-12.4425	middle | the
0.916667	-12.4427	i/j | mi/j
0.916667	-12.4427	cb | ip
0.916667	-12.4427	braceleftbigg | bracerightbigg
0.218009	-12.4427	12 | of
0.260274	-12.4427	in | 2008),
0.722222	-12.4428	find | solution
0.27907	-12.4432	that | outperform
0.297297	-12.4432	configuration | and
0.246914	-12.4434	the | nearest
0.116531	-12.4435	sentence, | the
0.235955	-12.4435	a | not.
0.143443	-12.4435	then | by
0.390625	-12.4436	and | classifiers.
0.295652	-12.4437	that | independent
0.238636	-12.444	a | scale
0.592593	-12.4442	words | polarity.
0.527778	-12.4443	approaches | problem
0.145263	-12.4443	initial | and
0.173134	-12.4444	of | (in
0.846154	-12.4447	seventh | under
0.228723	-12.4447	propositions | the
0.179739	-12.4449	can | such as
0.108235	-12.445	in | those
0.0880989	-12.4456	due to | the
0.106576	-12.4456	to | them
0.652174	-12.4456	reducing | number
0.358974	-12.4459	however, | much
0.207792	-12.4459	not | since
0.219512	-12.4459	a | change
0.652174	-12.446	often | multiple
0.156098	-12.4463	we used | for
0.0848921	-12.4463	selected | the
0.253247	-12.4463	a | relationship
0.16954	-12.4464	half | the
0.592593	-12.4464	to predict | given
0.151724	-12.4466	requires | of
0.652174	-12.4467	world | 23
0.764706	-12.447	documents | topics.
0.376812	-12.447	described | previous
0.456522	-12.447	not | reliable
0.527778	-12.4471	guaranteed | be
0.586207	-12.4471	linguistic | social
0.189781	-12.4471	the | generates
0.364865	-12.4472	may | multiple
0.32967	-12.4474	the | 10.
0.418182	-12.4475	integrates | and
0.276923	-12.4476	directly | as
0.268116	-12.4477	occurred | the
0.297297	-12.4478	the | position,
0.376812	-12.4481	however, | time
0.266187	-12.4483	mentions | are
0.364865	-12.4484	conversion | to
0.144351	-12.4485	the | entailment
0.592593	-12.4485	score | scoring
0.456522	-12.4487	large | corpus.
0.180328	-12.449	collection | and
0.260274	-12.449	hypotheses | a
0.418182	-12.4493	build | system
0.304762	-12.4495	compare | performance
0.322917	-12.4496	the | weights,
0.341176	-12.4497	et al., | 2003).
0.319588	-12.4498	in | iteration
0.18662	-12.4499	> | >
0.326087	-12.45	in fact, | a
0.221675	-12.4504	in | third
0.413793	-12.4506	system | moses
0.456522	-12.4508	five | were
0.19685	-12.4509	and | 2008),
1	-12.451	ith | ith
1	-12.451	low-level | high-level
1	-12.451	brody | elhadad
1	-12.451	@ | (me@hal3.name)
1	-12.451	chung | gildea
1	-12.451	sp2irl | iii
1	-12.451	semantics | 23,
1	-12.451	(crammer | singer,
1	-12.451	mohammad | turney
1	-12.451	homologene | ensembl
0.652174	-12.451	distribution | mean
0.319588	-12.4511	is | dirichlet
0.254902	-12.4513	acknowledgments | thank
0.425926	-12.4513	the | nps.
0.909091	-12.4513	character-level | word-level
0.909091	-12.4513	at | glance,
0.909091	-12.4513	prefixes | suffixes,
0.652174	-12.4514	require | knowledge
0.652174	-12.4514	anaphor | a
0.7	-12.4515	virtual | world
0.115127	-12.4515	table 2 | the
0.0987292	-12.4518	the | had
0.364865	-12.4519	verb | subject
0.369863	-12.452	acknowledgments | project
0.241176	-12.452	of | relations,
0.666667	-12.4522	clauses | clauses
0.7	-12.4524	one | completely
0.586207	-12.4526	associating | to
0.8	-12.4526	p(w|t)
0.625	-12.4527	accuracy | parsers
0.909091	-12.4529	r. | k.
0.456522	-12.4529	they | provide
0.413793	-12.453	results | summarized
0.166205	-12.453	2010; | al.,
0.246914	-12.4532	since | use
0.238372	-12.4532	creates | the
0.666667	-12.4535	some | might be
0.153302	-12.4538	annotations | and
0.326087	-12.4539	a | topic,
0.652174	-12.454	in | situation,
0.337209	-12.4541	the | 1994).
0.202479	-12.4542	the | category.
0.10702	-12.4543	table 1 | the
0.36	-12.4545	each | sentence,
0.44898	-12.4545	4 | generated
0.151163	-12.4545	to | relevant
0.8	-12.4545	on | sides.
0.258503	-12.4547	reports | of
0.235955	-12.4549	previous work | to
0.208889	-12.455	however, | such
0.545455	-12.455	only | entire
0.0298108	-12.4551	for | for
0.260274	-12.4551	will | if
0.7	-12.4552	weights | tuned
0.288136	-12.4554	the | zero.
0.382353	-12.4555	1996; | al.,
0.168091	-12.4555	to | certain
0.197628	-12.4556	to | entire
0.122542	-12.4558	describe | of
0.254902	-12.4559	components | a
0.390625	-12.4561	of | fluent
0.625	-12.4565	(radev | al.,
0.615385	-12.4566	using | search.
0.241176	-12.4568	representative | the
0.764706	-12.457	host | host
0.221106	-12.457	generally | and
0.846154	-12.4572	institute | infocomm
0.119253	-12.4572	this | also
0.846154	-12.4574	locality | hashing
0.615385	-12.4578	each | x,
0.5	-12.458	srilm | (stolcke,
0.396825	-12.4581	for instance, | can
0.456522	-12.4581	many | types
0.233333	-12.4582	most | have
0.5	-12.4582	bitext | and
0.7	-12.4583	supervised | semi-supervised
0.238372	-12.4585	the | substantially
0.258503	-12.4585	similar | used
0.35	-12.4586	which | result
0.195312	-12.4587	on | task.
0.226316	-12.4588	provide | with
0.319588	-12.4596	expresses | the
0.615385	-12.4599	tuning | tuning
0.846154	-12.46	candidate generation | candidate
0.652174	-12.4602	because | already
0.307692	-12.4604	that | unique
0.31	-12.4604	requirement | the
0.355263	-12.4604	there are | different
0.205128	-12.4604	in | whereas
0.105145	-12.4605	the | before
0.140562	-12.4606	of | articles
0.666667	-12.4607	(2006) | semantic
0.168605	-12.4607	every | is
0.211009	-12.4609	the | track
0.529412	-12.4611	involving | or
0.513514	-12.4614	trained | wsj
0.722222	-12.4614	words | images
0.592593	-12.4614	we | characters.
0.355263	-12.4618	the | infix
0.127907	-12.462	the | (in
0.22449	-12.462	are | (e.g.
0.0983284	-12.4623	of | terms
0.14966	-12.4624	of | observed
0.107351	-12.4624	three | in
0.28	-12.4625	we explore | of
0.301887	-12.4625	values | all
0.0995976	-12.4625	only | that
0.722222	-12.4626	alignment | mt
0.566667	-12.4629	into | groups
0.326087	-12.463	the | parses.
0.14693	-12.4633	using | training
0.162602	-12.4634	greater | the
0.0728548	-12.4634	the | class
0.256757	-12.4635	secondly, | the
0.154412	-12.4637	increase | of
0.345679	-12.4637	the | annotated.
0.545455	-12.4638	this | valuable
0.466667	-12.4639	average | across
0.846154	-12.464	spearman | correlation
0.384615	-12.4643	is | and/or
0.592593	-12.4644	candidate | sentence.
0.180602	-12.4645	and | techniques
0.142562	-12.4645	an | where
0.192453	-12.4647	about | or
0.5	-12.4648	techniques | models
0.301887	-12.4652	of | alignments.
0.566667	-12.4653	these, | are
0.234637	-12.4653	both | have
0.625	-12.4655	ranking | candidates
0.337209	-12.4656	but | less
0.0676937	-12.4657	their | and
0.204255	-12.4657	the | projected
0.615385	-12.4657	supervised | based on
0.410714	-12.4659	unlabeled | unlabeled
0.40678	-12.4659	shen | al.,
0.625	-12.4659	using | background
0.218447	-12.4659	acknowledgments | research
0.156566	-12.466	framework | a
0.548387	-12.4661	once | into
0.313131	-12.4662	of | questions.
0.513514	-12.4664	other | tasks.
0.128814	-12.4665	have | which
0.529412	-12.4666	is | multinomial distribution
0.212963	-12.4671	definitions | the
0.0790312	-12.4672	measure | the
0.425926	-12.4673	entities | their
0.35	-12.4673	for | domain,
0.277778	-12.4674	data | text
0.285714	-12.4675	speech | language
0.846154	-12.4677	restriction | type
0.103373	-12.4678	the | made
0.371429	-12.4679	increase | number
0.179402	-12.468	only | words
0.11502	-12.468	here | the | of
0.129534	-12.4681	the | likelihood
0.326087	-12.4682	probabilities, | the
0.8	-12.4682	in | past,
0.592593	-12.4682	for | term.
0.200837	-12.4683	despite | of
0.132969	-12.4684	first | from
0.337209	-12.4685	possible | use
0.236994	-12.4687	for | recognition
0.566667	-12.4688	turned | in
0.36	-12.4691	2we | the
0.333333	-12.4691	our | scheme
0.0998981	-12.4692	the | containing
0.299065	-12.4693	translation | bilingual
0.194553	-12.4695	an additional | the
0.4	-12.4696	this | clearly
0.127946	-12.4696	the | weighted
0.235632	-12.4697	the | variables.
0.316327	-12.4698	the | rooted
0.245283	-12.4701	to | domains.
0.195312	-12.4702	limited | and
0.113757	-12.4702	is | so
0.196787	-12.4703	major | of
0.316327	-12.4704	the | span.
0.294643	-12.4706	are | before
0.431373	-12.4707	finally, | models
0.163435	-12.471	and | arguments
0.211982	-12.4712	the | question.
0.200837	-12.4713	coverage | and
0.246835	-12.4714	this | similar to
0.8	-12.4715	representation | representations
0.200837	-12.4717	of | n-grams
0.44	-12.4717	devices | and
0.190299	-12.4721	a | path
0.236994	-12.4722	and | denote
0.5	-12.4724	and | respectively)
0.909091	-12.4725	node | children.
0.722222	-12.4726	attaching | to
0.513514	-12.4727	(we | for
0.197581	-12.4727	probabilistic | a
0.909091	-12.4728	matrix | rows
0.235632	-12.4729	as | reference
0.123824	-12.4729	of | (e.g.
0.217822	-12.473	using | learning
0.168116	-12.473	given | an
0.182131	-12.473	et | 2007),
0.222222	-12.4732	we | separate
0.410714	-12.4733	in | 2011),
0.390625	-12.4737	acknowledgments this | funded
0.529412	-12.474	a | product.
0.592593	-12.4747	phrases | contain
0.333333	-12.4748	the | bleu.
0.130282	-12.4749	thus | to
0.666667	-12.4751	compute | probabilities
0.456522	-12.4751	from | parsed
0.722222	-12.4751	sequence | words,
0.337349	-12.4751	conclusions | presented
0.418182	-12.4751	semantics | semantic
0.31	-12.4754	systematic | of
0.10148	-12.4757	the | weights
0.133455	-12.4758	the | quite
0.20614	-12.476	the | parameters.
0.277778	-12.4762	with | annotations
0.666667	-12.4763	output | output.
0.548387	-12.4765	tense | verb
0.5	-12.4768	of | synonyms.
0.378788	-12.4769	examining | of
0.47619	-12.4769	to | decision.
0.592593	-12.4769	selecting | best
0.20524	-12.4771	in | documents.
0.0885668	-12.4771	frequency | the
0.393443	-12.4772	∑ | where
0.200837	-12.4773	we | investigate
0.31	-12.4773	for | classifier.
0.722222	-12.4773	semi-supervised | semi-supervised
0.666667	-12.4774	but | little
0.2	-12.4775	assumption | a
0.592593	-12.4775	a | pattern,
0.456522	-12.4776	models | modeling
0.286957	-12.4776	sets | set
0.296296	-12.4777	that | expressed
0.101713	-12.4777	the | translated
0.4	-12.4778	wikipedia, | and
0.285714	-12.4781	typical | a
0.548387	-12.4782	example | like
0.431373	-12.4782	employs | to
0.384615	-12.4783	parsing | parse
0.268657	-12.4783	day | the
0.240964	-12.4784	dataset | from
0.456522	-12.4784	side | parallel
0.355263	-12.4785	of | (iii)
0.566667	-12.4787	yields | all
0.20524	-12.4789	a | hierarchical
0.122102	-12.479	use | which
0.299065	-12.4791	the | confident
0.160858	-12.4793	an | over
0.138614	-12.4794	alignments | the
0.19305	-12.4795	the | categories.
0.418182	-12.4797	the | labeler
0.514286	-12.48	in | expression,
0.846154	-12.4801	1 introduction | long
0.44	-12.4801	a | reviews.
0.666667	-12.4802	classifier, | classifier
0.222798	-12.4803	we show | for
0.212264	-12.4805	word | can be
0.355263	-12.4807	substitute | the
0.371429	-12.4808	is | and,
0.431373	-12.4809	“ | “
0.296296	-12.4809	however, | performance
0.337349	-12.481	relations | their
0.198347	-12.481	some | such
0.262774	-12.4811	in | ontology
0.529412	-12.4811	university | singapore
0.0902954	-12.4814	• | a
0.666667	-12.4815	dependent | dependent
0.25	-12.4815	moses | and
0.20524	-12.4819	etc. | of
0.0407637	-12.4821	the | has
0.337349	-12.4823	and | van
0.245283	-12.4823	by | we can
0.132727	-12.4824	we | whether
0.0680077	-12.4824	are | or
0.146018	-12.4826	in | through
0.410714	-12.4828	manually | corpus
0.283333	-12.4828	the | dependencies.
0.197581	-12.4829	formal | of
0.337349	-12.4831	(2010) | using
0.239521	-12.4832	that | appropriate
0.529412	-12.4833	news | web
0.393443	-12.4833	features | token
0.264706	-12.4833	since | will
0.196	-12.4835	is | further
0.75	-12.4836	why | so
0.446809	-12.4836	the | workshop.
0.258741	-12.4836	the | head.
0.341463	-12.4837	a | context-free
0.157623	-12.4838	give | of
0.105202	-12.484	we use | in
0.423077	-12.4841	entailment | between
0.174603	-12.4841	describe | for
0.909091	-12.4843	nouns | determiners
0.240964	-12.4846	for | domain.
0.592593	-12.4846	manager | and
0.289474	-12.4847	had | be
0.418182	-12.4847	both | high
0.240964	-12.4848	only | sentences
0.158031	-12.4849	to | individual
0.5	-12.4849	this | possible.
0.418182	-12.485	method | applied
0.138277	-12.4851	the | representing
0.431373	-12.4852	approach | able
0.245283	-12.4852	investigate | for
0.104664	-12.4853	point | the
0.206278	-12.4854	of | biomedical
0.378788	-12.4855	in | output,
0.384615	-12.4855	statement | a
0.123223	-12.4856	always | the
0.269231	-12.4856	model | models.
0.3125	-12.4857	nonterminal | a
0.529412	-12.4858	of | technology.
0.255172	-12.4858	extracting | in
0.280992	-12.4858	can | several
0.157216	-12.4858	to | words,
0.293578	-12.4859	the | generated.
0.350649	-12.486	up | words
0.393443	-12.486	coreference resolution | in
0.19084	-12.486	only | our
0.134981	-12.4861	the | corpora.
0.231638	-12.4863	computes | the
0.566667	-12.4863	various | these
0.299065	-12.4863	by | annotators
0.246835	-12.4864	is | complete
0.214286	-12.4869	of | (as
0.155	-12.4871	for example, | with
0.0763636	-12.4872	values | the
0.131532	-12.4873	a | space
0.384615	-12.4873	use | represent
0.289474	-12.4875	event, | the
0.341463	-12.4875	that | competitive
0.5	-12.4875	also | adding
0.104427	-12.4876	of | et | 2010).
0.666667	-12.4877	for | inputs.
0.346154	-12.4877	the | interest.
0.514286	-12.4878	model | jointly
0.350649	-12.488	denote | by
0.75	-12.488	capture | dependencies
0.423077	-12.4885	summary | sentences
0.120427	-12.4885	second, | the
0.571429	-12.4885	weighted | weights
0.625	-12.4886	model | simultaneously
0.387097	-12.4887	an | process.
0.909091	-12.4887	7 | tutor
0.25	-12.4889	index | of
0.170732	-12.4891	the | missing
0.104839	-12.4892	were | as
0.134216	-12.4893	use | model
0.592593	-12.4893	derivation | rules
0.231638	-12.4894	one | these
0.193676	-12.4896	also | different
0.341463	-12.4896	the | cited
0.191571	-12.4898	is | whose
0.3125	-12.4899	of | n.
0.465116	-12.4903	comparison | baseline
0.722222	-12.4903	impact | impact
0.25	-12.4905	is | grammatical
0.722222	-12.4905	measures | overlap
0.253425	-12.4906	relations | between
0.21028	-12.4907	typically | a
0.337349	-12.4907	parsing | sentence
0.592593	-12.4908	places | in
0.431373	-12.491	gives | results
0.232955	-12.491	all | single
0.0929603	-12.4913	2 | 2
0.166181	-12.4919	we | full
0.188192	-12.492	distributions | and
0.6	-12.492	sl | rl
0.350649	-12.4921	ranking | each
0.0705029	-12.4923	first | a
0.6	-12.4923	pittsburgh, | 15213,
0.135922	-12.4924	much | in
0.21028	-12.4926	p | f
0.341463	-12.4926	summationdisplay | (2)
0.300971	-12.4926	results | analysis
0.666667	-12.4928	occurring | times
0.393443	-12.4929	this | explained
0.423077	-12.4932	this | enables
0.625	-12.4932	another interesting | is
0.183099	-12.4933	in | resulting
0.47619	-12.4938	accuracy | correctly
0.25	-12.494	it | without
0.143478	-12.4941	report | of
0.116643	-12.4943	of | quality
0.4375	-12.4946	from | blog
0.274194	-12.4947	we | calculate
0.446809	-12.4947	was | final
0.378788	-12.4948	the | networks.
0.230337	-12.4948	section | describes | and
0.337349	-12.4949	sentences | document
0.3125	-12.4954	indicate | or
0.350649	-12.4957	by | via
0.446809	-12.4957	there | total
0.280992	-12.4957	on | corpora.
0.393443	-12.4964	and | names.
0.431373	-12.4967	using | seed
0.0989691	-12.4967	the | resulting
0.139059	-12.4969	mean | of
0.239264	-12.4969	the | pattern.
0.240741	-12.4969	novel | that
0.207207	-12.4969	a | probabilistic
0.195918	-12.497	of | identification
0.514286	-12.4973	generated | original
0.214634	-12.4973	once | a
0.403509	-12.4974	parsing | grammar
0.284483	-12.4974	our | perform
0.238095	-12.4974	methodology | and
0.53125	-12.4974	events, | event
0.125206	-12.4975	first | by
0.318681	-12.4976	only | very
0.337349	-12.4978	our | comparison
0.2	-12.4979	in | contexts
0.722222	-12.4983	update | parameters
0.077707	-12.4984	with | or
0.625	-12.4984	these | along with
0.131627	-12.4985	a | meaning
0.6	-12.4985	found | perform
0.17134	-12.4986	in | but | not
0.286957	-12.4987	of | ones.
0.236686	-12.4988	however, | different
0.53125	-12.4991	frequencies | words
0.213592	-12.4996	6.1 | the
0.410714	-12.5001	different | values
0.592593	-12.5001	using | tool
0.140167	-12.5002	our | over
0.387097	-12.5004	because | many
0.373134	-12.5005	is | predictive
0.170279	-12.5005	the | above.
0.5	-12.5005	those | government.
0.209302	-12.5006	setting | to
0.300971	-12.5007	for | computation
0.373134	-12.5009	(see | 1).
0.75	-12.5013	from | lexicon,
0.182456	-12.5013	system, | a
0.361111	-12.5014	of | reference.
0.403509	-12.5015	are | objects
0.17134	-12.5015	the | produces
0.279661	-12.5016	performance | systems
0.341772	-12.5016	centroid | the
0.722222	-12.5016	this | because,
0.246753	-12.5017	among | are
0.223404	-12.5018	core | and
0.514286	-12.5018	though | may
0.378788	-12.5018	quite | from
0.341772	-12.5018	immediate | of
0.415094	-12.5018	aligned | english
0.276423	-12.502	compute | between
0.6	-12.5021	synsets | synset
0.465116	-12.5024	was | french
0.290909	-12.5024	also, | a
0.487179	-12.5025	(see | 3).
0.213592	-12.5028	the | cluster.
0.333333	-12.5028	and | expressions.
0.446809	-12.5029	the | programme
0.26087	-12.5031	investigation | the
0.350649	-12.5032	a | desired
0.361111	-12.5033	has | meaning
0.185455	-12.5033	the | page
0.180556	-12.5035	for | models.
0.265152	-12.5038	definition | that
0.110677	-12.5038	examples | and
0.0990615	-12.5043	its | is
0.218274	-12.5044	we | 100
0.245161	-12.5044	acknowledgments we | for
0.393443	-12.5045	first | after
0.666667	-12.5047	size | length
0.361111	-12.5047	with | methods.
0.262774	-12.5048	we | moses
0.235294	-12.5048	each | value
0.351351	-12.5049	combination | different
0.199153	-12.5053	of | predicates
0.286957	-12.5053	uses | model
0.846154	-12.5054	calculated | equation
0.666667	-12.5054	such | etc.,
0.423077	-12.5054	of | daily
0.548387	-12.5056	this | caused by
0.548387	-12.5056	aligned | alignment
0.181185	-12.506	the | events.
0.909091	-12.5065	which | defined.
0.318681	-12.5066	include | features
0.548387	-12.5068	could | be used
0.237805	-12.5068	mechanism | the
0.415094	-12.5071	in | 1),
0.274194	-12.5072	calculating | of
0.159459	-12.5073	matches | the
0.119335	-12.5073	do not | to
0.6	-12.5074	documents | collection
0.269231	-12.5076	the | ending
0.235294	-12.5078	c 2011 | linguistics
0.387097	-12.5079	as | previously
0.666667	-12.508	representing | meaning
0.351351	-12.508	agent | a
0.231214	-12.5082	and | novel
0.846154	-12.509	accuracy, | accuracy
0.846154	-12.509	translating | morphologically
0.722222	-12.509	standard | dialect
0.175497	-12.5093	counts | and
0.636364	-12.5094	k. | m.
0.909091	-12.5094	(papineni | (doddington,
0.4375	-12.5096	gives | number
0.0838323	-12.5096	the | labels
0.282051	-12.5097	baseline | only
0.53125	-12.5097	using | default
0.341772	-12.5099	closer | to
0.17134	-12.5101	this | many
0.571429	-12.5101	report | scores
0.428571	-12.5102	these | capture
0.245161	-12.5102	grammar, | the
0.239264	-12.5103	• | features
0.341772	-12.5105	performed | using
0.487179	-12.5106	good | should
0.288288	-12.5107	that | increases
0.213592	-12.5109	model | better
0.300971	-12.5109	most common | of
0.210526	-12.5109	the | rule-based
0.487179	-12.5111	is | subgraph
0.221053	-12.5113	a | answer
0.785714	-12.5113	action | action.
0.100868	-12.5117	the | agreement
0.237805	-12.5119	calculated | a
0.3375	-12.512	original | translated
0.298077	-12.5121	the | parameters,
0.188462	-12.5122	section 2 | the
0.356164	-12.5123	multiple | different
0.322222	-12.5125	some | about
0.571429	-12.5125	definition | its
0.180556	-12.5127	association | 19-24,
0.373134	-12.5127	for | uniform
0.306122	-12.5128	the | two.
0.333333	-12.5128	one | multiple
0.267176	-12.5129	after | each
0.195833	-12.5135	relevance | the
0.0732685	-12.5137	first | in
0.403509	-12.5138	hence, | are
0.684211	-12.5138	takes | produces
0.446809	-12.5139	in | generation.
0.0579971	-12.514	this | is | to
0.684211	-12.514	given | text,
0.636364	-12.5142	which | explain
0.909091	-12.5142	compares | against
0.148936	-12.5142	the | ambiguous
0.6	-12.5144	sequence | tokens
0.846154	-12.5147	improvement | achieved
0.487179	-12.5149	over | last
0.189922	-12.5152	and | take
0.136546	-12.5155	data | training
0.224044	-12.5155	variations | the
0.333333	-12.5156	a | selects
0.551724	-12.5159	m. | j.
0.475	-12.5161	common | both
0.785714	-12.5163	shorter | longer
0.232558	-12.5163	object | a
0.548387	-12.5165	to | point,
0.341772	-12.5166	of | word’s
0.475	-12.5167	these | via
0.306122	-12.5168	of | assignments
0.114047	-12.5168	the last | the
0.475	-12.5169	in | evaluations.
0.341772	-12.517	assumption, | the
0.4375	-12.5172	the | connective.
0.315217	-12.5173	huge | of
0.551724	-12.5175	access | information
0.199134	-12.5176	section | describes
0.21608	-12.5177	easily | to
0.279661	-12.5178	2.2 | is
0.232558	-12.5178	is | predicted
0.514286	-12.5179	imbalance | the
0.4375	-12.5179	we | mechanical turk
0.475	-12.5181	method | bilingual
0.389831	-12.5181	contribution | each
0.423077	-12.5184	because | so
0.309278	-12.5184	tweets | are
0.322222	-12.5184	types | their
0.114407	-12.5187	need to | the
0.487179	-12.5187	set | 500
0.288288	-12.5189	syntax | a
0.298077	-12.519	specifications | the
0.279661	-12.5192	workshop | and | 2011.
0.255319	-12.5196	this | makes
0.5	-12.5196	are | nouns.
0.194332	-12.5197	special | of
0.309278	-12.5197	as | separate
0.178694	-12.5201	lattice | the
0.30303	-12.5201	experiments | english
0.2	-12.5201	that | always
0.341772	-12.5202	but | about
0.329412	-12.5203	include | words
0.375	-12.5204	figure 3 shows | of
0.4375	-12.5205	of | alone.
0.0969697	-12.5207	approaches | and
0.75	-12.5207	minimum | (mert)
0.362319	-12.5207	a | rule,
0.415094	-12.5209	enforce | of
0.185874	-12.521	discussion | in
0.551724	-12.521	citation | citation
0.189922	-12.5211	is | parameter
0.193548	-12.5212	in | domain.
0.428571	-12.5213	classification | classifier
0.261194	-12.522	a | boundary
0.487179	-12.522	is | ideal
0.163743	-12.5221	of | potential
0.636364	-12.5221	syntactic | grammatical
0.487179	-12.5222	classifying | or
0.101099	-12.5224	is | where | and
0.396552	-12.5224	and | synonyms
0.487179	-12.5225	systems, | system
0.909091	-12.5226	sub-word | sub-word
0.785714	-12.5233	dialogues | dialogues
0.232558	-12.5233	the | embedded
0.129562	-12.5234	with respect to | the
0.514286	-12.5235	that | signal
0.200873	-12.5236	et | 2005)
0.236364	-12.5236	constraint | in
0.168196	-12.5236	that | being
0.259259	-12.5237	the | chooses
0.6	-12.5238	method | various
0.178082	-12.5239	for | within
0.292453	-12.5241	to | possibly
0.341772	-12.5246	sense disambiguation | a
0.176271	-12.5246	in | dictionary
0.24359	-12.5248	such | may
0.298077	-12.5248	as | initial
0.121019	-12.525	the | identified
0.288288	-12.5252	number | topics
0.216495	-12.5253	techniques | for
0.329412	-12.5253	that | external
0.407407	-12.5254	word | appears
0.396552	-12.5255	these | require
0.233533	-12.5256	without | with
0.211823	-12.5258	in | article
0.2	-12.526	the | acoustic
0.265625	-12.5262	baselines | and
0.238994	-12.5262	stage | to
0.6	-12.5262	adjectives | verbs
0.292453	-12.5264	that | needed
0.25	-12.5264	a | element
0.156085	-12.5266	the | needs
0.176871	-12.5267	want | to
0.380952	-12.5269	the | involved.
0.515152	-12.5269	infer | from
0.265625	-12.527	though | not
0.684211	-12.527	project | framework
0.114943	-12.527	the same | for
0.318681	-12.5271	the | sequence,
0.684211	-12.5272	relatively | relatively
0.551724	-12.5273	british | corpus
0.126298	-12.5273	study | and
0.475	-12.5274	is | requiring
0.515152	-12.5275	selection | candidate
0.475	-12.5276	decoding | component
0.126079	-12.5277	the | appears
0.75	-12.5279	before | next
0.228571	-12.528	empirically | the
0.121212	-12.5283	the | (section
0.190476	-12.5283	the | runs
0.126298	-12.5285	of | models.
0.389831	-12.5285	yielding | of
0.114085	-12.5286	that | would
0.415094	-12.5287	and | generalization
0.184502	-12.5288	helps | the
0.396552	-12.5288	the | corpus).
0.53125	-12.5289	takes | its
0.109804	-12.5293	the | system,
0.321839	-12.5295	source | information
0.318681	-12.5296	landmark | the
0.195833	-12.5299	adapted | the
0.178947	-12.53	text | or
0.0843653	-12.5304	in | text
0.261194	-12.5305	bit | the
0.179577	-12.5306	as | example
0.346667	-12.5307	we | experiment.
0.263158	-12.531	(5) | is
0.0473402	-12.531	most | the
0.257353	-12.5313	recognized | the
0.75	-12.5315	assignments | all
0.240506	-12.5315	for | word,
0.12275	-12.5316	case | a
0.367647	-12.5316	so | to be
0.342105	-12.5316	utterance, | the
0.1777	-12.5317	of | article
0.0730858	-12.5319	the | documents
0.187023	-12.5321	the | cases.
0.230769	-12.5321	em | to
0.415094	-12.5324	normal | simple
0.179577	-12.5325	improved | of
0.325581	-12.5326	function, | a
0.225989	-12.5328	words | topic
0.396552	-12.5328	labelled | and
0.272727	-12.533	tag | word
0.198276	-12.533	is | when | are
0.357143	-12.533	unigrams | and
0.141304	-12.533	the | gene
0.4375	-12.5333	consistently | than
0.75	-12.5336	classifier | polarity
0.515152	-12.5336	conversely, | in
0.396552	-12.5338	5.2 | results
0.272727	-12.534	100 | with
0.42	-12.5342	the | (excluding
0.117293	-12.5343	the | represents
0.123746	-12.5344	labeled | and
0.321839	-12.5345	took | to
0.454545	-12.5348	model | log-linear
0.6	-12.5348	scores | summaries
0.206573	-12.5349	we | report
0.362319	-12.5349	unlike | our
0.463415	-12.535	in | ng
0.415094	-12.535	these | categories
0.3	-12.5352	structures | are
0.25	-12.5352	is | required
0.321839	-12.5353	and | subjects
0.154047	-12.5354	expected | of
0.75	-12.5355	only | considered.
0.25	-12.5356	of | nodes.
0.224719	-12.5357	lexical | semantic
0.909091	-12.5357	framenet | framenet
0.6	-12.5359	measure | reliability
0.515152	-12.536	corpora | corpora.
0.463415	-12.536	corpus | created
0.219251	-12.536	:= | :=
0.144495	-12.5361	n-grams | the
0.380952	-12.5362	and | only.
0.3	-12.5363	this | achieved
0.6	-12.5364	[np@1 | squaresmallsolid
0.75	-12.5366	3, | 5,
0.245033	-12.5367	the | adopted
0.198276	-12.5367	almost | of
0.636364	-12.5368	here | input
0.6	-12.5368	align | words
0.229412	-12.537	identification | in
0.346667	-12.5371	besides, | the
0.329268	-12.5372	are | predicted
0.515152	-12.5372	with | weights.
0.11828	-12.5373	the | rank
0.248276	-12.5373	6 | 8
0.383333	-12.5374	measuring | between
0.263566	-12.5376	on | right
0.357143	-12.5376	with | semantics
0.346667	-12.538	lin | and
0.785714	-12.5384	each | independently.
0.684211	-12.5384	among | first
0.177083	-12.5386	first | using
0.329268	-12.5388	evaluation | quality
0.475	-12.5388	resulting | all
0.463415	-12.5389	able | find
0.267717	-12.5391	documents | each
0.571429	-12.5392	this | surprising
0.6	-12.5392	designed | use
0.197425	-12.5393	the | effort
0.175084	-12.5395	extension | the
0.221622	-12.5396	further, | the
0.101947	-12.5398	of | being
0.42	-12.54	a | edges.
0.178322	-12.5407	all | but
0.270492	-12.5407	increased | to
0.156757	-12.5408	3.1 | a
0.0417865	-12.5408	given | the
0.833333	-12.5408	22 | 24
0.278261	-12.5409	the | affects
0.201794	-12.5409	that | appear
0.75	-12.541	tuned | dev
0.75	-12.541	it | noteworthy
0.29703	-12.541	which | important
0.2	-12.5411	(2007) | and
0.785714	-12.5412	re-ranking | n-best
0.333333	-12.5412	two | approaches
0.138075	-12.5413	only | which
0.909091	-12.5414	lexicons | lexicons
0.357143	-12.5414	words | topics
0.168224	-12.5415	however | the
0.389831	-12.5415	translation | translating
0.318182	-12.5417	these | form
0.636364	-12.5419	trained | europarl
0.206573	-12.542	a | procedure
0.5	-12.5422	pairwise | from
0.215385	-12.5424	2001; | et
0.221622	-12.5425	are | common
0.164671	-12.5425	learning | from
0.287037	-12.5425	selected | each
0.15445	-12.5426	a | term
0.205607	-12.5426	ability | to
0.101598	-12.5426	the | performed
0.30303	-12.5427	the | inventory
0.389831	-12.5427	the | german.
0.486486	-12.5431	employ | which
0.272727	-12.5432	the | option
0.212121	-12.5433	a | confidence
0.246575	-12.5435	restriction | the
0.833333	-12.5435	iteratively | until
0.123128	-12.5438	identified | the
0.272727	-12.5439	– | •
0.684211	-12.5439	incorporated | our
0.6	-12.544	less than | times
0.314607	-12.5442	extracts | and
0.146226	-12.5443	takes | of
0.181818	-12.5443	using | feature
0.329268	-12.5444	= | productdisplay
0.141907	-12.5444	in | data,
0.551724	-12.5444	we | 5-fold
0.251748	-12.5445	a | topic.
0.14554	-12.5447	is | without
0.909091	-12.5447	should not | interpreted
0.369231	-12.5447	use | labels
0.321839	-12.5448	great | in
0.311828	-12.5449	a | type,
0.576923	-12.5449	(mcclosky | al.,
0.53125	-12.5451	more | added
0.475	-12.5452	information, | or
0.833333	-12.5453	with | 0.01.
0.292453	-12.5455	our | further
0.705882	-12.5455	integrated | into
0.164671	-12.5457	so, | the
0.153646	-12.5458	for | 1.
0.212121	-12.5459	consistency | the
0.909091	-12.546	compute | expectations
0.261538	-12.5461	as | base
0.151134	-12.5462	local | and
0.65	-12.5463	policy | real
0.65	-12.5463	mwes | their
0.909091	-12.5463	(kilicoglu | bergler,
0.909091	-12.5463	filippova | strube,
0.909091	-12.5463	hgfc | agg
0.115899	-12.5463	of | do not
0.158774	-12.5464	the | reduced
0.515152	-12.5466	experiments | datasets
0.121911	-12.5467	in | corresponding
0.608696	-12.5469	multiword | parsing | (mwe
0.234568	-12.547	absolute | of
0.163205	-12.547	actions | the
0.251748	-12.5471	kept | the
0.18677	-12.5472	results | baseline
0.205607	-12.5476	a significant | of
0.785714	-12.5479	will | later.
0.259259	-12.5479	is | key
0.294118	-12.548	these | 1.
0.0696304	-12.5481	is | but
0.65	-12.5482	gesture | gesture
0.233129	-12.5484	lower | for
0.75	-12.5487	method | methods,
0.389831	-12.5487	propose | algorithm
0.141892	-12.5488	3 | 5
0.243421	-12.5489	of | arcs
0.270492	-12.5489	of | rule.
0.128913	-12.5489	the | towards
0.833333	-12.549	semeval-2010 | task
0.65	-12.5492	you | what
0.357143	-12.5493	model | capture
0.308511	-12.5495	the | lists.
0.428571	-12.5496	now | have
0.223464	-12.5497	efficient | of
0.255474	-12.5497	of | scientific
0.463415	-12.5499	used | support
0.369231	-12.5499	these | contain
0.16092	-12.55	second | that
0.194093	-12.55	in | e.g.
0.246575	-12.5502	such | would
0.515152	-12.5504	this | computationally
0.337662	-12.5506	hypothesis | the | hypothesis
0.342105	-12.5507	we | infer
0.833333	-12.5508	room | improvement.
0.65	-12.5509	to address | issue
0.515152	-12.5509	likely | contain
0.576923	-12.551	top | highest
0.428571	-12.5512	100 | 500
0.452381	-12.5513	which | sequences
0.434783	-12.5514	but | cannot
0.375	-12.5514	indicates that | are
0.150376	-12.5516	use | features
0.0947159	-12.5519	of | form
0.463415	-12.552	the | stream.
0.833333	-12.5521	had | success
0.215789	-12.5522	section 5 | the
0.705882	-12.5522	jhu | combination
0.073716	-12.5523	for each | and
0.5	-12.5523	combine | two
0.131021	-12.5525	help | the
0.705882	-12.5525	structured | algorithm
0.53125	-12.5527	tutor | tutor
0.137712	-12.5529	data | used
0.230769	-12.5529	influence | of
0.272727	-12.553	and | events.
0.636364	-12.5531	with | predefined
0.214286	-12.5531	the | suffix
0.152062	-12.5533	the | condition
0.434783	-12.5535	effect | different
0.2	-12.5535	somewhat | the
0.576923	-12.5536	university, | usa
0.122483	-12.5537	et | 2010),
0.284404	-12.5539	also | similar
0.318182	-12.5541	so | they
0.287037	-12.5541	a | function,
0.185328	-12.5542	the | input.
0.785714	-12.5545	run | experiment
0.65	-12.5547	through | analysis
0.176471	-12.5548	specifically, | and
0.261538	-12.5549	the | longest
0.165138	-12.555	a | token
0.75	-12.555	graph, | graph
0.126335	-12.555	ratio | the
0.452381	-12.5553	concepts | semantic
0.375	-12.5554	language | toolkit
0.213198	-12.5555	examples | from
0.347222	-12.5556	following | ,
0.705882	-12.5556	possible | under
0.407407	-12.5556	verb | object
0.308511	-12.556	our | (section
0.4	-12.5561	are | easy
0.411765	-12.5561	compare | baseline
0.576923	-12.5561	3 | architecture
0.117647	-12.5562	matching | the
0.318182	-12.5562	on | dev
0.551724	-12.5563	rank | according
0.183521	-12.5563	messages | the
0.0396022	-12.5564	the | its
0.75	-12.5564	negationslash= | negationslash=
1	-12.5564	bleu: | method
1	-12.5564	4th | (cslt
1	-12.5564	brianna | bedroom
1	-12.5564	finkel | manning
1	-12.5564	fowler | penn
1	-12.5564	(resnik | smith,
1	-12.5564	palmer | colorado
1	-12.5564	parallel | parallel.
1	-12.5564	grammaticality, | preservation,
1	-12.5564	(habash | rambow,
1	-12.5564	birke | sarkar
1	-12.5564	(stone | 1966)
1	-12.5564	protocol | protocol
0.22807	-12.5564	written | a
0.463415	-12.5564	that | triggers
0.473684	-12.5567	systems | output
0.9	-12.5568	(bojar | ˇzabokrtsk´y,
0.0753165	-12.5568	we | into
0.533333	-12.5569	limit | number
0.201835	-12.5572	are | being
0.305263	-12.5574	the | polarity.
0.114537	-12.5575	procedure | the
0.392857	-12.5575	we discuss | our
0.42	-12.5576	which | identified
0.551724	-12.5576	bionlp | (kim
0.109905	-12.5576	impact | the
0.148148	-12.5576	models | model
0.9	-12.5581	perceptron | (collins,
0.636364	-12.5583	ability | identify
0.216931	-12.5584	is | (i.e.,
0.515152	-12.5584	priori | to
0.151282	-12.5584	the | tables
0.257576	-12.5587	set | candidate
0.0955807	-12.5588	a | relation
0.202765	-12.5588	this | important
0.266129	-12.5589	optimal | is
0.347222	-12.5589	merging | and
0.705882	-12.559	degree | contribution
0.9	-12.5591	multilingual | multilingual
0.138528	-12.5592	these | have
0.325301	-12.5592	the | hand.
0.551724	-12.5592	proposed | classification
0.705882	-12.5593	1999; | 2002;
0.9	-12.5594	national | canada
0.218085	-12.5595	we | tweets
0.377049	-12.5595	for | summarization.
0.608696	-12.5595	in | experiments:
0.244898	-12.5598	three | one
0.230303	-12.56	the | vocabulary.
0.261538	-12.56	of | plus
0.636364	-12.5602	lexical | acoustic
0.705882	-12.5603	produced | quality
0.452381	-12.5603	perform | task
0.23871	-12.5605	f1 | on
0.202765	-12.5605	a | making
0.357143	-12.5605	task | events
0.636364	-12.5606	proportion | all
0.342105	-12.5608	aggregate | of
0.444444	-12.561	and | dutch
0.9	-12.5611	sort | readability
0.194093	-12.5611	is | effective
0.533333	-12.5612	the | were:
0.411765	-12.5613	conduct | on
0.9	-12.5615	cosine | angle
0.515152	-12.5616	creating | from
0.833333	-12.5616	deep | shallow
0.244898	-12.5617	translations | are
0.321429	-12.5618	in | chart
0.684211	-12.5618	drop | when
0.244898	-12.5618	similarly, | a
0.434783	-12.562	a | dependency parser
0.15051	-12.562	a | prior
0.608696	-12.562	method | i.e.,
0.785714	-12.5621	domains | target
0.218085	-12.5621	transition | to
0.515152	-12.5622	(such | (such
0.149367	-12.5622	difference | is
0.25	-12.5623	additionally, | to
0.145585	-12.5625	identify | in
0.337662	-12.5626	that | typical
0.473684	-12.5627	arguments | events
0.9	-12.5628	robust | detection
0.162242	-12.5628	tool | the
0.145585	-12.5628	a | corpus,
0.180812	-12.5628	and | produce
0.533333	-12.5628	acronyms | in
0.141892	-12.5631	corresponding | in
0.444444	-12.5631	and | use.
0.9	-12.5635	paired | resampling
0.0781893	-12.5635	the | the original
0.785714	-12.5636	i.e., | i.e.,
0.0871795	-12.5636	method | in
0.533333	-12.5636	groups | their
0.325301	-12.5638	5 conclusion | a
0.608696	-12.5638	an | x,
0.377049	-12.564	specificity | of
0.199095	-12.5642	test | using
0.266129	-12.5642	corrections | the
0.291262	-12.5642	two | 1.
0.0947581	-12.5644	of | documents
0.18251	-12.5646	difference | that
0.636364	-12.5649	sygal | wintner
0.288462	-12.565	when | can be
0.636364	-12.5651	we perform | experiments
0.0926641	-12.5653	measures | the
0.298969	-12.5653	in | matched
0.358209	-12.5656	mean | per
0.278261	-12.5658	the | (denoted
0.608696	-12.5658	code | at
0.608696	-12.566	dots | connected
0.705882	-12.5661	an | user.
0.318182	-12.5662	the | site
0.377049	-12.5664	for | documents,
0.684211	-12.5664	muc | b3
0.224138	-12.5664	(2009) | for
0.463415	-12.5664	study | these
0.576923	-12.5664	their | scores.
0.257576	-12.5665	exact | is
0.235669	-12.5666	the | a,
0.094898	-12.5669	comparison | the
0.705882	-12.5669	convey | same
0.102353	-12.5669	of | knowledge
0.207921	-12.567	the | parser,
0.608696	-12.5672	result | correct
0.4	-12.5673	a | investigation
0.42	-12.5673	active | active
0.370968	-12.5673	a | sign
0.311111	-12.5674	the | stems
0.377049	-12.5678	and | indicators
0.321429	-12.5678	take | account | of
0.112536	-12.568	an | model
0.444444	-12.5681	the | operation.
0.24026	-12.5684	using | model.
0.551724	-12.5685	what | might
0.231707	-12.5688	best | this
0.152231	-12.5689	have | but
0.9	-12.5689	topics | jst
0.261905	-12.5689	cn | cn
0.576923	-12.569	is | decay
0.21	-12.569	of | strings
0.608696	-12.5692	approaches | 2007).
0.275862	-12.5693	translating | to
0.4	-12.5694	show | improvements
0.337662	-12.5696	work | framework
0.271186	-12.5696	in | adjective
0.243243	-12.5697	allowed | to
0.515152	-12.5697	whereas | has
0.259542	-12.5697	these | new
0.377049	-12.5698	in | paragraph
0.4	-12.5698	translation | monolingual
0.298969	-12.5699	c. | and
0.463415	-12.5699	argmax | argmax
0.0782548	-12.57	documents | the
0.208955	-12.5702	provide | that
0.167192	-12.5706	who | a
0.9	-12.5709	2010 | american
0.112392	-12.5709	we also | a
0.515152	-12.5709	such as | structure
0.325301	-12.571	pronoun | a
0.833333	-12.571	unaligned | words.
0.705882	-12.5711	methods | learning,
0.337662	-12.5712	system | reference
0.101163	-12.5715	this | but
0.0399349	-12.5715	the | as | as
0.576923	-12.5716	alignment | authority
0.65	-12.5717	stanford | (toutanova
0.100811	-12.5717	of | features.
0.425532	-12.5718	achieved | results
0.358209	-12.5718	a | node,
0.118211	-12.5721	2008; | al.,
0.162162	-12.5722	we use | with
0.181818	-12.5722	generation | in
0.188525	-12.5725	measured | of
0.515152	-12.5726	tagging | tagging
0.178832	-12.5726	the | challenge
0.257576	-12.5726	and | quality.
0.555556	-12.5728	range | including
0.298969	-12.5728	second, | are
0.174216	-12.5729	the | book
0.9	-12.5729	of acl | pages
0.201878	-12.5729	the | year
0.576923	-12.5729	term, | a
0.833333	-12.5732	many | remain
0.266129	-12.5733	for | optimal
0.25	-12.5733	show | when
0.259542	-12.5733	model | uses
0.227545	-12.5737	designed to | of
0.358209	-12.5738	spread | the
0.295918	-12.5739	the | continues
0.201878	-12.5741	are | which are
0.0720614	-12.5742	our | is | and
0.515152	-12.5742	it | crucial
0.411765	-12.5744	we | adopt
0.311111	-12.5744	four | each
0.377049	-12.5747	research | (afrl)
0.65	-12.5748	relevant | query
0.23871	-12.5748	author | and
0.833333	-12.5748	among | candidates
0.103659	-12.5749	4 | for
0.0946083	-12.5749	several | of
0.188525	-12.5749	3, | and
0.392857	-12.575	a | billion
0.120861	-12.575	the | consists of
0.370968	-12.575	word | dictionary
0.125224	-12.5751	it | possible
0.352113	-12.5751	characterize | of
0.12305	-12.5752	start | the
0.473684	-12.5752	a | authority
0.329114	-12.5758	features | include
0.126838	-12.5758	all | are | in
0.302083	-12.5759	analyzing | in
0.608696	-12.5759	procedure | 1.
0.284404	-12.576	state-of-the-art | on
0.463415	-12.5762	consider | example
0.232704	-12.5766	the | drop
0.434783	-12.5766	a | formalism
0.425532	-12.5766	at | rate
0.171717	-12.5767	for | sentence.
0.785714	-12.5768	from | logs.
0.125899	-12.5768	a | defined
0.135881	-12.5769	the | matches
0.257576	-12.5769	that | semantically
0.311111	-12.577	has | possible
0.5	-12.5771	each | t,
0.230303	-12.5772	a | learner
0.116641	-12.5773	the | monolingual
0.515152	-12.5773	able | extract
0.533333	-12.5773	significant | performance.
0.102138	-12.5773	a | topic
0.463415	-12.5774	we combine | with
0.403846	-12.5775	what | you
0.342466	-12.5776	by: | =
0.288462	-12.5776	whereas | for
0.352113	-12.5776	largest | in
0.705882	-12.5778	to classify | into
0.21978	-12.5778	reflects | the
0.143868	-12.5779	developed | of
0.9	-12.578	be | implied,
0.311111	-12.5781	of | equivalents
0.463415	-12.5783	∗ | ∗
0.463415	-12.5783	suggested | by
0.347222	-12.5784	although | results
0.248227	-12.5784	of | engineering
0.555556	-12.5785	impact | performance.
0.253731	-12.5786	estimated | a
0.0979098	-12.5787	the | version
0.266667	-12.5791	we calculate | of
0.291262	-12.5795	the | ngrams
0.0869191	-12.5797	the | during
0.325301	-12.58	wall | journal
0.317647	-12.58	a | immediately
0.9	-12.5804	cz-en | fr-en
0.403846	-12.5804	that | unseen
0.237179	-12.5804	sampling | for
0.190871	-12.5805	the | “the
0.434783	-12.5808	into | categories
0.170569	-12.5809	most | which
0.131213	-12.5809	a | including
0.176895	-12.581	we | determine
0.196429	-12.5812	which | then
0.144928	-12.5812	categories | and
0.9	-12.5814	day | –
0.266667	-12.5818	but | any
0.179775	-12.5818	identification | and
0.234177	-12.5819	entities | are
0.081203	-12.5819	however, | is
0.212435	-12.582	and | accuracy.
0.253731	-12.5821	calculation | the
0.370968	-12.5821	by | absolute
0.23125	-12.5822	the | relies on
0.9	-12.5824	(whether | or
0.224138	-12.5824	the | datasets.
0.411765	-12.5824	and | mentions.
0.313953	-12.5824	the | answer.
0.177536	-12.5825	of | sequences
0.425532	-12.5826	sixth | on
0.0466154	-12.5828	the | task
0.516129	-12.583	translate | english
0.224852	-12.583	that | help
0.298969	-12.583	boolean | the
0.302083	-12.5832	top | each
0.162162	-12.5832	to | users
0.228916	-12.5834	experiment | with
0.452381	-12.5836	was | made
0.705882	-12.5837	is | costly
0.705882	-12.5837	j. | m.
0.313953	-12.5837	comparison, | of
0.215054	-12.584	curve | the
0.608696	-12.584	and | minutes
0.11746	-12.5842	we | those
0.342466	-12.5843	supports | and
0.150649	-12.5844	output | is
0.608696	-12.5844	runs | runs
0.217391	-12.5845	open | and
0.268908	-12.5845	particularly | in
0.244755	-12.5849	can | either | or
0.833333	-12.5849	subjective | problems
0.17	-12.5852	we | binary
0.516129	-12.5852	with | sizes
0.317647	-12.5853	frequency | which
0.65	-12.5854	build | models.
0.0747126	-12.5854	the | error
0.576923	-12.5854	a | (about
0.347222	-12.5856	sent | to
0.485714	-12.5857	for | generation,
0.705882	-12.5858	wordnet: | lexical
0.473684	-12.5858	described | based on
0.576923	-12.5858	translated | translations
0.9	-12.5862	hurt | performance
0.347222	-12.5862	grants | and
0.197309	-12.5865	achieves | of
0.358209	-12.5866	of | statements
0.65	-12.5867	belonging | same
0.241611	-12.5867	of | running
0.232704	-12.5868	a | interaction
0.129845	-12.5868	the | built
0.123894	-12.5871	using | model
0.229814	-12.5874	the | tweets.
0.0395328	-12.5874	et al., | and
0.197309	-12.5876	of | translations.
0.0601161	-12.5876	example | the
0.583333	-12.5877	exploit | features
0.9	-12.5879	inherited | its
0.9	-12.5879	based on | experience
0.833333	-12.588	in | subsection.
0.576923	-12.588	these | providing
0.352941	-12.5883	by | selecting
0.44186	-12.5883	is | parameter.
0.705882	-12.5883	bacteria | bacteria
0.733333	-12.5884	wsj | 23
0.24	-12.5885	otherwise, | the
0.298969	-12.5886	use | knowledge
0.516129	-12.5887	textual | text
0.434783	-12.5888	results | indicate
0.313953	-12.5888	the | author’s
0.311111	-12.5889	limitation | the
0.385965	-12.5889	model | built
0.347222	-12.589	minimizing | the
0.392857	-12.589	for | answering
0.232704	-12.5891	is | real
0.1491	-12.5892	of | known
0.452381	-12.5895	crf | crf
0.833333	-12.5896	exponentially | number
0.485714	-12.5896	instances | relations
0.533333	-12.5898	for | averaging
0.705882	-12.5899	work | award
0.1475	-12.5899	understanding | the
0.358209	-12.5899	quantitative | and
0.24	-12.59	the | model’s
0.416667	-12.5903	that | chance
0.246479	-12.5903	recall | recall
0.0855932	-12.5906	the | result
0.365079	-12.5907	are | estimated
0.266667	-12.5909	the | extracted.
0.65	-12.5909	for | tuning.
0.271186	-12.5909	for | training.
0.352941	-12.591	that | accurately
0.461538	-12.591	results | experiments.
0.166134	-12.5911	analyses | the
0.139013	-12.5912	4 | 4
0.337838	-12.5913	translated | original
0.268908	-12.5914	each | corresponds
0.133891	-12.5916	nlp | and
0.193833	-12.5918	a | span
0.65	-12.5919	optimal | optimal
0.0746461	-12.592	label | the
0.136364	-12.5921	row | the
0.555556	-12.5922	training | perceptron
0.244755	-12.5923	that | nouns
0.268908	-12.5924	the | genre
0.159763	-12.5925	with | probability
0.207071	-12.5925	if | any
0.241379	-12.5927	of | terms.
0.171233	-12.5928	the | results,
0.785714	-12.593	evaluations | systems
0.425532	-12.5931	counts | all
0.280374	-12.5931	our system, | the
0.0732919	-12.5934	this | our
0.155367	-12.5935	filter | the
0.785714	-12.5936	adverb | adjective
0.313953	-12.5937	a | elementary
0.161094	-12.5938	built | and
0.207071	-12.5938	projection | the
0.161094	-12.5942	are | for each
0.365079	-12.5943	constituency | and
0.184314	-12.5944	this work | in
0.365079	-12.5945	for | applications,
0.576923	-12.5945	topic | document.
0.461538	-12.5946	however, | limited
0.283019	-12.5947	assumption | are
0.785714	-12.5948	dn |=
0.555556	-12.5948	into | framework.
0.705882	-12.5948	contributions | follows:
0.283019	-12.5948	the | changed
0.44186	-12.5949	measure | performance
0.175	-12.5949	of | alignments
0.109116	-12.5954	discussion | the
0.295918	-12.5956	a | node.
0.0603121	-12.5958	also | in
0.156695	-12.5961	system | translation
0.666667	-12.5965	evaluate | components
0.0517669	-12.5966	evaluation | the
0.248175	-12.5966	in addition, | is
0.666667	-12.5967	min | max
0.193833	-12.5968	so | not
0.396226	-12.5968	words | wordnet
0.317647	-12.5968	that | potentially
0.785714	-12.5969	~ | “look
0.358209	-12.5969	an unsupervised | for
0.213904	-12.597	hence, | of
0.733333	-12.5971	represent | content
0.283019	-12.5971	(callison-burch | al.,
0.285714	-12.5972	the | component.
0.100711	-12.5972	new | in
0.185484	-12.5974	has | but
0.473684	-12.5979	task | participants
0.365079	-12.5979	achieved | best
0.121739	-12.598	finally, | and
0.233766	-12.5981	the | achieving
0.65	-12.5984	basis | basis
0.619048	-12.5985	each | once.
0.1168	-12.5985	the | collection
0.533333	-12.5985	using | method,
0.099768	-12.5988	has | with
0.304348	-12.599	model | phrases
0.485714	-12.599	inclusion of | in
0.163522	-12.5992	categories | in
0.133891	-12.5993	discourse | and
0.185484	-12.5998	those | by
0.202899	-12.5998	with | small
0.337838	-12.6001	the | cluster,
0.243056	-12.6002	thus | with
0.29	-12.6004	and | metrics.
0.313953	-12.6004	our | contains
0.174377	-12.6007	performed | a
0.259843	-12.6008	about | words
0.0978628	-12.6008	high | of
0.285714	-12.6011	generative | model
0.619048	-12.6011	funding | grant
0.533333	-12.6011	building | system
0.5	-12.6012	system | correctly
0.666667	-12.6014	an | event,
0.37931	-12.6014	and | reference.
0.301075	-12.6017	result | not
0.277778	-12.6019	how | use
0.207071	-12.602	nodes | are
0.403846	-12.6021	contains | pairs
0.44186	-12.6022	with | potentially
0.140878	-12.6022	any | that
0.666667	-12.6022	which | in.
0.239726	-12.6026	learning | using
0.153005	-12.6027	the | levels
0.359375	-12.6027	al., | wang
0.9	-12.6028	kneser-ney | (kneser
0.431818	-12.6029	moreover, | only
0.2	-12.6029	types of | are
0.396226	-12.603	while | less
0.159639	-12.603	string | and
0.425532	-12.6032	which | properties
0.238411	-12.6034	all | possible
0.583333	-12.6036	first | detection
0.347826	-12.6037	proceed | the
0.142518	-12.6037	this | work
0.533333	-12.6038	about | points
0.320988	-12.6039	participants | with
0.9	-12.6039	over | alphabet
0.172535	-12.604	pairs | by
0.0621879	-12.6043	can be | and
0.301075	-12.6044	evaluation | gold
0.516129	-12.6044	the | offering
0.1875	-12.6045	4.2 | we
0.325	-12.6048	the | obtained.
0.131846	-12.605	to | up
0.280374	-12.605	described | which
0.555556	-12.6051	research | center
0.705882	-12.6051	incorporates | into
0.9	-12.6056	found | dictionary,
0.103666	-12.6056	annotated | and
0.485714	-12.6056	which | semantically
0.705882	-12.6057	reranking | generative
0.385965	-12.6057	this | reasonable
0.372881	-12.6059	a | tag.
0.199074	-12.6062	the | operation
0.461538	-12.6062	rules | those
0.347826	-12.6064	moves | to
0.516129	-12.6064	a | white
0.333333	-12.6066	advantage | it
0.185484	-12.6067	for | data,
0.292929	-12.6068	maxent | the
0.139908	-12.6072	the | tasks.
0.29	-12.6072	by | linear
0.65	-12.6072	into | categories,
0.238095	-12.6072	in | especially
0.583333	-12.6073	dependency | (nivre
0.9	-12.6074	usually | usually
0.229814	-12.6075	from | languages
0.239726	-12.6078	an | value
0.37931	-12.6079	is | modern
0.287129	-12.608	phrase-based | system
0.19457	-12.6083	the | agent
0.619048	-12.6083	4-gram | model
0.9	-12.6084	statistical | (mt)
0.408163	-12.6084	for | learning,
0.255814	-12.6084	for | documents.
0.533333	-12.6084	in | spanish.
0.733333	-12.6085	for | tuning,
0.359375	-12.6086	they | often
0.9	-12.6088	theoretical | practical
0.253846	-12.6091	labeled | or
0.235294	-12.6093	links | in
0.833333	-12.6093	highest | achieved
0.461538	-12.6096	perform | better
0.306818	-12.6096	to | code
0.103535	-12.6096	seen | the
0.0897921	-12.6097	which | not
0.9	-12.6098	p(w|d)
0.516129	-12.6099	specific | domain
0.416667	-12.61	3: | 5:
0.44186	-12.6101	for example, | pair
0.9	-12.6102	types | subtypes
0.0581905	-12.6102	were | to
0.388889	-12.6103	used | testing
0.274336	-12.6104	realized | the
0.408163	-12.6105	that is, | an
0.65	-12.6108	are | for.
0.163009	-12.6108	presented | for
0.328947	-12.6109	a | learns
0.408163	-12.6109	excerpt | the
0.408163	-12.6109	comparison | other
0.705882	-12.6109	supported | fellowship
0.403846	-12.6109	(2010) | approach
0.12594	-12.611	of | categories
0.535714	-12.6113	but | ones
0.388889	-12.6113	and | parsers.
0.0988506	-12.6113	in | following
0.403846	-12.6115	introduces | in
0.833333	-12.6116	specificity | content
0.485714	-12.6116	also | algorithms
0.9	-12.612	distinct | types
0.416667	-12.612	that | input.
0.328947	-12.6121	system | able
0.9	-12.6123	stability | confidence
0.705882	-12.6123	2009 | 2009
0.130522	-12.6123	the | german
0.297872	-12.6126	the | performed.
0.140845	-12.6127	linguistic | in
0.174545	-12.6129	a | metric
0.197248	-12.613	detail | the
0.705882	-12.6131	able | distinguish
0.583333	-12.6132	after | further
0.320988	-12.6133	reducing | to
0.5	-12.6135	provides | about
0.0912659	-12.6136	text | a
0.583333	-12.6139	conclusion | material
0.833333	-12.6139	material | foundation.
0.171329	-12.614	the | power
0.325	-12.6142	we | developing
0.301075	-12.6144	cues | to
0.236486	-12.6145	percentage | that
0.262295	-12.6146	the | interpretation.
0.583333	-12.6146	you | my
0.328947	-12.6149	it | already
0.408163	-12.6149	gi | empirical
0.45	-12.6149	our | transition
0.174545	-12.615	score | each
0.159159	-12.6152	of | matrix
0.44186	-12.6152	in | scheme,
0.294737	-12.6152	accuracy | different
0.226994	-12.6152	see that | in
0.090998	-12.6154	we | some
0.320988	-12.6155	agent | to
0.269565	-12.6155	selecting | for
0.485714	-12.6156	corpora | available
0.516129	-12.6157	various | methods
0.0848329	-12.6158	the | corpora
0.619048	-12.6158	merge | into
0.320988	-12.6159	of | learners
0.0734355	-12.6159	the | within
0.555556	-12.616	analysis | content
0.313953	-12.616	all | annotated
0.583333	-12.6162	for | test,
0.583333	-12.6162	selecting | selecting
0.833333	-12.6162	word-based | model.
0.833333	-12.6165	sentences | sentences)
0.461538	-12.6167	construct | each
0.236486	-12.6168	of | (table
0.416667	-12.6169	using | above
0.535714	-12.617	journal of | and
0.705882	-12.617	michael | michael
0.216667	-12.6172	use | lexical
0.9	-12.6173	putting | into
0.9	-12.6173	tasks | domains.
0.162939	-12.6173	in | future
0.277778	-12.6176	humans | to
0.388889	-12.6177	6.3 | and
0.171329	-12.6177	the | completely
0.258065	-12.6177	is | (i.e.
0.396226	-12.6178	containing | words
0.112275	-12.6178	has | be
0.833333	-12.6178	indicate | arguments
0.159639	-12.6178	we then | and
0.193694	-12.6179	as | case
0.219653	-12.6179	conference | 27–31,
0.0707071	-12.618	which | are
0.65	-12.6182	study | effect
0.150943	-12.6183	limited | of
0.388889	-12.6184	2006; | 2007).
0.0757162	-12.6186	the | through
0.583333	-12.6187	because | require
0.160991	-12.6188	that | user
0.153846	-12.6191	training | model
0.272727	-12.6192	future work | this
0.306818	-12.6192	brief | of
0.328947	-12.6194	features | selected
0.396226	-12.6195	to | complementary
0.45	-12.6195	native | english
0.769231	-12.6196	complexity | l2
0.535714	-12.6199	adaptation | adaptation
0.0998811	-12.62	further | of
0.337838	-12.62	we | bigrams
0.5	-12.6202	in | samples.
0.083612	-12.6203	in | task
0.619048	-12.6205	inference, | that
0.555556	-12.6208	appear | times
0.306818	-12.6213	and | summaries.
0.0894386	-12.6213	that | use
0.209424	-12.6214	is | relatively
0.833333	-12.6214	this | comprises
0.122523	-12.6216	linear | of
0.666667	-12.6217	10 | 17
0.185185	-12.6219	is | n-gram
0.151099	-12.622	the | 2004).
0.20197	-12.6222	the | files
0.269565	-12.6222	should | should
0.365079	-12.6223	developed | which
0.18595	-12.6224	closer | the
0.178707	-12.6224	the | detected
0.733333	-12.6229	use | language model.
0.275229	-12.623	compare | two
0.272727	-12.6231	for | retrieval
0.306818	-12.6231	the | contributes
0.20297	-12.6233	features | lexical
0.090999	-12.6233	probability | a
0.619048	-12.6234	factors | factor
0.317073	-12.6234	the | reciprocal
0.342857	-12.6236	models | latent
0.125468	-12.6237	consider | in
0.196262	-12.6237	the | type,
0.555556	-12.6239	where | ith
0.230769	-12.6239	of | approaches.
0.555556	-12.6241	build | classifier
0.485714	-12.6241	entity | entity.
0.101485	-12.6243	however, | are
0.619048	-12.6244	both | neither
0.187234	-12.6244	the | supporting
0.342857	-12.6244	it | easily
0.232258	-12.6245	the | lemmas
0.190476	-12.6245	can be | such
0.324675	-12.6245	from | national
0.139211	-12.6246	phrases | a
0.145408	-12.6246	corpus | which
0.769231	-12.6248	original | modified
0.133192	-12.6249	only | not
0.08829	-12.6249	an | or
0.333333	-12.625	we assign | to
0.189655	-12.6251	from | document
0.535714	-12.6254	statistical | linguistic
0.153409	-12.6254	approach | which
0.37931	-12.6255	wu | and
0.19457	-12.6256	to find | a
0.56	-12.6256	ne | ne
0.56	-12.6256	stanford | lbj
0.0572325	-12.6256	one | a
0.619048	-12.6259	parser | reordering
0.472222	-12.6259	a | constraint.
0.291667	-12.6259	of | patterns,
0.215909	-12.6261	translation | bleu
0.19457	-12.6263	by | through
0.110947	-12.6263	and | relations
0.253846	-12.6263	training, | for
0.5	-12.6263	more generally, | of
0.173913	-12.6264	is | directly
0.0860692	-12.6264	the | relevant
0.121265	-12.6265	in | documents
0.246377	-12.6265	had | with
0.2	-12.6265	tested | and
0.251908	-12.6267	all | baseline
0.310345	-12.6268	our | achieved
0.485714	-12.6268	demonstrate | performance
0.222892	-12.6269	selected | as
0.170732	-12.627	identified | and
0.267241	-12.627	a | in | sentence,
0.191964	-12.627	method | using
0.112805	-12.6271	the | al., 2009).
0.555556	-12.6272	systems, | different
0.160494	-12.6272	functions | and
0.583333	-12.6273	answers | question
0.37931	-12.6274	best | scores
0.833333	-12.6274	results | demonstrating
0.833333	-12.6274	23 | wsj
0.22561	-12.6274	of | parent
0.485714	-12.6275	classes | their
0.733333	-12.6276	gain | baseline
0.264957	-12.6277	tagset | the
0.253968	-12.6277	loss | is
0.516129	-12.6278	and | em.
0.347826	-12.6279	all | labeled
0.381818	-12.628	order | obtain
0.320988	-12.6281	annotating | of
0.146982	-12.6282	per | and
0.516129	-12.6284	so | them
0.0606327	-12.6286	the | scores
0.262712	-12.6288	being | or
0.347826	-12.6289	model | probabilistic
0.9	-12.629	design | implementation
0.9	-12.629	techniques | techniques.
0.0991736	-12.629	simple | of
0.328947	-12.6291	the | noise.
0.208333	-12.6291	using | data.
0.262712	-12.6292	values | each
0.251908	-12.6293	identity | the
0.45	-12.6293	optimizing | in
0.137931	-12.6295	3.2 | of
0.555556	-12.6296	experimental setup | data
0.291667	-12.6296	the | encountered
0.3	-12.6297	and | detection.
0.154728	-12.6297	while | it
0.20297	-12.6298	presents | for
0.485714	-12.6298	since | must
0.208333	-12.6299	to | previously
0.118443	-12.6299	the | estimate
0.5	-12.6299	4 experiments | on
0.431818	-12.63	two | were | were
0.769231	-12.63	girl | toys
0.324675	-12.63	entity | type
0.9	-12.6301	part | survey
0.472222	-12.6302	faced | a
0.381818	-12.6305	utterances | not
0.198113	-12.6307	improve | by
0.408163	-12.6308	hmm | with
0.134199	-12.6312	system | an
0.152113	-12.6313	below | the
0.56	-12.6314	is | expensive.
0.9	-12.6315	heuristics | identify
0.208333	-12.6315	and | remaining
0.408163	-12.6316	simplicity | the
0.28866	-12.6316	improvement | baseline
0.733333	-12.6317	speed | accuracy
0.163462	-12.6318	resulting | and
0.45	-12.6318	got | and
0.264957	-12.6319	all | words in
0.5	-12.6319	given | out
0.535714	-12.6321	provide | good
0.56	-12.6321	texts | several
0.381818	-12.6322	comment | in
0.56	-12.6323	each | group.
0.229299	-12.6324	these | some
0.0301969	-12.6325	the | and | are
0.396226	-12.6325	with | discussion
0.555556	-12.6327	i, | i
0.281553	-12.6327	successfully | in
0.256	-12.6329	for | parsers
0.272727	-12.6329	generalized | and
0.555556	-12.6329	training | required
0.0689655	-12.6329	the | of | pairs
0.5	-12.6331	able | learn
0.555556	-12.6332	however, | makes
0.241135	-12.6332	if | pair
0.138889	-12.6333	proceedings of | workshop
0.150273	-12.6333	new | with
0.833333	-12.6333	berlin | kay
0.177606	-12.6334	technique | of
0.431818	-12.6337	2011), | oregon,
0.16	-12.6338	attributes | the
0.165563	-12.6338	approach | an
0.320513	-12.6341	data set | from
0.306818	-12.6342	narrative | and
0.165563	-12.6342	tag | a
0.9	-12.6344	(r) | (f)
0.294737	-12.6344	the | retained
0.150685	-12.6347	system | using
0.9	-12.6348	variant | given
0.516129	-12.6348	system | answers
0.333333	-12.6348	from | via
0.167235	-12.6349	we | annotations
0.328947	-12.6349	containing | or
0.472222	-12.6349	essentially | that
0.105053	-12.6353	those | in
0.472222	-12.6353	removed | from
0.177606	-12.6354	and | 2011).
0.233333	-12.6355	effective | for
0.769231	-12.6362	go | process
0.619048	-12.6362	we | augmented-loss
0.372881	-12.6364	for | adapted
0.0868373	-12.6364	standard | and
0.313253	-12.6364	on, | the
0.20098	-12.6365	when | one
0.0808882	-12.6366	the | head
0.087156	-12.6368	the | parameters
0.472222	-12.6368	and | out.
0.172662	-12.6368	whether | are
0.9	-12.6369	8 | 1 .
0.215909	-12.637	our | rules
0.472222	-12.637	classifier | identify
0.3	-12.6372	employ | for
0.278846	-12.6372	similarly | and
0.253968	-12.6373	since | was
0.0822368	-12.6374	the | uses
0.165563	-12.6375	the | verb.
0.5	-12.6375	for | prediction.
0.170819	-12.6379	as | previous
0.769231	-12.638	gender | case
0.239437	-12.6381	of | iterations
0.353846	-12.6381	modeled | and
0.769231	-12.6383	requires | complex
0.275229	-12.6384	unfortunately, | of
0.0933192	-12.6385	we first | the
0.100737	-12.6385	annotations | the
0.56	-12.6386	extraction | extraction.
0.181452	-12.6387	4.4 | the
0.198068	-12.6387	of | hypotheses
0.317073	-12.6388	a | that contains
0.260504	-12.6388	only | while
0.121864	-12.639	properties | the
0.45	-12.6391	scores | computed
0.3	-12.6394	some | also
0.0705669	-12.6395	the | size
0.213483	-12.6395	the | participating
0.131635	-12.6395	we | features.
0.242857	-12.6396	to | 2002).
0.342857	-12.6397	language | linguistic
0.166667	-12.64	parses | the
0.211957	-12.64	factor | a
0.4	-12.6401	familiar | and
0.251969	-12.6403	system | uses
0.0127182	-12.6403	a | of
0.733333	-12.6405	under | program.
0.769231	-12.6405	coherent | candidates
0.439024	-12.6405	from | media
0.199029	-12.6405	and | 2000;
0.222892	-12.6407	the | noun.
0.18	-12.6409	previously | and
0.0899209	-12.6409	features | are
0.229299	-12.6409	most | features
0.360656	-12.641	were | available
0.4	-12.6411	by | speaker
0.256	-12.6412	use | rules
0.210811	-12.6413	that | difficult
0.366667	-12.6413	and | wikipedia.
0.284314	-12.6414	5 | experiments
0.296703	-12.6417	using | previous
0.278846	-12.6418	identify | which
0.590909	-12.6419	0.7 | cost
0.485714	-12.6421	value | k
0.333333	-12.6423	and | agreement.
0.285714	-12.6425	8 | 10
0.0331029	-12.6425	the | at
0.222892	-12.6425	the | clark
0.28866	-12.6425	wi | a
0.294737	-12.6426	the | string,
0.408163	-12.6426	approximate | with
0.333333	-12.6426	linguistic | language
0.197115	-12.6427	of | texts.
0.181452	-12.6427	supported by | and
0.408163	-12.6429	complete | which
0.0897436	-12.6431	improvement | the
0.472222	-12.6431	sys | 0.62
0.209677	-12.6434	link | and
0.431818	-12.6436	contains | words.
0.3	-12.6438	similar | different
0.144703	-12.6439	individual | and
0.4	-12.644	(2006), | a
0.388889	-12.6441	for | alignment.
0.213483	-12.6444	with | labels
0.4	-12.6445	two | tasks
0.375	-12.6446	only | case
0.619048	-12.6447	textual | (dagan
0.6875	-12.6448	forward | hear
0.535714	-12.645	an | program
0.134956	-12.645	unlike | the
0.156627	-12.6451	et | 2006)
0.472222	-12.6452	nps | are
0.360656	-12.6454	the | operators.
0.372881	-12.6458	the | translated.
0.278846	-12.6458	for | biomedical
0.413043	-12.6458	once | each
0.413043	-12.6459	summaries | each
0.129897	-12.646	detection | and
0.166102	-12.6461	general | is
0.305882	-12.6462	research, | and
0.583333	-12.6465	we | scripts
0.439024	-12.6467	readers | a
0.158055	-12.6467	of | context.
0.439024	-12.6469	entire | set
0.517241	-12.6469	retrieve | that
0.590909	-12.6471	able | correctly
0.484848	-12.6473	evaluated | evaluation
0.353846	-12.6479	suitable | for
0.316456	-12.6484	edits | the
0.191781	-12.6484	with | previous
0.116086	-12.6485	the | expression
0.9	-12.6486	t(f|e)
0.217647	-12.6486	the | clean
0.25	-12.6486	yield | of
0.535714	-12.6487	search | beam
0.27619	-12.6487	from | collection
0.0783133	-12.6487	model | with
0.221557	-12.6488	1 introduction | has
0.56	-12.6488	topics, | topics
0.0978774	-12.6488	corpora | and
0.631579	-12.6489	corpus | summaries
0.101394	-12.649	combined | the
0.291667	-12.649	size | each
0.333333	-12.6492	i, | and
0.666667	-12.6493	analysis | analyze
0.278846	-12.6493	a | absolute
0.375	-12.6493	evaluation | methods
0.157576	-12.6494	that | (e.g.,
0.769231	-12.6495	using | estimation.
0.172161	-12.6495	given | each
0.14433	-12.6496	in | sequence
0.117057	-12.6496	measures | of
0.0880682	-12.6496	final | the
0.09375	-12.6497	therefore, | the
0.381818	-12.6498	theories | and
0.125483	-12.6499	3 | 3
0.422222	-12.6499	is | i-th
0.246269	-12.6499	final | with
0.209677	-12.6501	he | a
0.179283	-12.6502	all | these
0.278846	-12.6503	ten | and
0.258333	-12.6503	are | frequent
0.28866	-12.6504	the | 1995).
0.56	-12.6504	features | used,
0.381818	-12.6505	algorithm | any
0.375	-12.6505	polarity | words
0.619048	-12.6507	a | attribute.
0.6875	-12.6507	visual | image
0.631579	-12.6507	university | faculty
0.4	-12.6508	displays | of
0.45	-12.6509	over | space
0.119929	-12.651	the | in | way
0.10963	-12.651	this | will
0.1841	-12.6511	the | decisions
0.0203256	-12.6511	the | of | for
0.256198	-12.6513	simply | as
0.517241	-12.6513	although | less
0.226415	-12.6514	files | the
0.366667	-12.6516	clustering | each
0.28866	-12.6516	large | corpus
0.353846	-12.6517	little | on
0.733333	-12.6518	induce | using
0.205263	-12.6518	much | on
0.413043	-12.6518	each | score.
0.333333	-12.6518	notice that | to
0.4	-12.6519	careful | of
0.134831	-12.6519	constraint | the
0.147453	-12.652	attribute | the
0.146277	-12.6521	of | sentence,
0.583333	-12.6521	it | fast
0.631579	-12.6523	scisumm | a
0.0891283	-12.6525	the | text.
0.6875	-12.6526	definitions | definitions
0.116086	-12.6527	(2010) | of
0.15407	-12.6528	finding | of
0.265487	-12.6529	present | approach
0.5	-12.6529	work | ours
0.535714	-12.653	in | each.
0.818182	-12.6531	acknowledgement | supported
0.439024	-12.6531	a | found.
0.167832	-12.6532	table 4: | the
0.6875	-12.6532	that | of.
0.769231	-12.6533	but | expense
0.392157	-12.6533	of | meanings.
0.472222	-12.6534	computed | score
0.253968	-12.6536	a | framework.
0.619048	-12.6538	german | french
0.324675	-12.6538	of | terms,
0.769231	-12.6539	gains | points
0.28866	-12.6539	a | x,
0.459459	-12.654	4: | 6:
0.348485	-12.6541	6 experiments | the
0.0584771	-12.6541	scores | the
0.392157	-12.6541	that | constant
0.343284	-12.6541	results | corpora
0.142857	-12.6541	of | nouns
0.180723	-12.6542	this | available
0.360656	-12.6542	the | synset.
0.619048	-12.6543	summary | generated
0.413043	-12.6543	in | plural
0.631579	-12.6544	these | encode
0.631579	-12.6544	by | person.
0.251969	-12.6545	goal | is | a
0.124283	-12.6546	having | the
0.666667	-12.6547	models | capturing
0.139423	-12.6547	where | or
0.320513	-12.655	researchers have | to
0.375	-12.6551	conference | pp.
0.484848	-12.6552	surprising | is
0.422222	-12.6552	term | document
0.121771	-12.6555	if | word
0.142145	-12.6556	dependency | the | dependency
0.338028	-12.6557	function | defined
0.074352	-12.6557	if | is | is
0.535714	-12.6559	lower | lower
0.170909	-12.6559	the | adjectives
0.484848	-12.6561	2007; | 2009),
0.222222	-12.6562	tweets | a
0.205263	-12.6566	of | segment
0.619048	-12.6566	generalized | model
0.333333	-12.6566	now, | of
0.27619	-12.6568	inconsistent | the
0.176923	-12.6568	typically | and
0.535714	-12.6572	distribution | labels
0.28866	-12.6574	as | concept
0.733333	-12.6575	relevant | current
0.135945	-12.6575	parts | the
0.199005	-12.6575	3.2 | we
0.439024	-12.6576	as | cover
0.263158	-12.6577	movie | and
0.157895	-12.6577	bilingual | in
0.590909	-12.658	parser | correct
0.1425	-12.658	the | entry
0.818182	-12.6581	representation | 2010),
0.428571	-12.6581	when | full
0.619048	-12.6583	5.3 | evaluation
0.296703	-12.6583	of | graphs.
0.422222	-12.6583	it | sufficient
0.253968	-12.6583	boundary | and
0.472222	-12.6586	probability | computed
0.248062	-12.6587	to | performance,
0.818182	-12.6588	morphologically | languages.
0.278846	-12.6588	boundary | a
0.27027	-12.6588	on | mt
0.484848	-12.6589	these | actually
0.343284	-12.659	arguments | same
0.631579	-12.659	computational | learning:share
0.413043	-12.6591	a | landmark
0.091954	-12.6592	a | document
0.305882	-12.6594	of | node,
0.338028	-12.6595	million | from
0.135135	-12.6596	there | between | and
0.590909	-12.6599	distortion | translation
0.590909	-12.6599	how | create
0.290323	-12.66	complementary | of
0.309524	-12.6601	the | links.
0.459459	-12.6602	we | longest
0.56	-12.6602	findings | our
0.101542	-12.6604	figure | shows | the
0.733333	-12.6605	(brown | 1992)
0.333333	-12.6605	we consider | be
0.0962199	-12.6606	a | b
0.9	-12.6607	compounds | light
0.9	-12.6607	european | no.
0.324324	-12.6609	describe | system
0.590909	-12.6611	for instance, | example
0.196172	-12.6613	rich | of
0.439024	-12.6614	operates | a
0.328767	-12.6614	a | notion
0.28	-12.6615	that | syntax
0.472222	-12.6617	nouns | semantic
0.9	-12.6618	(knight | lapata,
0.9	-12.6618	sent’s | len
0.9	-12.6618	common | (lcs)
0.9	-12.6618	stanford, | 94305
0.258333	-12.6618	of | content.
0.20202	-12.662	these | then
0.28866	-12.6623	of | clusters.
0.17029	-12.6624	however, | have
0.159236	-12.6624	is | every
0.619048	-12.6624	is | sentiment.
0.535714	-12.6625	but | limited
0.733333	-12.6626	many | exist
0.244444	-12.6626	evaluate | by
0.818182	-12.6628	give | brief
0.818182	-12.6631	their | frequencies.
0.227273	-12.6631	though | are
0.328767	-12.6631	correctly | by
0.328767	-12.6634	frequency | each
0.818182	-12.6635	cosine | euclidean
0.338235	-12.6635	0.5 | 0.8
0.187773	-12.6635	and | structures
0.404255	-12.6638	there | strong
0.167832	-12.6639	in | range
0.25	-12.664	dialect | the
0.230263	-12.664	generates | of
0.0805802	-12.6644	present | of
0.517241	-12.6645	which | never
0.124756	-12.6647	is | important
0.348485	-12.665	get | from
0.472222	-12.665	of | needed.
0.360656	-12.6651	of | normally
0.222222	-12.6653	we | construct
0.324324	-12.6653	users | more
0.189189	-12.6654	corpus | sentences
0.224359	-12.6655	a | database
0.271028	-12.6655	with | above
0.733333	-12.6656	this | confirmed
0.666667	-12.6656	chinese | english.
0.172932	-12.6657	models | using
0.149171	-12.6657	that | structure
0.484848	-12.6657	research | processing
0.0513743	-12.6659	different | of
0.0337332	-12.6659	the | used | the
0.343284	-12.6661	word | part-of-speech
0.538462	-12.6661	14 | 15
0.212644	-12.6662	person | a
0.17029	-12.6663	of | spoken
0.208791	-12.6663	if | were
0.170909	-12.6663	effective | of
0.159236	-12.6666	arc | the
0.225806	-12.6667	of | rare
0.217647	-12.6667	it has been | the
0.0465805	-12.6669	are | with
0.484848	-12.6669	classifier | label
0.271028	-12.6669	which | particular
0.343284	-12.667	evaluation | accuracy
0.26087	-12.6671	in | graph.
0.484848	-12.6673	used | name
0.268519	-12.6673	and | formal
0.5	-12.6673	have | greater
0.199005	-12.6675	modification | the
0.360656	-12.6675	on | sets,
0.666667	-12.6676	nouns | proper
0.517241	-12.6677	1 introduction | become
0.338235	-12.6678	for | 23
0.124031	-12.6678	although | a
0.818182	-12.6679	unlike | approach,
0.375	-12.6682	the | anthology
0.154303	-12.6682	the | symbol
0.12549	-12.6684	a | like
0.6875	-12.6685	accuracy | identifying
0.174905	-12.6685	often | as
0.5	-12.6687	system | translate
0.422222	-12.6689	presented | new
0.320513	-12.6691	we | bias
0.381818	-12.6692	summationdisplay | (4)
0.0654155	-12.6693	and | were
0.0597681	-12.6693	/ | /
0.0918691	-12.6695	3 | for
0.277228	-12.6695	dt | dt
0.535714	-12.6698	10000 | 1e+06
0.818182	-12.6699	wtm | book
0.439024	-12.6699	algorithm | multiple
0.439024	-12.6699	figure 1: | sentence
0.110769	-12.6699	this | into
0.413043	-12.67	lexicon | used
0.23913	-12.6701	for | task, | of
0.324324	-12.6703	the | objects.
0.302326	-12.6704	that | reliable
0.769231	-12.6707	further | impact
0.258333	-12.6707	is | property
0.120438	-12.6707	the | procedure
0.0906282	-12.6709	so | to
0.19802	-12.6711	the | articles.
0.268519	-12.6712	the | ignoring
0.590909	-12.6714	without | accuracy
0.305882	-12.6715	the | segment.
0.179592	-12.6716	that | cannot
0.818182	-12.6716	words | men.
0.137767	-12.6716	a | another
0.5	-12.6716	(2007) | et al.
0.193396	-12.6717	furthermore, | in
0.0801925	-12.6718	were | for
0.113269	-12.6721	their | are
0.113452	-12.6724	items | the
0.769231	-12.6726	each | receives
0.769231	-12.6726	ideally, | should
0.0881489	-12.6728	first | for
0.4	-12.673	government | and
0.227273	-12.6731	of | articles.
0.218182	-12.6731	in | online
0.2	-12.6732	the | annotators.
0.216867	-12.6734	may be | that
0.103914	-12.6734	for | both | and
0.375	-12.6734	model | models,
0.631579	-12.6734	introduce | variables
0.287234	-12.6735	a | adjacent
0.631579	-12.6736	problem | identifying
0.205405	-12.6738	also | when
0.769231	-12.6739	trigger | arguments.
0.404255	-12.674	3 | 4.
1	-12.6742	mbdp-1 | ngs-u
1	-12.6742	gravano | hirschberg
1	-12.6742	bp | bf
1	-12.6742	conditional | fields.
1	-12.6742	(paskin, | manning,
1	-12.6742	textinfer | 2011,
1	-12.6742	x-axis | y-axis
1	-12.6742	roukos, | ward,
1	-12.6742	feature: | boolean
1	-12.6742	oovs | diff.
1	-12.6742	(freund | schapire,
1	-12.6742	huang | sagae,
1	-12.6742	[arg0 | [arg1
1	-12.6742	(chiao | zweigenbaum,
1	-12.6742	zollmann | venugopal,
1	-12.6742	(agirre | soroa,
1	-12.6742	rl | sim.user:
0.127572	-12.6743	a | measure
0.225806	-12.6743	the | noun,
0.517241	-12.6744	approach | clustering
0.126761	-12.6746	their | with
0.888889	-12.6746	meting | asociation
0.235714	-12.6746	may | some
0.439024	-12.6747	10: | for
0.098063	-12.6747	the | useful
0.375	-12.6748	instances | relation
0.0682614	-12.6749	for | data
0.5	-12.675	propose | automatically
0.888889	-12.6753	a., | s.
0.404255	-12.6753	computational linguistics | processing
0.26087	-12.6755	large-scale | and
0.252033	-12.6756	parser | dependency
0.888889	-12.6757	experts | expert
0.0657609	-12.6757	the | obtained
0.185841	-12.6759	address | in
0.32	-12.676	research | at
0.538462	-12.676	from | web,
0.470588	-12.6762	of | twitter.
0.19802	-12.6763	lexical | features
0.0661886	-12.6763	our | is | to
0.631579	-12.6763	parsing | collins,
0.154545	-12.6764	we now | the
0.298851	-12.6765	test, | the
0.185841	-12.6766	specifically | the
0.201031	-12.6767	(2007) | a
0.888889	-12.6768	and | wiebe.
0.202073	-12.6769	there is | on
0.28	-12.677	in | table.
0.439024	-12.6772	biomedical | extraction
0.126506	-12.6773	the | choice
0.538462	-12.6773	all | outperform
0.256198	-12.6774	status | the
0.538462	-12.6775	keep | from
0.470588	-12.6776	five | two
0.328767	-12.6776	the | origin
0.22973	-12.6777	3.2 | for
0.268519	-12.6777	as | ranking
0.888889	-12.6779	among | things,
0.888889	-12.6779	(blei | lafferty,
0.277228	-12.6779	aspect | is
0.368421	-12.678	k. | and
0.375	-12.6781	learn | words
0.631579	-12.6781	parsers | trained
0.439024	-12.6782	of | headword
0.888889	-12.6783	never | before.
0.25	-12.6783	and | little
0.293478	-12.6784	t, | and
0.0947012	-12.6784	e | e
0.590909	-12.6785	many | problems
0.404255	-12.6785	make | about
0.23741	-12.6786	the | newspaper
0.139024	-12.6786	during | and
0.362069	-12.6786	8 | .
0.538462	-12.6787	designed | information
0.266055	-12.6788	criteria | and
0.818182	-12.6788	game | action
0.138015	-12.6789	tagger | the
0.56	-12.679	light | light
0.219512	-12.679	is | weighted
0.0985222	-12.6793	for | english
0.459459	-12.6793	technique | used
0.145946	-12.6793	mean | and
0.0949657	-12.6793	the | matrix
0.248062	-12.6794	marked | in
0.888889	-12.6794	free | free
0.0700993	-12.6796	structure | the
0.13835	-12.6797	meaning | in
0.459459	-12.6797	is | sharing
0.631579	-12.6797	due | constraints,
0.484848	-12.6798	average | ratings
0.183761	-12.68	a | rate
0.168459	-12.6804	other | have
0.375	-12.6804	we | off
0.517241	-12.6804	200 | each
0.631579	-12.6805	scores: | and
0.404255	-12.6806	conceptual | in
0.137681	-12.6807	clear | the
0.175781	-12.6807	are | (e.g.,
0.19802	-12.6808	system | its
0.888889	-12.6809	observation | observation
0.246154	-12.681	of | length.
0.590909	-12.6812	build | corpus.
0.5	-12.6812	keyword | keyword
0.2	-12.6813	result | with
0.131868	-12.6814	often | a
0.158228	-12.6814	if | only
0.362069	-12.6815	to | purely
0.0578059	-12.6815	system | a
0.126	-12.6816	in | speech
0.22293	-12.6816	relation | between
0.273585	-12.6817	we | 1000
0.302326	-12.6818	when | accuracy
0.368421	-12.6818	our | produces
0.368421	-12.6818	of | identification.
0.769231	-12.6819	what | think
0.293478	-12.6821	framework | which
0.0658617	-12.6822	the second | the
0.888889	-12.6824	7: | 15:
0.404255	-12.6824	the | entailment.
0.517241	-12.6824	need | need
0.32	-12.6825	number | unique
0.219512	-12.6825	are | useful
0.181435	-12.6826	of | methods.
0.181435	-12.6826	the | available.
0.28	-12.6826	ordered | of
0.210227	-12.6826	the | summaries.
0.295455	-12.6828	tuple | a
0.1133	-12.6829	the | distributions
0.185841	-12.683	a | grammatical
0.0425337	-12.683	information | the
0.316456	-12.6831	transitive | of
0.888889	-12.6831	graphical | ngram
0.631579	-12.6831	tokyo | japan
0.178862	-12.6831	figures | the
0.769231	-12.6832	with | emphasis on
0.362069	-12.6832	 | 
0.157233	-12.6832	in | framework
0.538462	-12.6832	described | section,
0.470588	-12.6833	as | moves
0.15625	-12.6834	a | collection
0.125	-12.6834	classifier | a
0.268519	-12.6835	cases | or
0.6875	-12.6837	tilburg | netherlands
0.0711538	-12.6838	will | in
0.470588	-12.6839	parallel | aligned
0.447368	-12.6839	can | detect
0.067876	-12.684	other | in
0.148459	-12.684	and | systems.
0.266055	-12.6841	a | game
0.25641	-12.6841	analyze | in
0.0824034	-12.6841	the | term
0.219512	-12.6843	3 | two
0.590909	-12.6845	in | file,
0.6875	-12.6845	patterns | text.
0.20904	-12.6846	involves | of
0.144	-12.6846	in | graph
0.154545	-12.6847	including | in
0.484848	-12.6847	setting | all
0.404255	-12.6848	in: | and
0.888889	-12.685	recognizing | challenge
0.631579	-12.685	user | current
0.112721	-12.685	the | taken
0.19898	-12.6851	for | text.
0.167857	-12.6852	define | =
0.206522	-12.6852	the | acceptable
0.25	-12.6853	arbitrary | of
0.888889	-12.6858	tailored | specific
0.214286	-12.6858	of | variation
0.470588	-12.686	a | passage
0.343284	-12.686	select | set
0.150725	-12.6861	for | output
0.302326	-12.6862	therefore, | can
0.428571	-12.6862	lin, | and
0.128964	-12.6862	the | (and
0.242424	-12.6863	the | tagset
0.284211	-12.6864	cause | in
0.164912	-12.6865	these | into
0.631579	-12.6866	units | words
0.418605	-12.6867	however, | additional
0.34375	-12.6868	keep | as
0.0798696	-12.6869	for | was
0.392157	-12.687	values | feature
0.404255	-12.687	main | its
0.377358	-12.6872	used | construct
0.375	-12.6872	acknowledgements | supported
0.413043	-12.6873	for | problems.
0.362069	-12.6873	for | and/or
0.0751073	-12.6874	the | model,
0.266055	-12.6876	is | distribution over
0.377358	-12.6877	this | yields
0.395833	-12.6877	all other | are
0.226667	-12.6877	to | classification.
0.6875	-12.6877	process | automatically
0.5	-12.6878	edit | edit
0.470588	-12.688	each | topic.
0.145946	-12.688	techniques | to
0.5	-12.688	en | de
0.631579	-12.6882	etc. | also
0.565217	-12.6882	relations | e.g.
0.173077	-12.6883	process, | the
0.271028	-12.6883	tools | for
0.418605	-12.6884	for | randomly selected
0.375	-12.6885	also | linguistic
0.244275	-12.6887	clearly | in
0.194175	-12.6887	preliminary | of
0.22293	-12.6888	cues | and
0.0827526	-12.6889	the | tag
0.888889	-12.6891	objects | scene
0.308642	-12.6892	words | dictionary
0.590909	-12.6897	our | (which
0.384615	-12.6897	and | group.
0.888889	-12.6899	scenario | scenario
0.176707	-12.69	the | returns
0.0594595	-12.6901	we | use
0.103734	-12.6901	the | pattern
0.377358	-12.6902	the | unchanged.
0.377358	-12.6902	second | second
0.377358	-12.6902	predicted | be
0.196078	-12.6902	many | such
0.349206	-12.6903	lexicon | word
0.298851	-12.6903	independent | for
0.22973	-12.6904	in | training | test
0.349206	-12.6906	six | for
0.428571	-12.6906	compute | compute
0.0343637	-12.6906	is | with
0.565217	-12.6908	decreases | by
0.5	-12.6909	a | connects
0.22973	-12.691	for | support
0.459459	-12.6911	“play | role
0.178138	-12.6912	this | simple
0.268519	-12.6912	certain | are
0.484848	-12.6913	relation | pairs
0.074928	-12.6914	that | only
0.308642	-12.6917	of | 50%
0.315789	-12.6917	and | spans
0.254237	-12.6917	these | very
0.20904	-12.6919	learn | for
0.769231	-12.692	author(s) | or
0.447368	-12.692	tool | which
0.447368	-12.692	with | f-measure
0.277228	-12.692	all | corresponding
0.484848	-12.6921	trained | sections
0.292135	-12.6921	i.e., | that
0.459459	-12.6924	76 | and
0.714286	-12.6925	tratz | hovy
0.5	-12.6926	pairs | wikipedia
0.115059	-12.6927	an | using
0.368421	-12.693	if | certain
0.631579	-12.693	f-measure | when
0.290323	-12.6931	to | (cf.
0.349206	-12.6931	features | values
0.126273	-12.6931	part | a
0.206704	-12.6935	in | derivation
0.2	-12.6935	that | improved
0.25641	-12.6936	note that | this
0.184211	-12.6937	always | a
0.263636	-12.6937	of | challenges
0.631579	-12.6938	collection | texts
0.631579	-12.6938	corpus | abstracts
0.315789	-12.694	5.2 | with
0.392157	-12.694	··· | ···
0.202128	-12.6941	4.2 | for
0.284211	-12.6941	if | >
0.138614	-12.6942	for | ∈
0.151786	-12.6944	the | message
0.565217	-12.6944	predicates | argument
0.131696	-12.6945	new | that
0.19697	-12.6945	often | with
0.631579	-12.6946	first | arabic
0.163265	-12.6947	the | examples.
0.404255	-12.6947	automatically | parallel
0.368421	-12.6947	in | semeval
0.888889	-12.6948	k., | &
0.484848	-12.6948	extract | relations
0.214286	-12.6949	since | data
0.818182	-12.695	match | against
0.590909	-12.695	all | matched
0.246032	-12.695	has | many
0.125253	-12.695	to | sentences.
0.368421	-12.6951	2009; | 2010)
0.230769	-12.6951	features | baseline
0.211429	-12.6953	the | judged
0.213018	-12.6953	information | used
0.206704	-12.6954	his | and
0.714286	-12.6959	0.64 | 0.68
0.818182	-12.696	prevented | from
0.171103	-12.696	in | section.
0.221519	-12.696	the | wordnet.
0.135714	-12.6961	this | how
0.3125	-12.6962	indicator | for
0.188073	-12.6963	the | nearly
0.377358	-12.6965	but | improvement
0.590909	-12.6965	tree | (collins
0.888889	-12.6967	• | y)
0.246032	-12.6972	the | variability
0.287234	-12.6973	the | pruned
0.395833	-12.6973	to | hypothesis,
0.133949	-12.6973	the | split
0.220126	-12.6975	we | optimization
0.328571	-12.6976	is | inside
0.133333	-12.6977	mt | and
0.355932	-12.6977	uses | same
0.517241	-12.6977	are | 3).
0.470588	-12.6978	drops | to
0.714286	-12.698	questions | answers.
0.459459	-12.6981	← | ←
0.888889	-12.6982	input, | outputs
0.328571	-12.6982	such | question
0.240602	-12.6982	results obtained | the
0.377358	-12.6983	we | only those
0.152695	-12.6983	we | learn
0.261261	-12.6984	a | compression
0.818182	-12.6985	vector | vector,
0.354839	-12.6985	only | per
0.211765	-12.6986	around | of
0.191388	-12.6987	that | times
0.244094	-12.6987	in | stage
0.186364	-12.6988	the | ten
0.888889	-12.699	assumption | hold
0.5	-12.699	school | edinburgh
0.517241	-12.6992	many | specific
0.173077	-12.6993	that, | a
0.20904	-12.6993	efforts | the
0.261261	-12.6993	2.2 | for
0.06469	-12.6994	the first | of
0.153374	-12.6995	model | over
0.225166	-12.6996	detect | in
0.248	-12.6998	annotate | of
0.298851	-12.6999	colour | and
0.428571	-12.7001	word | occurs
0.888889	-12.7005	p(s|e)
0.418605	-12.7005	2001), | et
0.159468	-12.7006	uses | for
0.258621	-12.7007	rule | an
0.362069	-12.7007	all | had
0.190476	-12.7007	in this paper, we | on
0.176707	-12.7008	explore | to
0.210227	-12.701	model | alignment
0.368421	-12.701	and | recognition.
0.19598	-12.7011	we | sampling
0.202128	-12.7012	department | computer | university
0.333333	-12.7012	li | al.,
0.183406	-12.7015	asked to | the
0.538462	-12.7015	when | too
0.151335	-12.7016	problem | that
0.166065	-12.7016	to | end
0.888889	-12.7016	laboratory | cambridge
0.714286	-12.7017	aspects | consumers’
0.284211	-12.7017	but | much
0.0838651	-12.7018	the | probabilities
0.18894	-12.7019	module | and
0.05078	-12.702	the | the | data.
0.171103	-12.7021	train | on
0.565217	-12.7021	across | sets.
0.631579	-12.7022	2 | designed
0.447368	-12.7023	and | worked
0.6875	-12.7023	is | straightforward.
0.355932	-12.7023	full | set
0.136038	-12.7023	from | into
0.435897	-12.7023	where | learned
0.295455	-12.7025	two | respectively.
0.0586667	-12.7026	rules | the
0.0956316	-12.7027	this work | the
0.888889	-12.7028	were asked | rate
0.349206	-12.7028	in | flat
0.565217	-12.7028	lack | information
0.428571	-12.703	evaluate | proposed
0.328571	-12.703	yi | the
0.277228	-12.703	we | examine
0.25641	-12.7031	of | 2000).
0.5	-12.7031	corpus | tagged
0.288889	-12.7032	and | ...
0.181818	-12.7034	is | cluster
0.0895833	-12.7035	the | above
0.244094	-12.7035	a | filter
0.590909	-12.7036	estimation | parameters
0.152439	-12.7037	model | trained
0.301205	-12.7039	example | given
0.238806	-12.7043	mt | with
0.205556	-12.7043	the | giving
0.100654	-12.7044	the | base
0.160535	-12.7044	of | elements
0.769231	-12.7045	focuses on | between
0.134118	-12.7045	therefore, | to
0.150888	-12.7046	and | results.
0.6	-12.7046	is | vital
0.163763	-12.7046	the | context,
0.248	-12.7049	for | application
0.115318	-12.705	of | individual
0.131818	-12.705	across | and
0.5	-12.705	variants | or
0.6	-12.7051	work | closely related
0.6	-12.7051	task, | given
0.32	-12.7053	the | development.
0.107303	-12.7054	of | argument
0.133333	-12.7055	we | general
0.201058	-12.7056	have | over
0.223684	-12.7056	verb | noun
0.220126	-12.7058	in | texts.
0.278351	-12.7058	the | phase.
0.590909	-12.7058	see | further
0.409091	-12.706	examining | that
0.888889	-12.7062	models | compact
0.192308	-12.7064	is | final
0.0958084	-12.7064	the | lexicon
0.338462	-12.7064	this | actual
0.6875	-12.7065	has | attention.
0.147309	-12.7065	the | translation,
0.138272	-12.7065	for example, | are
0.457143	-12.7066	algorithms, | in
0.323944	-12.7067	et al., 2008) | a
0.180258	-12.7068	that | local
0.192308	-12.7069	the | length.
0.11883	-12.707	utterance | the
0.261261	-12.707	in | early
0.565217	-12.707	(riedel | al.,
0.368421	-12.7071	and | 90%
0.368421	-12.7071	evaluates | of
0.222222	-12.7074	similarities | the
0.269231	-12.7074	by taking | of
0.362069	-12.7074	study | different
0.483871	-12.7074	when | utterance
0.447368	-12.7077	army | under
0.0901379	-12.7077	of | (e.g.,
0.565217	-12.7077	identify | relevant
0.201058	-12.7081	strategies | and
0.769231	-12.7081	difficult | current
0.244094	-12.7082	use | following
0.404255	-12.7082	is | low.
0.301205	-12.7084	on | corpora,
0.238806	-12.7085	of | phrases,
0.387755	-12.7086	another | was
0.714286	-12.7087	limit | maximum
0.631579	-12.7087	but | worse
0.125	-12.7088	web | and
0.105036	-12.7089	of | through
0.447368	-12.7089	coverage | an
0.6	-12.709	in | tables,
0.301205	-12.7091	recent | have
0.124498	-12.7091	it | other
0.6875	-12.7091	coverage | precision
0.328571	-12.7091	that | become
0.111111	-12.7092	a | .
0.242188	-12.7092	of | dialogues
0.338462	-12.7093	argument | role
0.081266	-12.7095	the | graph
0.888889	-12.7096	rich | morphological
0.328571	-12.7096	results | individual
0.457143	-12.7097	and | resnik,
0.144414	-12.7097	we | extract
0.102881	-12.7097	and | features.
0.261261	-12.7097	too | for
0.404255	-12.7098	of | combinations.
0.278351	-12.7099	all | nodes
0.246032	-12.7099	the | position.
0.483871	-12.7099	model | uniform
0.404255	-12.71	a | realistic
0.288889	-12.71	analysis | results
0.12766	-12.7102	the | loss
0.384615	-12.7102	precision | f-score
0.404255	-12.7103	a | screen
0.6	-12.7103	5 conclusions | have
0.183406	-12.7105	statistics | in
0.214724	-12.7105	the | newswire
0.333333	-12.7105	vectors, | the
0.377358	-12.7106	segmentation | tagging
0.470588	-12.7106	4 | 19
0.102921	-12.7107	task, | the
0.518519	-12.7109	some | problems
0.631579	-12.7111	designing | that
0.483871	-12.7111	in | mode.
0.0529971	-12.7112	a | language
0.0849057	-12.7114	rules | and
0.483871	-12.7115	a | sds
0.304878	-12.7115	tasks | or
0.101198	-12.7117	n | n | n
0.5	-12.7117	words | punctuation
0.176955	-12.7118	3.4 | the
0.0878788	-12.7118	in this paper, | the
0.409091	-12.7118	only | local
0.235294	-12.7118	nouns | as
0.888889	-12.7119	constructs | graph
0.6	-12.7119	union | union
0.301205	-12.7122	retrieve | of
0.123752	-12.7122	reordering | and
0.483871	-12.7122	of | np.
0.25	-12.7123	free | and
0.248	-12.7123	and | 1998;
0.25	-12.7124	has | very
0.5	-12.7124	structural | speech
0.435897	-12.7128	they | appear
0.818182	-12.7129	google | api
0.328571	-12.713	for | versus
0.248	-12.713	of | (cf.
0.769231	-12.7131	german | swiss
0.818182	-12.7133	parameters | parameters.
0.447368	-12.7133	found that | better
0.538462	-12.7133	dev | devtest
0.565217	-12.7133	in | corpus;
0.769231	-12.7134	epsilon1 | δ
0.214724	-12.7134	for | entire
0.225166	-12.7134	will | different
0.217391	-12.7135	to | greater
0.315789	-12.7135	in | processing.
0.888889	-12.7135	steps | build
0.355932	-12.7135	employed | for
0.184685	-12.7136	this approach | and
0.229167	-12.7136	another | by
0.631579	-12.7136	are | (including
0.457143	-12.7137	input: | set
0.538462	-12.7138	the | realisations
0.107251	-12.7141	is | large
0.647059	-12.7141	after | iteration
0.288889	-12.7142	phrase | lexical
0.304878	-12.7143	is | case.
0.565217	-12.7143	for | students.
0.435897	-12.7143	next, | which
0.483871	-12.7145	a | scale,
0.404255	-12.7147	using | tagger
0.159468	-12.7147	of | (or
0.631579	-12.7147	paraphrases | monolingual
0.395833	-12.715	cannot | this
0.238806	-12.7151	choosing | of
0.19598	-12.7151	presented | that
0.447368	-12.7154	terminal | terminal
0.714286	-12.7155	understanding | human
0.631579	-12.7158	these | and,
0.377358	-12.716	for | scores,
0.271845	-12.7161	which | found
0.395833	-12.7164	event | entities
0.226027	-12.7164	the | acquired
0.111111	-12.7164	just | the
0.483871	-12.7164	5.2 | used
0.37037	-12.7165	early | on
0.246032	-12.7166	frequency | word
0.435897	-12.7166	for | correction.
0.266667	-12.7167	that | (and
0.214724	-12.7167	ratio | in
0.292135	-12.7167	technology | to
0.714286	-12.7168	we ran | over
0.271845	-12.7168	annotations | are
0.118959	-12.7169	further | to
0.219355	-12.717	acknowledgements | and
0.101902	-12.717	automatic | and
0.470588	-12.7171	the | has,
0.818182	-12.7172	utilizing | information
0.395833	-12.7174	matching | an
0.6875	-12.7175	somewhat | but
0.146893	-12.7176	and | during
0.202186	-12.7176	this | good
0.395833	-12.7177	given | generate
0.319444	-12.7177	considerable | in
0.12685	-12.7178	of | expressions
0.37037	-12.7179	from | root
0.292135	-12.718	(a | is
0.311688	-12.7183	furthermore, | can
0.244094	-12.7183	the | interval
0.0906149	-12.7183	it | an
0.229167	-12.7184	the | changing
0.5	-12.7184	, | ...
0.888889	-12.7185	do | job
0.435897	-12.7185	index | that
0.457143	-12.7186	(2005) | using
0.6	-12.7186	for | learners.
0.377358	-12.7189	all | i.e.
0.425	-12.7191	2011 | institute
0.246032	-12.7191	a | quality.
0.210526	-12.7191	which | similar
0.338462	-12.7193	a | consecutive
0.377358	-12.7197	3) | for
0.311688	-12.7198	we propose | approach
0.264151	-12.7198	overall, | and
0.242188	-12.7198	most | only
0.307692	-12.7199	model | local
0.258929	-12.7199	we | keep
0.647059	-12.72	is | ours,
0.647059	-12.72	(2005) | their
0.5	-12.72	two | systems,
0.244094	-12.7201	incorporate | in
0.0535041	-12.7202	the | of | corpus
0.470588	-12.7202	goal | was
0.818182	-12.7204	annotation | adjudication
0.319444	-12.7204	the | sections.
0.226027	-12.7204	and | extraction.
0.6	-12.7205	she | had
0.355932	-12.7206	known | or
0.338462	-12.7206	which, | of
0.292135	-12.7207	the | mode
0.173228	-12.7207	position | a
0.216049	-12.7207	despite | and
0.447368	-12.7209	levels | from
0.435897	-12.721	of | region
0.307692	-12.7212	nevertheless, | to
0.888889	-12.7212	below | above
0.183036	-12.7213	use | semantic
0.5	-12.7213	gives | best
0.101626	-12.7214	they are | the
0.377358	-12.7214	a | italian
0.818182	-12.7214	(e.g., | etc.),
0.139175	-12.7215	types of | in
0.301205	-12.7216	the | behavior.
0.174089	-12.7216	and | support
0.292135	-12.7216	a | pcfg
0.205556	-12.7217	but | was
0.470588	-12.7218	within | document
0.363636	-12.722	use | existing
0.647059	-12.722	graphs | classification
0.538462	-12.722	gibbs sampling | for
0.0977444	-12.7221	the | points
0.483871	-12.7221	highlighted | in
0.538462	-12.7222	strings | length
0.483871	-12.7223	you | me
0.304878	-12.7224	the | similar,
0.125786	-12.7226	item | the
0.198953	-12.7226	forms | in
0.6875	-12.7226	pairs | judged
0.214724	-12.7226	particularly | to
0.483871	-12.7227	lexical | prosodic
0.271845	-12.7229	networks | and
0.161512	-12.723	value | for
0.109005	-12.7232	and | research
0.447368	-12.7232	which | frequently
0.6875	-12.7232	every | assigned
0.127706	-12.7232	and | corpora
0.323944	-12.7234	result | using
0.377358	-12.7235	[<s> squaresmallsolid | squaresmallsolid
0.888889	-12.7235	robot | human
0.27551	-12.7235	we used | corpus
0.631579	-12.7235	used | texts.
0.0902256	-12.7235	question | the
0.307692	-12.7236	of | research,
0.457143	-12.7236	normalized | normalized
0.818182	-12.7236	but | reverse
0.0703583	-12.7238	over | and
0.222222	-12.7239	not | specific
0.0632578	-12.7239	sentence | a
0.483871	-12.724	tested | system
0.171875	-12.7242	denote | and
0.647059	-12.7242	finally, section | concludes
0.0987179	-12.7243	problem | to
0.6	-12.7244	opinion | product
0.377358	-12.7245	that | specified
0.301205	-12.7245	we | generic
0.193069	-12.7247	100 | for
0.565217	-12.7247	this | adequate
0.282609	-12.7248	was | about
0.219355	-12.7249	like | et al.,
0.714286	-12.7249	benefits | from
0.186047	-12.725	2000; | al.,
0.195876	-12.725	from | annotated
0.6875	-12.725	different | semantically
0.210526	-12.725	the | formula
0.219355	-12.7253	sparsity | the
0.355932	-12.7256	these | particular
0.28125	-12.7257	paragraph | the
0.25	-12.7258	the | small,
0.159864	-12.7258	as | baseline
0.238462	-12.7258	a | valid
0.888889	-12.7259	reliability | multiple
0.038674	-12.726	it | a
0.818182	-12.7261	from | figure,
0.818182	-12.7261	corpus | lemmatized
0.135593	-12.7261	we believe | the
0.24031	-12.7261	there are | this
0.11204	-12.7264	in | classification
0.245902	-12.7266	maximize | of
0.377358	-12.7267	the | generator
0.123232	-12.7268	vocabulary | the
0.338462	-12.7268	keeps | the
0.301205	-12.7269	a, | of
0.160959	-12.727	predicate | and
0.101648	-12.727	the | and | features.
0.230216	-12.727	even | than
0.25	-12.7272	the | value,
0.206897	-12.7273	some | may
0.185185	-12.7274	the | classifier,
0.244094	-12.7275	clusters, | the
0.153125	-12.7275	that | particular
0.261682	-12.7276	refers to | a
0.714286	-12.7277	2008 | 2009
0.151057	-12.7278	among | in
0.149701	-12.7278	when | it
0.157895	-12.7278	with | syntactic
0.418605	-12.7278	adding | feature
0.328571	-12.7279	a | jointly
0.213415	-12.7279	giza++ | and
0.565217	-12.7281	b, | b
0.395833	-12.7283	pre-processing | to
0.121569	-12.7283	4.3 | the
0.6	-12.7284	sense | word,
0.37037	-12.7285	apparent | the
0.212121	-12.7285	extra | the
0.146132	-12.7285	that | correct
0.136476	-12.7287	the last | of
0.483871	-12.7291	concept | its
0.0781759	-12.7292	all | with
0.210843	-12.7292	2011 | uk,
0.301205	-12.7293	to | problems.
0.35	-12.7293	a | satisfies
0.164286	-12.7293	edinburgh, | july
0.518519	-12.7294	for | ≈
0.0994764	-12.7294	abstract | a
0.409091	-12.7295	to | determiner
0.818182	-12.7297	multilingual | analysis
0.344262	-12.7299	where | correct
0.159322	-12.73	ordering | the
0.109524	-12.7303	and | 2006).
0.6875	-12.7304	,t | ,precedesequal
0.714286	-12.7308	phone | using
0.647059	-12.7308	referring | generation
0.130337	-12.7308	n-gram | and
0.172691	-12.731	criteria | the
0.25	-12.731	also | better
0.647059	-12.731	furthermore, | often
0.303797	-12.7311	the | large.
0.227586	-12.7311	to | his
0.109855	-12.7312	which | more
0.209302	-12.7313	i | where
0.27	-12.7313	because | information
0.483871	-12.7314	twitter | users
0.311688	-12.7317	secondly, | to
0.457143	-12.7317	an | event.
0.226027	-12.7317	an | state
0.294118	-12.7321	experiment | which
0.169884	-12.7323	np | vp
0.363636	-12.7325	going | to
0.176471	-12.7325	trigram | the
0.0814159	-12.7327	proposed | and
0.135802	-12.7328	the | preposition
0.5	-12.7329	was | entire
0.647059	-12.733	is | game.
0.444444	-12.7331	world | coreference
0.125523	-12.7332	considered | of
0.042589	-12.7332	are | as
0.409091	-12.7334	verbs | class
0.315068	-12.7336	move | a
0.179487	-12.7337	runs | the
0.171206	-12.7339	position | in
0.307692	-12.7342	and | separately
0.5	-12.7342	figure, | and
0.714286	-12.7343	fahrr¨ader | sind
0.297619	-12.7343	and | samples
0.167939	-12.7344	the | domains.
0.155844	-12.7344	of | rules.
0.1033	-12.7344	classes | the
0.290698	-12.7345	by | algorithm.
0.230216	-12.7346	perform | with
0.323944	-12.7347	dictionary | word
0.0741012	-12.7348	the | reference
0.344262	-12.7351	collecting | and
0.155844	-12.7351	the | upper
0.206897	-12.7352	hence | of
0.387755	-12.7354	instances, | and
0.172691	-12.7356	line | a
0.294118	-12.736	generic | for
0.172	-12.7361	experiments | using
0.0933489	-12.7362	tokens | the
0.187793	-12.7362	the | helps
0.219355	-12.7363	on | other hand, | is
0.128319	-12.7363	each | two
0.27957	-12.7364	matrix | with
0.236641	-12.7365	propose | model
0.37037	-12.7366	wordnet | words
0.196891	-12.7369	of | phrases.
0.333333	-12.7369	our | experiments.
0.173387	-12.7369	there | many
0.303797	-12.737	are | written
0.254386	-12.737	for | methods.
0.084121	-12.7371	in | 1.
0.226027	-12.7372	relation | two
0.252174	-12.7372	of | nominal
0.221477	-12.7372	et | 2002)
0.483871	-12.7376	histogram | in
0.124481	-12.7378	we also | that
0.153355	-12.7378	et | 2007)
0.165441	-12.738	metric | a
0.4	-12.7384	furthermore, | also
0.457143	-12.7385	english | native
0.363636	-12.7385	material | necessarily
0.647059	-12.7385	disambiguation | very
0.290698	-12.7386	dimensionality | the
0.205714	-12.7389	a | window
0.35	-12.7389	with | default
0.264151	-12.7389	(2010) | an
0.357143	-12.739	data | million
0.75	-12.7391	helpfulness | peer-review
0.46875	-12.7392	words | distinct
0.282609	-12.7392	te | a
0.210843	-12.7393	resulting | for
0.172	-12.7396	could | or
0.435897	-12.7396	auxiliary | auxiliary
0.75	-12.7397	reach | level
0.136364	-12.7397	the | segment
0.5	-12.74	division | of
0.818182	-12.7402	modules | through
0.282609	-12.7402	in | resources.
0.20339	-12.7403	with | additional
0.236641	-12.7404	table | presents
0.221477	-12.7406	an | form
0.565217	-12.7406	were | received
0.290698	-12.7407	latent | model
0.164835	-12.7408	the | namely
0.24031	-12.7408	that | required
0.112436	-12.7409	a | multiple
0.37037	-12.7412	and | aspects.
0.153355	-12.7412	that | across
0.206897	-12.7412	correct | by
0.541667	-12.7413	accordingly, | in
0.102564	-12.7414	to | knowledge
0.387755	-12.7414	describe | (section
0.162544	-12.7414	model | its
0.191176	-12.7415	dimensions | the
0.241935	-12.7416	specialized | the
0.25	-12.7416	did | any
0.444444	-12.7416	parallel | pairs.
0.565217	-12.7416	abstract | present | method
0.180617	-12.7417	the | arbitrary
0.267327	-12.7417	system’s | of
0.188406	-12.7419	all | use
0.267327	-12.7419	we assume | that
0.129545	-12.742	a | statistical
0.22695	-12.742	of | annotations.
0.35	-12.7422	set | binary
0.46875	-12.7423	even | even
0.147929	-12.7423	five | and
0.323944	-12.7424	diversity | in
0.0329206	-12.7424	and | et al.,
0.272727	-12.7426	of | participating
0.303797	-12.7426	there | ways
0.121272	-12.7426	classification | a
0.27	-12.7426	of | query.
0.319444	-12.7427	extended | by
0.1875	-12.7427	of | better than
0.75	-12.7427	age | gender
0.35	-12.7427	text | documents
0.425	-12.7428	in | representation,
0.541667	-12.7428	the | service.
0.387755	-12.743	1, | 1,
0.38	-12.743	in | causal
0.714286	-12.7432	age | group
0.222973	-12.7433	representing | in
0.0537383	-12.7433	the | about
0.75	-12.7434	computation | prefix
0.444444	-12.7437	i.e., | all
0.297619	-12.7437	a | determines
0.0800344	-12.7438	results | are
0.818182	-12.7439	constraints | ensure
0.565217	-12.7439	if | previously
0.328358	-12.7441	utilized | and
0.0652547	-12.7441	a | corpus
0.154839	-12.7441	new | are
0.27551	-12.7442	list | each
0.307692	-12.7443	those | other
0.46875	-12.7444	systems | those
0.243902	-12.7444	a | lattice
0.22	-12.7445	a | message
0.234848	-12.7446	the | functions.
0.75	-12.7448	c&c | (clark
0.267327	-12.7448	following | al.
0.4	-12.7448	of | important.
0.37037	-12.7449	to | automatically.
0.171315	-12.745	to | via
0.202247	-12.745	analysis, | and
0.177215	-12.745	are | available
0.387755	-12.7451	location, | and
0.245902	-12.7451	all | types
0.0739098	-12.7453	to | new
0.127473	-12.7456	there is | for
0.188406	-12.7458	summationdisplay | where
0.37037	-12.746	since | could
0.0608782	-12.746	the | part | the
0.372549	-12.746	patterns | semantic
0.6	-12.7462	hmm | models
0.414634	-12.7463	are | distributed
0.363636	-12.7463	estimate | using
0.207101	-12.7464	50% | the
0.172	-12.7465	by | based on
0.228571	-12.7465	approach | learning
0.233083	-12.7466	the | leaves
0.202247	-12.7466	in | sample
0.363636	-12.7466	calculated | all
0.178261	-12.7466	form | that
0.818182	-12.7468	where | constant.
0.319444	-12.7468	a | encoding
0.109656	-12.7468	a | words.
0.363636	-12.7469	the | schemes.
0.2	-12.747	variable | is
0.1171	-12.7471	taken | the
0.122951	-12.7472	the | called
0.27	-12.7473	of | lengths
0.2	-12.7473	the | scored
0.4	-12.7475	simply | all
0.5	-12.7476	approaches | 2008).
0.387755	-12.7477	using | corpus,
0.75	-12.7478	alternatively, | could
0.425	-12.7479	among | different
0.363636	-12.7479	is | verb,
0.444444	-12.7479	original | new
0.184332	-12.7481	models | data
0.0721868	-12.7483	generated | the
0.166038	-12.7483	the | reflect | of
0.290698	-12.7485	2.4 | of
0.518519	-12.7485	function | content
0.112436	-12.7487	this | when
0.144068	-12.7487	basic | and
0.130841	-12.7488	and | further
0.75	-12.7488	p(y|x)
0.218543	-12.7489	the | signal
0.137405	-12.7489	studies | and
0.207101	-12.749	word, | a
0.541667	-12.7491	only | context.
0.205714	-12.7492	the | feature,
0.245902	-12.7492	this | difficult
0.483871	-12.7493	we | hyperparameters
0.818182	-12.7494	occur | part
0.6	-12.7494	that | recovered
0.565217	-12.7495	provide | best
0.565217	-12.7495	many | errors
0.4	-12.7496	a | v,
0.16129	-12.7496	the | toolkit
0.261682	-12.7497	standard | using
0.121514	-12.7497	provide | to
0.272727	-12.7498	discussions | the
0.391304	-12.7501	different | approaches
0.818182	-12.7501	individual | within
0.435897	-12.7501	reducing | by
0.177489	-12.7502	using | models
0.518519	-12.7504	questions, | are
0.518519	-12.7504	something | or
0.387755	-12.7505	the | element.
0.33871	-12.7505	from | dirichlet
0.22695	-12.7505	until | is
0.483871	-12.7506	temporal | time
0.175732	-12.7507	although | on
0.303797	-12.7507	from | domain.
0.0149317	-12.7507	we | the | of
0.185714	-12.7507	the | lemma
0.435897	-12.7507	le | le
0.153355	-12.7508	the | behavior
0.571429	-12.7508	operation | which
0.119843	-12.7509	introduced | the
0.33871	-12.751	performance | evaluation
0.116236	-12.751	possible | for
0.37037	-12.7512	wn | wn
0.208333	-12.7512	the | kept
0.483871	-12.7512	hybrid | which
0.179825	-12.7513	summaries | and
0.136364	-12.7513	social | and
0.210843	-12.7513	the | threshold.
0.035876	-12.7514	number of | the
0.130536	-12.7515	which | one
0.310811	-12.7516	automatic | system
0.186603	-12.7516	the | possible.
0.344262	-12.7518	a | proposition
0.571429	-12.7519	† | ‡
0.483871	-12.7519	variants | which
0.185714	-12.7519	finite | of
0.17284	-12.7521	we | five
0.202247	-12.7522	threshold | to
0.10271	-12.7523	set | =
0.228571	-12.7524	structure, | of
0.225352	-12.7526	the | syntactically
0.0529524	-12.7527	as | or
0.319444	-12.7527	categories | or
0.287356	-12.7527	of | proposition
0.391304	-12.7528	introduces | for
0.319444	-12.7528	which | selected
0.132212	-12.7529	the | predictions
0.142077	-12.753	figure | figure
0.254386	-12.753	the | predictions.
0.409091	-12.753	acquisition | for
0.483871	-12.753	when | increase
0.213836	-12.753	et | 2005),
0.714286	-12.7531	set | frontier
0.241935	-12.7531	includes | that
0.10479	-12.7532	to | quality
0.483871	-12.7532	use | help
0.164794	-12.7533	weighted | and
0.231343	-12.7535	these | lexical
0.259259	-12.7536	a | normal
0.310811	-12.7537	in | indeed
0.315068	-12.7539	system | 2007).
0.409091	-12.7539	task | 2011
0.444444	-12.7539	outputs | each
0.483871	-12.7539	where | elements
0.319444	-12.754	the | made.
0.4	-12.7541	suppose | are
0.4	-12.7541	these | usually
0.174274	-12.7541	the | operations
0.207101	-12.7542	best | all
0.409091	-12.7545	w | d
0.3	-12.7546	we | derive
0.714286	-12.7548	topics | sms
0.146199	-12.7548	values | for
0.113677	-12.7548	an example | the
0.33871	-12.7548	is | crucial
0.192893	-12.7549	(i.e., | a
0.287356	-12.7549	few | have
0.647059	-12.755	rooted | node
0.196809	-12.755	baseline | our
0.194737	-12.7552	evaluating | and
0.076251	-12.7552	used | as
0.166038	-12.7553	later | the
0.0708447	-12.7554	5 | and
0.294118	-12.7555	of | roles.
0.571429	-12.7555	several | systems.
0.888889	-12.7556	popularity | context
0.315068	-12.7557	sentences | selected
0.273684	-12.7559	the | label,
0.225352	-12.7559	for | users
0.518519	-12.756	each | has | associated
0.222973	-12.756	from | user
0.273684	-12.7562	of | utterances.
0.6	-12.7562	lexical | hierarchical
0.195767	-12.7562	with | per
0.457143	-12.7564	contains | million
0.818182	-12.7564	noun phrases | concepts
0.267327	-12.7566	as | already
0.204545	-12.7567	model, | with
0.714286	-12.7567	performance | drops
0.457143	-12.7568	most | resources
0.425	-12.7569	to | match.
0.75	-12.7569	analyses | main
0.372549	-12.7569	of | similarity,
0.276596	-12.7569	the | model).
0.14939	-12.757	reduction | the
0.818182	-12.7571	cc | mle
0.198925	-12.7572	bilingual | for
0.38	-12.7573	that | hold
0.518519	-12.7574	and | bigrams.
0.16996	-12.7574	that | present
0.3	-12.7576	built | with
0.264706	-12.7576	one way | the
0.714286	-12.7577	focus | what
0.647059	-12.7577	translated | english.
0.165414	-12.7579	students | the
0.287356	-12.7579	all | documents
0.175214	-12.7579	the | creating
0.391304	-12.7581	our | analysis.
0.148485	-12.7582	the | ordering
0.296296	-12.7582	of | greatly
0.247863	-12.7583	integrate | and
0.391304	-12.7584	using | matching
0.6	-12.7584	scores | teachers
0.202247	-12.7585	of | phenomena
0.0860324	-12.7585	the | treebank
0.518519	-12.7586	correlated | with
0.159722	-12.7587	syntax | the
0.276596	-12.7589	properties | are
0.541667	-12.7589	parameters | algorithm
0.483871	-12.7589	analysis | showed
0.176724	-12.759	this | allows
0.225352	-12.7591	the | together.
0.425	-12.7594	slightly | but
0.647059	-12.7594	edinburgh | uk
0.150943	-12.7595	issues | the
0.888889	-12.7596	in | projects,
0.13285	-12.7597	in | reference
0.233083	-12.76	this | contains
0.0901001	-12.76	for | target
0.160714	-12.7602	for | selected
0.071831	-12.7602	we | both
0.310811	-12.7603	easily | by
0.24	-12.7604	matching | for
0.264706	-12.7605	typed | and
0.303797	-12.7606	these | correct
0.256881	-12.7606	an | has been
0.162455	-12.7607	the | place
0.165414	-12.7607	in | left
0.156997	-12.7607	example | from
0.414634	-12.7608	turn-taking | a
0.0636766	-12.7608	the | without
0.207101	-12.7609	formed | the
0.073991	-12.7609	task | to
0.273684	-12.7611	tag, | the
0.541667	-12.7611	from | literature
0.714286	-12.7612	nouns | associated
0.202247	-12.7614	change | to
0.4	-12.7614	it | structure.
0.276596	-12.7614	users | on
0.115809	-12.7614	that | well
0.666667	-12.7615	edits | edits
0.518519	-12.7617	story | which
0.541667	-12.7618	times | many
0.262136	-12.7619	analysis | different
0.264706	-12.7619	a | describing
0.363636	-12.7622	a | pick
0.425	-12.7623	of | 1998),
0.225352	-12.7627	strategies | to
0.888889	-12.7628	across | whole
0.46875	-12.7628	more specifically, | is
0.138298	-12.763	nullnull | nullnull
0.0778146	-12.7632	analysis | and
0.169291	-12.7632	assumed | the
0.306667	-12.7633	the | component,
0.22963	-12.7634	to | reasonable
0.571429	-12.7635	achieve | higher
0.6	-12.7635	displays | for
0.372549	-12.7636	using | wordnet
0.160714	-12.7636	applications | and
0.197802	-12.7637	the | file
0.205882	-12.7639	the | thereby
0.202247	-12.764	below | of
0.0874605	-12.764	proceedings of | 2011.
0.0891304	-12.7641	measure | of
0.21118	-12.7642	investigated | and
0.159011	-12.7644	vp | np
0.647059	-12.7644	asr | speech
0.444444	-12.7644	list | english
0.166038	-12.7644	it | new
0.425	-12.7646	documents | given
0.611111	-12.7646	that | actions.
0.647059	-12.7647	documents | author
0.108065	-12.7647	the | concepts
0.105828	-12.7648	specific | and
0.666667	-12.7649	and | style.
0.666667	-12.7649	(s | )
0.323529	-12.7649	constraint | be
0.328358	-12.765	rule | source
0.0388612	-12.765	each | a
0.518519	-12.7652	a | objective.
0.0826291	-12.7652	of | those
0.262136	-12.7653	a | document,
0.363636	-12.7653	semantic | type
0.157534	-12.7653	the | denoted
0.167315	-12.7654	a | see
0.425	-12.7656	of | frequency.
0.571429	-12.7656	used | technique
0.207101	-12.7657	detect | and
0.0540977	-12.7657	are | from
0.287356	-12.7657	ter | ter
0.647059	-12.7659	data: | data
0.303797	-12.7661	which | difficult
0.130641	-12.7661	the | section.
0.17004	-12.7661	of | tweets
0.323529	-12.7664	suppose | is
0.174468	-12.7664	to | real
0.208333	-12.7664	suggests that | of
0.75	-12.7665	selecting | selection
0.647059	-12.7665	and | (right)
0.0870934	-12.7666	instance | the
0.208333	-12.7668	occurrences | in
0.22963	-12.7672	that | limited
0.292683	-12.7673	appear | or
0.267327	-12.7673	describe | which
0.202247	-12.7675	follows | of
0.137931	-12.7675	2 | our
0.611111	-12.7679	interactive | which
0.241935	-12.7679	above | are
0.372549	-12.7679	the | window.
0.27957	-12.7679	the | datasets,
0.25	-12.768	and | yield
0.166023	-12.7681	or | their
0.195767	-12.7681	problem | an
0.0595412	-12.7682	some | of
0.2125	-12.7683	to | reading
0.611111	-12.7685	texts | various
0.38	-12.7686	describe | method
0.208589	-12.7686	de | and
0.6	-12.7687	have | meanings
0.186603	-12.7688	the | move
0.33871	-12.7688	further | using
0.350877	-12.7689	approach | unsupervised
0.483871	-12.769	accuracy | reported
0.222222	-12.7691	et | 2004),
0.75	-12.7692	structure, | only
0.296296	-12.7693	the | head,
0.184834	-12.7693	to | short
0.666667	-12.7694	support | machines.
0.33871	-12.7694	several | were
0.5	-12.7694	part-of-speech | part-of-speech
0.363636	-12.7695	using | stanford
0.444444	-12.7695	include | target
0.315068	-12.7698	of | causes
0.372549	-12.7698	in | columns
0.75	-12.7699	measuring | lexical
0.267327	-12.77	moreover, | are
0.296296	-12.77	notion | and
0.647059	-12.77	classifier | classes
0.425	-12.7701	while | result
0.571429	-12.7701	learning | students
0.231343	-12.7702	pick | the
0.190955	-12.7702	having | to
0.647059	-12.7703	are | following:
0.187192	-12.7703	unique | and
0.454545	-12.7704	metric | quality
0.124731	-12.7704	question | and
0.16129	-12.7704	additional | for
0.6	-12.7707	are | removed.
0.0937881	-12.7708	this | two
0.0736842	-12.7709	while | a
0.133005	-12.7709	the | marked
0.225352	-12.7709	of | respective
0.19	-12.771	different | than
0.611111	-12.7713	maximum | achieved
0.243697	-12.7713	distribution | word
0.382979	-12.7714	a | analyzer
0.169355	-12.7718	these | are | with
0.207101	-12.7719	settings | and
0.666667	-12.7719	automated | detection
0.666667	-12.7719	and | (1998).
0.222222	-12.7719	is | long
0.159011	-12.7719	a | e.g.,
0.252252	-12.7719	in | changes
0.414634	-12.772	their | within
0.425	-12.772	zhang, | and
0.127563	-12.772	we define | of
0.5	-12.7721	human | judgments
0.310811	-12.7722	with | smoothing
0.350877	-12.7722	all | included
0.391304	-12.7724	with | rules,
0.425	-12.7726	1, | 3
0.280899	-12.7727	that | scope
0.432432	-12.7729	for | n.
0.264706	-12.773	note that | our
0.46875	-12.7733	documents | topics
0.4	-12.7733	the | range.
0.243697	-12.7734	the | paraphrases.
0.571429	-12.7735	about | hours
0.414634	-12.7737	for | reasons,
0.666667	-12.7737	words | question.
0.571429	-12.7737	loss | when
0.75	-12.774	fitting | model
0.414634	-12.7743	past | that
0.382979	-12.7743	formally, | set of
0.137203	-12.7743	is | original
0.357143	-12.7744	class | instance
0.444444	-12.7744	left | side
0.245763	-12.7746	the | red
0.365385	-12.7746	powerful | to
0.414634	-12.7747	and | cues.
0.382979	-12.7748	disjoint | and
0.454545	-12.775	journal | penn
0.0696416	-12.775	is | two
0.164179	-12.775	task | which
0.145833	-12.7751	known | and
0.611111	-12.7752	if | available,
0.22963	-12.7752	proposed | using
0.171429	-12.7752	of | gene
0.33871	-12.7752	perspective | a
0.13555	-12.7752	of | ranking
0.454545	-12.7754	(2004) | with
0.222222	-12.7754	proposition | the
0.302632	-12.7754	the | constituents.
0.219178	-12.7755	0 | 10 | 20
0.302632	-12.7755	of | points.
0.141274	-12.7756	the | order.
0.273684	-12.7757	they | similar
0.247863	-12.7758	phrase | its
0.714286	-12.7759	functions | functions.
0.46875	-12.7759	in | johnson,
0.454545	-12.776	that | e1
0.482759	-12.776	large | corpus,
0.0892857	-12.7762	the | arguments
0.247863	-12.7764	that | (or
0.17004	-12.7765	pruning | the
0.127273	-12.7765	the | extended
0.432432	-12.7765	for | summarization,
0.181395	-12.7766	we | get
0.571429	-12.7766	consisting of | their
0.223776	-12.7768	a | 1).
0.204678	-12.7768	methods | or
0.571429	-12.7769	a | heuristics.
0.350877	-12.7769	is | simpler
0.190722	-12.7769	10 | 20
0.197802	-12.7771	also | same
0.318841	-12.7772	for | synsets
0.129108	-12.7772	details | the
0.5	-12.7774	characters | words
0.164122	-12.7775	(2010), | the
0.284091	-12.7776	similarity | cosine
0.6	-12.7776	outputs | systems
0.571429	-12.7777	for | draft
0.284091	-12.7778	1997; | al.,
0.323529	-12.7778	with | character
0.647059	-12.7778	from | 2000
0.155405	-12.778	the | cause
0.100567	-12.778	a | sequence
0.482759	-12.778	is | manner.
0.262136	-12.7781	task | semantic
0.19171	-12.7782	training data | for
0.382979	-12.7782	i=1 | j=1
0.14497	-12.7782	build | and
0.231343	-12.7784	alignment | translation
0.328125	-12.7785	of | conversations
0.391304	-12.7785	is | greedy
0.571429	-12.7785	of | developed.
0.571429	-12.7785	when | actually
0.611111	-12.7786	dividing | by
0.611111	-12.7786	2001; | 2005).
0.404762	-12.7789	prep | prep
0.714286	-12.7789	despite | large
0.482759	-12.7789	these | especially
0.314286	-12.779	linear | where
0.328125	-12.7791	in | area of
0.75	-12.7792	it | out,
0.171548	-12.7794	first | our
0.152104	-12.7795	20 | 20
0.166667	-12.7798	strategy | to
0.167315	-12.7799	the | zero
0.146707	-12.7799	8 | and
0.204678	-12.78	corpora, | and
0.180556	-12.78	together | a
0.183962	-12.7802	several | are
0.8	-12.7802	lattice | word-par
0.157895	-12.7804	denotes | of
0.227941	-12.7805	ambiguous | in
0.382979	-12.7805	event | sentiment
0.0875133	-12.7806	system | with
0.137931	-12.7806	classes | and
0.234375	-12.7806	also | source
0.234375	-12.7807	scope | in
0.382979	-12.7808	semantic | semantics
0.306667	-12.7808	simplest | of
0.160584	-12.7808	is | lower
0.146789	-12.7809	represents | and
0.190722	-12.7809	that | learned
0.126984	-12.781	way | a
0.254545	-12.7811	annotated | data
0.25	-12.7811	r, | the
0.357143	-12.7812	in | ordered
0.372549	-12.7814	filtering | from
0.647059	-12.7814	moreover, | often
0.5	-12.7814	all | strong
0.5	-12.7814	andrew | and
0.357143	-12.7815	varied | of
0.264706	-12.7815	trained | parallel
0.310811	-12.7817	performs | better
0.23622	-12.7817	author | a
0.211538	-12.7818	for instance, | as
0.333333	-12.7819	(and | as
0.52	-12.7819	as | redundant
0.131707	-12.782	lexicon | and
0.75	-12.782	pronunciation | speech
0.444444	-12.782	compared | standard
0.14011	-12.7821	achieved | of
0.404762	-12.7823	recent | et al.
0.289157	-12.7825	generating | for
0.8	-12.7825	fisher’s | test
0.136126	-12.7828	property | the
0.171548	-12.7828	we | aligned
0.404762	-12.7829	content | text
0.454545	-12.7832	nist | translation
0.25	-12.7833	a | exists
0.714286	-12.7835	senses | clustering
0.0974633	-12.7836	generation | the
0.203488	-12.7836	processing, | 27–31,
0.186275	-12.7837	would | more
0.268041	-12.7838	progress | the
0.666667	-12.7839	an | environment.
0.273684	-12.7839	that | whole
0.310811	-12.784	a | coherent
0.46875	-12.7841	instantiate | and
0.262136	-12.7841	sometimes | a
0.714286	-12.7842	parameters | optimized
0.714286	-12.7842	times | seen
0.372549	-12.7842	with | becomes
0.166023	-12.7842	are | words.
0.153333	-12.7843	a | mapping
0.314286	-12.7844	nevertheless, | and
0.201149	-12.7844	sentences | corpus
0.350877	-12.7844	a | error.
0.25	-12.7845	of | oov
0.129412	-12.7847	of | measures
0.197802	-12.7847	scored | the
0.8	-12.7847	reviews | domains.
0.357143	-12.7848	various | using
0.888889	-12.7849	scores, | scores.
0.611111	-12.7851	automatically | linguistic
0.208589	-12.7852	of | grammars
0.0899772	-12.7852	for | sentences
0.310811	-12.7854	questions | by
0.8	-12.7854	(liu | liu,
0.444444	-12.7855	partial | it
0.482759	-12.7856	step | algorithm
0.298701	-12.7857	is | issue
0.666667	-12.7858	projecting | from
0.888889	-12.7858	later | section.
0.647059	-12.7859	three | tasks.
0.149682	-12.7859	building | of
0.382979	-12.786	following: | a
0.8	-12.7862	nist | 2002)
0.46875	-12.7863	constraints | relation
0.136126	-12.7864	the | collected
0.75	-12.7865	less | 1%
0.114815	-12.7865	find | in
0.239669	-12.7866	accuracy | over
0.180556	-12.7866	the | collocation
0.666667	-12.7867	used | testing,
0.318841	-12.7868	from | corpora,
0.8	-12.7869	unlabeled | (uas)
0.25	-12.787	is | exact
0.571429	-12.787	than | baselines.
0.259615	-12.7872	performance | task
0.365385	-12.7873	counts | words
0.197802	-12.7874	choices | the
0.160584	-12.7874	a | before
0.714286	-12.7875	das | das
0.52	-12.7875	draw | each
0.0908033	-12.7878	of | linguistic
0.292683	-12.7879	 | 
0.227941	-12.7879	to | reliable
0.571429	-12.788	action | one
0.647059	-12.788	for | discussion,
0.284091	-12.7882	to | influence
0.454545	-12.7883	were | participants
0.296296	-12.7884	and | id
0.113553	-12.7885	additional | to
0.571429	-12.7886	models | efficient
0.432432	-12.7886	number | characters
0.404762	-12.7887	system | state-of-the-art
0.8	-12.7888	each | individually,
0.414634	-12.7889	for | w,
0.328125	-12.7892	also | quality
0.214286	-12.7892	model | language model
0.211538	-12.7892	of | properties of
0.414634	-12.7893	to | differences.
0.333333	-12.7893	model | estimated
0.52	-12.7894	make | less
0.432432	-12.7896	model | distortion
0.183575	-12.7896	to | relations.
0.149682	-12.7896	the | returned
0.75	-12.7897	only | satisfy
0.314286	-12.7899	corpora | english
0.888889	-12.7899	increases, | more
0.2	-12.7899	scale | to
0.482759	-12.7899	method | directly
0.147239	-12.79	message | the
0.365385	-12.79	contiguous | in
0.265306	-12.7903	which | high
0.0974633	-12.7903	by | number
0.541667	-12.7903	• | features:
0.8	-12.7903	katakana | english
0.0759804	-12.7904	average | of
0.75	-12.7904	decoding | seconds
0.212903	-12.7906	then | one
0.19171	-12.7907	tuple | the
0.23622	-12.7907	our | additional
0.179724	-12.7908	set | pairs
0.391304	-12.7909	the | art.
0.33871	-12.7909	the | evaluated.
0.227941	-12.7913	the | yielding
0.142442	-12.7915	find | for
0.0860441	-12.7916	that | will
0.404762	-12.7917	moreover, | also
0.318841	-12.7917	strict | and
0.15016	-12.792	our | baseline
0.888889	-12.792	web-derived | preference
0.289157	-12.792	for | lda
0.296296	-12.792	phonetic | of
0.382979	-12.7921	is | denoting
0.382979	-12.7921	implements | and
0.75	-12.7921	algorithm | parameters.
0.8	-12.7922	obtains | performance
0.129496	-12.7923	for | large
0.344828	-12.7924	is | noun.
0.112701	-12.7924	the | having
0.274725	-12.7925	in | absence
0.182243	-12.7925	consistent | of
0.8	-12.7926	procedure | finding
0.328125	-12.7927	however, | significant
0.323077	-12.7928	we | down
0.109402	-12.7928	resources | the
0.277778	-12.7928	the | object,
0.0197687	-12.7929	a | of | and
0.365385	-12.7929	children, | the
0.262136	-12.793	task | given
0.365385	-12.7931	nsf | grant
0.328125	-12.7932	for | nonterminal
0.5	-12.7933	this | inherent
0.114878	-12.7938	words, | and
0.217687	-12.7939	thank | and | their
0.365385	-12.7941	resulting | an
0.454545	-12.7942	1, | 3,
0.666667	-12.7942	f. | usa
0.75	-12.7943	identified | being
0.140845	-12.7944	view | the
0.375	-12.7944	system | floor
0.611111	-12.7945	university | united
0.227941	-12.7946	activation | the
0.262136	-12.7947	additionally, | in
0.375	-12.7947	using | heuristic
0.108514	-12.7948	first, | of
0.432432	-12.7948	in | participants.
0.611111	-12.7951	one | missing
0.5	-12.7951	on | ace
0.129187	-12.7952	the | 2)
0.265306	-12.7954	firstly, | the
0.262626	-12.7955	so | only
0.146341	-12.7956	that | found
0.454545	-12.7957	perform | over
0.137838	-12.7957	monolingual | and
0.302632	-12.7957	these | phrases
0.289157	-12.7959	set | value
0.230769	-12.7961	classifiers | on
0.170124	-12.7962	with | large
0.154639	-12.7963	mt | to
0.298701	-12.7963	statistical | models
0.259615	-12.7964	addressed | of
0.323077	-12.7964	a | spanning
0.234375	-12.7964	and | variants
0.259615	-12.7966	addresses | the
0.545455	-12.7966	forest | derivations
0.545455	-12.7966	used | mark
0.154639	-12.7967	this | better
0.155709	-12.797	frequencies | the
0.285714	-12.7971	and | constructions
0.432432	-12.7971	differences | language
0.314286	-12.7972	verbs | or
0.375	-12.7972	all | forms
0.223022	-12.7973	the | heads
0.111901	-12.7973	for | pair
0.289157	-12.7974	of | search.
0.110721	-12.7974	query | and
0.5	-12.7975	controlling | of
0.314286	-12.7976	for example, | english
0.350877	-12.7977	some | new
0.0601233	-12.7977	the | was | in
0.151815	-12.798	is | representation
0.254717	-12.798	of | list.
0.323077	-12.7981	above | which
0.212903	-12.7981	it | those
0.0632839	-12.7982	are | which
0.414634	-12.7983	identifies | that
0.234375	-12.7984	our | unsupervised
0.157143	-12.7985	in | nlp
0.571429	-12.7985	concerns | that
0.285714	-12.7989	of | another.
0.666667	-12.7989	cover | types
0.395349	-12.7991	that | occurred
0.318841	-12.7993	same | those
0.237705	-12.7994	reducing | of
0.52	-12.7994	average | points
0.391304	-12.7995	has been | where
0.277778	-12.7996	no | were
0.208861	-12.7997	2006), | et
0.666667	-12.7998	present | called
0.119192	-12.7999	the | frame
0.086262	-12.8	consider | of
0.8	-12.8001	policy | no
0.158273	-12.8003	4.2 | a
0.323077	-12.8004	we | regularization
0.130435	-12.8004	semantic | from
0.545455	-12.8004	iteration | algorithm
0.666667	-12.8005	expression | expressions
0.541667	-12.8005	select | top
0.158845	-12.8005	is | local
0.131514	-12.8005	of | nlp
0.382979	-12.8007	a | derivation.
0.350877	-12.8007	which | learned
0.178899	-12.8007	to | complete
0.252336	-12.8009	a | domain,
0.254717	-12.8009	while | some
0.129808	-12.801	for example, | and | in
0.083004	-12.8012	experiment | the
0.121339	-12.8012	a | automatically
0.302632	-12.8012	outcome | the
0.157143	-12.8016	this | case
0.0566728	-12.8016	in | but
0.161049	-12.8016	to | right
0.0997151	-12.8017	a | corresponding
0.541667	-12.8018	umass | stanford
0.186869	-12.8018	starting | and
0.611111	-12.802	ten | were
0.232558	-12.802	accurately | the
0.262626	-12.802	with | weights
0.203593	-12.802	experiments | results
0.254717	-12.8021	the | factor.
0.421053	-12.8024	classic | and
0.18593	-12.8025	specify | the
0.344828	-12.8025	a | phone
0.111504	-12.8027	several | in
0.0830861	-12.8028	2 | 1
0.257143	-12.8029	measure | two
0.75	-12.8031	2005) | syntactic
0.8	-12.8031	subset | characters
0.140805	-12.8032	preposition | the
0.5	-12.8032	from | distribution,
0.75	-12.8035	in | mail
0.201183	-12.8035	specifically, | in
0.404762	-12.8036	markers | in
0.666667	-12.8036	combine | models.
0.441176	-12.8039	system | to generate
0.309859	-12.8039	for | modified
0.202381	-12.804	shows | performance
0.611111	-12.8041	manually | list
0.545455	-12.8041	linear | size
0.241667	-12.8043	the | model, which
0.666667	-12.8046	• | indicate
0.229008	-12.8047	select | as
0.482759	-12.8047	always | same
0.12766	-12.8048	of | name
0.75	-12.8049	approach | simultaneously
0.0610605	-12.8049	case | the
0.482759	-12.8049	i=1 | (1)
0.0918984	-12.805	system, | the
0.206061	-12.805	substitution | and
0.257143	-12.805	mwe | a
0.338983	-12.8055	the | unchanged
0.382979	-12.8056	all | occur
0.289157	-12.8056	translation, | 2011.
0.0263002	-12.8058	the | model
0.314286	-12.8059	use | automatically
0.344828	-12.8059	difference | significant
0.14121	-12.8061	] | ]
0.358491	-12.8062	for | algorithms.
0.118474	-12.8063	translation | language
0.0726722	-12.8064	but | is
0.382979	-12.8065	normally | in
0.0881696	-12.8066	were | with
0.298701	-12.8068	the | outputs.
0.298701	-12.8068	also, | that
0.52	-12.8068	features | speaker
0.121795	-12.8069	numbers | the
0.245614	-12.8071	so | all
0.25	-12.8071	model | random
0.309859	-12.8071	a | answering
0.545455	-12.8072	r. | a.
0.151316	-12.8073	on | accuracy
0.432432	-12.8073	parallel | corpora.
0.083004	-12.8075	several | and
0.245614	-12.8076	additionally | the
0.224638	-12.8077	the | st
0.75	-12.8077	seven | were
1	-12.8077	nelken | shieber
1	-12.8077	interaction | interaction.
1	-12.8077	multiword expressions: | nlp.
1	-12.8077	substitutions, | deletions
1	-12.8077	(biemann | giesbrecht,
1	-12.8077	[3/ | [4/
1	-12.8077	(filimonov | harper,
1	-12.8077	“glad | angry,”
1	-12.8077	null1 | null2,
1	-12.8077	fin153 | fin200
1	-12.8077	goyal | g.s.
1	-12.8077	promoted | japan).
1	-12.8077	morfessor | lagus,
1	-12.8077	(chambers | jurafsky,
1	-12.8077	birke | sarkar’s
1	-12.8077	ceaf(m) | ceaf(e)
1	-12.8077	entrainment | entrainment
1	-12.8077	genres, | 23,
1	-12.8077	(nrf) | (mda)
0.5	-12.8077	is | present.
0.5	-12.8077	bf | bg
0.338983	-12.8079	knowledge base | a
0.186869	-12.8079	then, | and
0.125	-12.808	∧ | ∧
0.75	-12.8081	how | compares
0.395349	-12.8081	level | level.
0.875	-12.8081	el | es
0.875	-12.8081	“that’s | she
0.156028	-12.8082	the | student
0.454545	-12.8084	for | direction.
0.8	-12.8084	specifically, | preceding
0.8	-12.8084	iff | such
0.5	-12.8084	express | same
0.164706	-12.8085	settings | the
0.344828	-12.8085	is | learning.
0.875	-12.8085	2-21 | wsj
0.875	-12.8085	entailment | contradiction
0.875	-12.8085	geometrical | semantics,
0.118474	-12.8086	the | tree.
0.178082	-12.8088	source | translation
0.571429	-12.8088	instances | found
0.875	-12.8089	(barbosa | 2010)
0.114723	-12.8089	we | able
0.441176	-12.809	make | possible
0.109948	-12.8092	3.3 | the
0.0787825	-12.8092	the | considered
0.298701	-12.8092	our | topics
0.545455	-12.8092	simplifying | that
0.454545	-12.8093	table 5 | results
0.230769	-12.8093	future, | to
0.351852	-12.8093	a | value,
0.875	-12.8094	yejin | and
0.318182	-12.8094	the | improvements.
0.454545	-12.8095	hierarchical | rules
0.183575	-12.8096	it | large
0.245614	-12.8097	the | drops
0.545455	-12.8098	in | memory.
0.395349	-12.8103	derivation | tree
0.227273	-12.8103	we | evaluating
0.382979	-12.8104	task | detection
0.2	-12.8104	how | an
0.611111	-12.8105	verbal | argument
0.26	-12.8106	reveal | the
0.149837	-12.8106	entities | a
0.875	-12.8106	strict | match
0.875	-12.8106	speed | memory
0.172414	-12.8108	segments | the
0.611111	-12.8108	expression | transcription
0.168033	-12.8108	of | factors
0.204819	-12.8108	learner | a
0.75	-12.811	arc | added
0.178082	-12.8112	evaluating | of
0.175676	-12.8112	an | for each
0.344828	-12.8113	from | english.
0.230769	-12.8116	thus, | are
0.571429	-12.8118	user | system | user
0.875	-12.8119	entity | (rel)
0.302632	-12.812	our | goal
0.224638	-12.8123	project | to
0.571429	-12.8123	then | result
0.375	-12.8123	4th | on
0.545455	-12.8124	word | occurrence
0.432432	-12.8125	a | stores
0.211921	-12.8125	in | network
0.367347	-12.8125	is | strictly
0.271739	-12.8126	domain | which
0.545455	-12.8126	which | approximately
0.204969	-12.8127	that is, | to
0.195531	-12.8128	as | measure
0.351852	-12.8128	annotation | their
0.213333	-12.8131	approaches, | the
0.666667	-12.8132	ran | iterations
0.0580488	-12.8132	systems | the
0.25	-12.8133	of | candidates.
0.404762	-12.8134	choosing | from
0.328125	-12.8134	the | participants,
0.875	-12.8135	opi | opi
0.21831	-12.8136	when | they
0.358491	-12.8137	yielded | of
0.26	-12.8137	to | content.
0.148387	-12.8137	wide | of
0.75	-12.8138	updating | model
0.0660377	-12.8138	and | the same
0.611111	-12.814	dependency parsers | dependency
0.119835	-12.8142	the | of | senses
0.282353	-12.8143	operators | and
0.571429	-12.8143	1the | are
0.382979	-12.8144	like | where
0.875	-12.8144	joshua | (li
0.875	-12.8144	questions | yahoo!
0.309859	-12.8145	of | input,
0.285714	-12.8147	is | informative
0.26	-12.8148	dialogue | user
0.875	-12.8148	reader | referred
0.108997	-12.8148	number | each
0.466667	-12.8148	 | 
0.333333	-12.8149	topics | words
0.0688011	-12.8149	node | the
0.103555	-12.815	in | 3.
0.171674	-12.815	of | ambiguous
0.545455	-12.815	is | time-consuming
0.666667	-12.8151	some | needed
0.5	-12.8151	in | positions.
0.875	-12.8152	numbers | parentheses
0.875	-12.8152	indicate | presence
0.75	-12.8153	acknowledge | darpa
0.365385	-12.8153	with | lowest
0.203593	-12.8153	for | combined
0.666667	-12.8154	dictionaries | dictionary
0.386364	-12.8154	ambiguous | are
0.233871	-12.8154	features, | are
0.0946822	-12.8154	a | form
0.068	-12.8154	the | features.
0.0662843	-12.8155	is | only
0.109185	-12.8155	the | time.
0.344828	-12.8155	the | material.
0.291139	-12.8156	word | tagged
0.666667	-12.8157	descriptions | same
0.375	-12.8158	if | get
0.17094	-12.8158	to | a | system.
0.404762	-12.8159	visual | are
0.611111	-12.816	augment | from
0.100145	-12.816	algorithm | for
0.229008	-12.8161	yet | to
0.158088	-12.8161	a | study
0.611111	-12.8163	4.1 | some
0.545455	-12.8163	same | true
0.194444	-12.8164	in the | section,
0.482759	-12.8165	e, | in
0.0618102	-12.8167	the | output
0.611111	-12.8169	(2010) | wikipedia
0.410256	-12.8171	a | lexeme
0.75	-12.8171	abstract | rich
0.328125	-12.8172	generates | for
0.309859	-12.8172	construct | with
0.8	-12.8172	one, | one
0.0733802	-12.8174	for | other
0.666667	-12.8176	weights | maximize
0.216783	-12.8177	while | was
0.213333	-12.8177	to make | in
0.454545	-12.8178	develop | model
0.441176	-12.8178	a | interface.
0.52	-12.818	knowledge | form
0.333333	-12.8181	head | dependency
0.454545	-12.8182	several | single
0.8	-12.8184	prior | distribution.
0.165323	-12.8186	the | phase
0.5	-12.8189	topics, | topic
0.75	-12.8192	languages | arabic,
0.254717	-12.8194	of | paper | organized
0.375	-12.8194	model | discourse
0.174107	-12.8195	used | feature
0.421053	-12.8195	enabling | to
0.196629	-12.8195	on | sets.
0.143713	-12.8195	of | already
0.104762	-12.8196	counts | the
0.138122	-12.8196	for | average
0.227273	-12.8197	of | half
0.52	-12.8198	if | becomes
0.285714	-12.8199	only | type
0.179245	-12.8199	not | so
0.271739	-12.8199	the | alphabet
0.241379	-12.82	2010), | 2010),
0.611111	-12.8201	includes | includes
0.358491	-12.8203	is | multilingual
0.8	-12.8203	measure | ability
0.692308	-12.8204	following | (6)
0.5	-12.8204	annotators | asked
0.5	-12.8204	(15) | is
0.13587	-12.8205	with | data.
0.545455	-12.8205	unsupervised | 2009).
0.103774	-12.8205	the | cost
0.875	-12.8207	reordering, | reordering
0.0708029	-12.821	the | labeled
0.382979	-12.821	with | memory
0.345455	-12.8213	we | decision tree
0.666667	-12.8215	employing | features
0.666667	-12.8215	min | strategy
0.0789474	-12.8218	of | three
0.545455	-12.8218	design | such
0.5	-12.8218	associate | to
0.185	-12.8219	the | precision.
0.134921	-12.822	in | experiments,
0.163424	-12.822	of | person
0.2	-12.822	predicts | the
0.157509	-12.8221	to | appropriate
0.26	-12.8221	words | sentence.
0.75	-12.8221	mappings | syntactic
0.367347	-12.8222	minor | and
0.291139	-12.8224	by | parser.
0.291139	-12.8224	languages | have
0.5	-12.8225	lines | lines
0.545455	-12.8226	by | typed
0.298701	-12.8227	given | text
0.482759	-12.8227	determine | text
0.75	-12.8228	editing | text
0.52	-12.8233	because | occur
0.0555807	-12.8234	the | type
0.179245	-12.8234	our work | to
0.2	-12.8234	see that | is
0.367347	-12.8235	using | regression
0.75	-12.8236	reduces | performance
0.185	-12.8237	further | by
0.124153	-12.8237	can | different
0.27907	-12.8238	the | turn.
0.441176	-12.8238	here, | an
0.274725	-12.824	a | record
0.2875	-12.824	this | structured
0.122807	-12.8242	and | additional
0.229008	-12.8243	of | subjects
0.375	-12.8244	1we | the
0.404762	-12.8245	both | perform
0.875	-12.8245	adjective | adjectives
0.875	-12.8245	posts | responses
0.0311526	-12.8246	a | from
0.8	-12.8246	workshop | 1–9,
0.351852	-12.8246	these | allow
0.571429	-12.8247	dialogue | applications
0.466667	-12.8247	(as | section
0.327869	-12.8247	models | statistical
0.421053	-12.8248	(an | of
0.351852	-12.8248	difference | performance
0.0686813	-12.8249	the | the | system.
0.75	-12.825	order | exploit
0.309859	-12.8251	the | ignored.
0.410256	-12.8252	in | intelligent
0.323077	-12.8253	however | not
0.52	-12.8253	given | match
0.875	-12.8253	markov | (hmms)
0.578947	-12.8253	in | field,
0.454545	-12.8254	for example, | means
0.75	-12.8254	is | straightforward,
0.233871	-12.8254	actual | in
0.25	-12.8254	features | whether
0.204969	-12.8254	the | subjectivity
0.358491	-12.8255	overall | system
0.20625	-12.8256	of | documents,
0.8	-12.8257	obtain | optimal
0.262626	-12.8259	reduction | a
0.11546	-12.8261	no | for
0.358491	-12.8262	dataset | used
0.275862	-12.8264	implicitly | the
0.140762	-12.8264	to | final
0.117886	-12.8264	of | arguments
0.233871	-12.8265	specifically | of
0.166667	-12.8266	after | with
0.52	-12.8266	with | experience
0.323077	-12.8267	constitute | of
0.611111	-12.8269	in | success.
0.482759	-12.8269	results | support
0.5	-12.8269	find | good
0.875	-12.827	wizard: | lied
0.367347	-12.827	use | weights
0.1	-12.8271	compare | and
0.421053	-12.8271	increases | number
0.309859	-12.8271	rules | →
0.13764	-12.8272	we | four
0.395349	-12.8273	in | constituency
0.20625	-12.8274	analyzing | of
0.285714	-12.8275	an | version
0.611111	-12.8275	d | wi
0.185	-12.8275	in | long
0.571429	-12.8275	case | multiple
0.351852	-12.8275	that | correct.
0.291139	-12.8277	and | baselines
0.202454	-12.8277	given | training
0.692308	-12.8278	our | lookahead
0.2875	-12.8278	by | either | or
0.115686	-12.8279	the | derivation
0.27907	-12.8279	the | falls
0.441176	-12.8279	conversion | dependency
0.211921	-12.828	and | (i.e.
0.578947	-12.8281	provide | knowledge
0.410256	-12.8282	more | modeling
0.611111	-12.8284	other | methods,
0.182266	-12.8284	used | models
0.141593	-12.8284	for | development
0.578947	-12.8284	of | story.
0.222222	-12.8285	interactions | and
0.441176	-12.8285	now | all
0.161538	-12.8285	of | response
0.275862	-12.8287	of | cell
0.127711	-12.8287	the | to obtain
0.265957	-12.8288	give | on
0.75	-12.829	common | present
0.305556	-12.829	amount | data.
0.245455	-12.829	moreover, | that
0.121739	-12.8294	makes | of
0.194444	-12.8294	in | algorithm.
0.265957	-12.8295	features | classifier
0.692308	-12.8295	(these | are
0.0848356	-12.8298	lexical | in
0.578947	-12.8298	carnegie | 15213
0.0698503	-12.8299	the | problem
0.482759	-12.83	to | discourse.
0.358491	-12.8302	an | towards
0.611111	-12.8302	on | metrics,
0.545455	-12.8303	emotional | emotional
0.0643594	-12.8304	we | only
0.265957	-12.8304	but, | the
0.153846	-12.8305	(6) | the
0.0778175	-12.8306	the | significantly
0.1875	-12.8307	strategy | in
0.133508	-12.8307	the | distinct
0.309859	-12.8307	long | that
0.625	-12.8309	there | variations
0.482759	-12.8309	the | intractable.
0.410256	-12.831	their | meaning
0.172566	-12.831	student | the
0.129032	-12.8311	× | ×
0.2875	-12.8311	we | looking
0.345455	-12.8312	the | difficult.
0.233871	-12.8313	produce | that
0.875	-12.8313	penn | (maamouri
0.454545	-12.8314	which | suitable
0.220588	-12.8315	explicitly | in
0.0916976	-12.8315	pair | and
0.257426	-12.8316	our | ranking
0.271739	-12.8318	monolingual | on
0.578947	-12.832	n, | number
0.143293	-12.8321	with | best
0.113636	-12.8323	after | a
0.75	-12.8323	tree-to-string | tree
0.395349	-12.8325	presented | novel
0.386364	-12.8326	with | strings
0.268817	-12.8326	(which | in
0.454545	-12.8327	from | 30
0.75	-12.8327	from | users,
0.692308	-12.8328	training: | and
0.282353	-12.833	arc | to
0.8	-12.8331	as | misses
0.395349	-12.8332	(2008) | al.
0.207792	-12.8332	from | data,
0.268817	-12.8332	are | binary
0.692308	-12.8332	summarization, | summary
0.545455	-12.8332	first, | so
0.428571	-12.8334	communications | the
0.351852	-12.8335	original | was
0.195402	-12.8337	limited | in
0.666667	-12.8338	whose | threshold
0.106312	-12.8338	standard | in
0.428571	-12.8338	defined | (3)
0.294872	-12.8338	is | improving
0.386364	-12.8341	summarize | a
0.163347	-12.8341	synsets | the
0.11368	-12.8345	the | finding
0.0946124	-12.8345	features, | the
0.345455	-12.8346	5.4 | and
0.333333	-12.8347	share | with
0.8	-12.8347	present | series
0.8	-12.8347	words | collocations
0.8	-12.8347	tries | find
0.351852	-12.8347	is | promising
0.351852	-12.8347	the | bus
0.188482	-12.8348	that | individual
0.666667	-12.8348	problem | mt
0.666667	-12.8348	sample | generated
0.75	-12.8349	and | utility.
0.25	-12.8349	for | iterations
0.692308	-12.8349	selecting | correct
0.692308	-12.8349	obtain | results.
0.188482	-12.8351	in | (or
0.692308	-12.8352	adapts | to
0.263158	-12.8353	framework, | and
0.428571	-12.8354	the | neutral.
0.282353	-12.8355	denote | set
0.10678	-12.8357	a | graph
0.189189	-12.8358	for | next
0.875	-12.836	official | department
0.275862	-12.8363	showing | in
0.138418	-12.8363	3.3 | the | of
0.117284	-12.8364	we | additional
0.254902	-12.8366	of | investigated
0.150502	-12.8366	function | =
0.228346	-12.8366	the | organization
0.8	-12.8367	liblinear | al.,
0.297297	-12.8368	first | documents
0.875	-12.8369	…, | then
0.5	-12.8369	where | replaced
0.124711	-12.837	we | compare
0.4	-12.8371	received | from
0.21831	-12.8371	3 | different
0.12619	-12.8374	good | and
0.268817	-12.8374	however, | large
0.191257	-12.8374	several | have
0.27907	-12.8375	in | synchronous
0.428571	-12.8375	construct | using
0.692308	-12.8376	values | averages
0.4	-12.8377	1995) | of
0.190217	-12.8378	is | correctly
0.180488	-12.838	and | contexts
0.0954301	-12.8381	data | by
0.323077	-12.8382	all | words,
0.333333	-12.8382	and | co-occur
0.245455	-12.8382	compare | using
0.611111	-12.8383	a | recognizer.
0.377778	-12.8383	are | mostly
0.441176	-12.8384	syntax-based | translation
0.282353	-12.8385	however, | so
0.0844773	-12.8388	significantly | the
0.164659	-12.8388	of | highly
0.294872	-12.8389	health | and
0.0678547	-12.8391	most | in
0.247706	-12.8391	a | dictionary.
0.5	-12.8391	conditions | which
0.283951	-12.8392	generation | which
0.167364	-12.8392	if | was
0.228346	-12.8392	rouge | the
0.625	-12.8392	simply | label
0.222222	-12.8393	verbs | are
0.223881	-12.8393	(2009) | on
0.135135	-12.8394	person | the
0.142424	-12.8394	first | word
0.27907	-12.8396	a | value.
0.339286	-12.8396	next, | for
0.206452	-12.8397	parse trees | the
0.0613839	-12.8397	over | of
0.578947	-12.8397	algorithm | rules.
0.134771	-12.8398	directly | to
0.191257	-12.8398	new | or
0.291139	-12.84	element | is
0.395349	-12.84	and | pronunciation
0.692308	-12.84	one | think
0.8	-12.8402	rules | syntactically
0.692308	-12.8403	4the | is
0.36	-12.8404	words | concepts
0.126492	-12.8404	to | full
0.75	-12.8404	as | alternative,
0.124424	-12.8405	has | which
0.10045	-12.8405	of | nodes
0.26	-12.8407	will | any
0.875	-12.8407	hashtags | twitter
0.395349	-12.8408	both | general
0.75	-12.8411	report | accuracy.
0.377778	-12.8411	can | capture
0.545455	-12.8412	increase | recall
0.367347	-12.8412	tasks | task
0.428571	-12.8413	research | automatic
0.692308	-12.8413	what | done
0.116466	-12.8414	the | names
0.8	-12.8414	srl | preposition
0.75	-12.8415	guidelines | guidelines
0.875	-12.8416	cross-lingual | translations
0.197674	-12.8416	system | over
0.263158	-12.8417	tried | and
0.327869	-12.8417	the | assumption.
0.257426	-12.8418	of | lexicons
0.451613	-12.8418	therefore, | will
0.545455	-12.8418	is | viable
0.297297	-12.8419	head | its
0.175926	-12.8421	the | follows
0.578947	-12.8422	improved | quality
0.75	-12.8423	wordnet | wordnet.
0.173516	-12.8423	starts | the
0.233333	-12.8423	the | team
0.220588	-12.8425	and | 2).
0.257426	-12.8425	a | fragment
0.441176	-12.8428	(cohen | al.,
0.243243	-12.8429	using | classifier
0.189189	-12.8429	also | into
0.375	-12.8429	improve | results
0.451613	-12.8429	is | faster than
0.75	-12.843	date | time
0.377778	-12.843	while | output
0.410256	-12.8431	architecture | system
0.0444055	-12.8432	the | score
0.118896	-12.8432	larger | of
0.254902	-12.8433	= | |
0.611111	-12.8434	used | study.
0.75	-12.8434	0.72 | 0.78
0.333333	-12.8434	store | and
0.11368	-12.8435	the | action
0.152778	-12.8435	concept | and
0.228346	-12.8435	third, | the
0.133333	-12.8435	a | string
0.666667	-12.8436	tasks | question
0.153571	-12.8436	combined | a
0.247706	-12.8436	label | each
0.120614	-12.8437	specific | in
0.692308	-12.8437	designed | provide
0.0629042	-12.8437	rule | the
0.386364	-12.8438	there | clear
0.188172	-12.8438	are | nodes
0.5	-12.8438	president | in
0.119149	-12.8439	from | sentence
0.0885781	-12.844	a | n
0.0693848	-12.8441	english | and
0.8	-12.8441	43rd | of
0.351852	-12.8441	type, | is
0.0720372	-12.8442	of | rules
0.666667	-12.8443	with | coming
0.410256	-12.8443	in | notation
0.386364	-12.8443	features | combined
0.275862	-12.8444	words | them
0.254902	-12.8445	information | about
0.0803485	-12.8445	is | over
0.131783	-12.8446	baseline | with
0.268817	-12.8446	abstract | semantic
0.875	-12.8446	target | targets.
0.421053	-12.8448	symmetric | and
0.578947	-12.8448	string | some
0.0336191	-12.8449	a | or
0.8	-12.8449	malt+mst | significant
0.220588	-12.845	features | are | features
0.11194	-12.845	of | overall
0.297297	-12.845	systematically | the
0.545455	-12.845	in | situations.
0.0459408	-12.8452	this | with
0.666667	-12.8453	apply | technique
0.36	-12.8454	we | modifiers
0.222222	-12.8454	are | e.g.,
0.333333	-12.8454	the | tail
0.578947	-12.8454	(q | )
0.377778	-12.8454	i | go
0.625	-12.8456	lists | both
0.521739	-12.8457	analyzes | and
0.578947	-12.8457	constituent | tree
0.165289	-12.8458	network | and
0.164	-12.8462	we | explore
0.142424	-12.8462	two | were
0.0977011	-12.8463	subject | the
0.367347	-12.8463	write | for
0.451613	-12.8465	average | difference
0.112381	-12.8466	nodes | and
0.207792	-12.8466	a | occurs
0.27907	-12.8467	of | books
0.232	-12.8468	to | automated
0.327869	-12.8468	(one | and
0.611111	-12.847	extracted from | text
0.173516	-12.8472	documents, | the
0.254902	-12.8473	of | colour
0.521739	-12.8473	and | quickly.
0.333333	-12.8474	proved | in
0.150171	-12.8475	trained | with
0.188172	-12.8476	through | that
0.144654	-12.8476	to | social
0.193182	-12.8476	to | elements
0.260417	-12.8477	two | corresponding
0.202454	-12.8477	5.1 | in
0.322581	-12.8478	compute | similarity
0.313433	-12.8479	these | additional
0.339286	-12.848	was | during
0.428571	-12.8481	extraction | patterns
0.666667	-12.8482	these | confirm
0.138329	-12.8482	the | tuning
0.254902	-12.8483	has | been | for
0.318182	-12.8483	tuning | with
0.183673	-12.8485	a | line
0.128079	-12.8485	state | a
0.333333	-12.8485	all | highest
0.481481	-12.8486	with | target.
0.269663	-12.8487	while | does
0.183673	-12.8489	a | built
0.875	-12.8489	state: | [slot1],
0.178744	-12.849	names | and
0.224806	-12.8492	of | generic
0.367347	-12.8492	with | depth
0.0482471	-12.8493	a | of | which
0.260417	-12.8494	a detailed | of
0.578947	-12.8494	had | correct
0.2	-12.8495	in | direct
0.210884	-12.8495	with | models.
0.272727	-12.8496	a | file
0.327869	-12.8496	for | weighting
0.243243	-12.8498	uses | which
0.428571	-12.8498	suggests that | more
0.123853	-12.8501	experiments | with
0.239316	-12.8502	the | (to
0.120879	-12.8503	of | verbs
0.0931436	-12.8503	training | with
0.197605	-12.8504	of | computation
0.345455	-12.8504	says | is
0.14375	-12.8505	one | all
0.0207277	-12.8505	the | is | in
0.377778	-12.8506	(2010) | they
0.140719	-12.8507	being | in
0.322581	-12.8507	is | (cf.
0.089697	-12.8507	the | metric
0.189189	-12.8508	for | m
0.241071	-12.8509	in | extraction.
0.327869	-12.8512	recognize | in
0.195402	-12.8514	overlap | of
0.578947	-12.8514	if | returns
0.128079	-12.8516	in | we present
0.351852	-12.8518	this | reduced
0.308824	-12.8519	exercise | the
0.297297	-12.8519	a | vocabulary.
0.30137	-12.852	these | suggest
0.265957	-12.852	results | without
0.215278	-12.852	in | database
0.625	-12.8521	mentions | names
0.625	-12.8521	while | former
0.245455	-12.8521	assuming | and
0.521739	-12.8522	consists | following
0.521739	-12.8522	in | tweet.
0.189189	-12.8523	words | context
0.103728	-12.8525	the | sentences,
0.131579	-12.8525	observation | the
0.160156	-12.8526	in | translation.
0.441176	-12.8526	had | their
0.101382	-12.8527	to | out
0.466667	-12.8527	size | 100
0.283951	-12.8528	facilitate | of
0.217391	-12.8528	already | for
0.625	-12.8528	situations | or
0.451613	-12.8528	negative | overall
0.226562	-12.8528	generates | and
0.243243	-12.8529	the | proteins
0.5	-12.853	automatically | lexical
0.101721	-12.8531	portland, | june
0.2875	-12.8532	at | node
0.140299	-12.8533	term | a
0.206452	-12.8536	words | english
0.481481	-12.8538	need | large
0.252427	-12.8538	p: | p:
0.237288	-12.8541	records | the
0.8	-12.8541	following | sentence:
0.184615	-12.8543	design | to
0.308824	-12.8544	problem | into
0.451613	-12.8546	edges | graph
0.545455	-12.8547	we | state-ofthe-art
0.352941	-12.8551	ontological | to
0.352941	-12.8551	are | completely
0.140299	-12.8551	the | reported in
0.272727	-12.8551	set | labels
0.17619	-12.8551	and | it.
0.36	-12.8552	method | extract
0.0962343	-12.8554	general | and
0.135359	-12.8556	of | algorithms
0.228346	-12.8557	accuracy | test
0.481481	-12.8558	algorithm | return
0.545455	-12.8558	frame | frames
0.545455	-12.8558	content | was
0.158491	-12.8559	is | term
0.666667	-12.8559	finally, | generated
0.223077	-12.8559	2.4 | the
0.119565	-12.8563	4. | of
0.174312	-12.8563	people | to
0.441176	-12.8563	to | grammars.
0.30137	-12.8564	empirical | on
0.129442	-12.8564	of | detection
0.245455	-12.8564	exact | for
0.100304	-12.8565	is | function
0.0425648	-12.8565	data | and
0.265957	-12.8565	a | score,
0.121622	-12.8565	we | measure
0.8	-12.8565	classifier | detect
0.327869	-12.8566	while | higher
0.36	-12.8567	deal | with
0.322581	-12.8567	to | onto
0.257732	-12.8567	state | an
0.466667	-12.8569	using | exact
0.466667	-12.8569	= | nproductdisplay
0.308824	-12.857	hyperparameters | the
0.231405	-12.857	environment | the
0.179612	-12.8571	graphs | of
0.4	-12.8571	we leave | work.
0.283951	-12.8573	while | many
0.451613	-12.8573	is | threshold,
0.155556	-12.8575	therefore, | is
0.122449	-12.8575	that | likely
0.367347	-12.8575	relative position | the
0.31746	-12.8578	in | near
0.692308	-12.8578	tuple | its
0.265957	-12.858	cell | and
0.129442	-12.858	explicitly | the
0.252427	-12.858	of | remains
0.243243	-12.8581	distance | between
0.113281	-12.8582	trained | on
0.101721	-12.8583	model | an
0.107965	-12.8584	the | change
0.410256	-12.8584	is | hpsg
0.481481	-12.8584	words | phrases.
0.578947	-12.8584	make | sense
0.141141	-12.8585	particular | to
0.481481	-12.8587	(wiebe | al.,
0.322581	-12.8589	and | regarding
0.252427	-12.8589	automatically | using
0.0642946	-12.8591	we | their
0.521739	-12.8592	show | impact
0.333333	-12.8593	not | produce
0.30137	-12.8594	moreover, | not
0.297297	-12.8594	initial | set
0.545455	-12.8594	summaries | both
0.11609	-12.8596	the best | for
0.272727	-12.8597	ned | the
0.173516	-12.8597	to | known
0.4	-12.8597	our | underlying
0.377778	-12.8597	topic | (blei
0.481481	-12.8599	a | lattice.
0.0471222	-12.86	more | to
0.189189	-12.86	the | independently
0.18617	-12.8602	filtered | the
0.238938	-12.8603	and | suggest
0.229508	-12.8603	to | input.
0.193182	-12.8603	for | five
0.875	-12.8603	should | mind
0.134615	-12.8603	we | including
0.277108	-12.8603	discourse | or
0.13089	-12.8604	study | in
0.466667	-12.8604	european | grant
0.410256	-12.8606	designed | this
0.173516	-12.8606	the | intermediate
0.185185	-12.8606	this | system.
0.625	-12.8606	toolkit | train
0.333333	-12.8607	all, | in
0.154412	-12.8608	the | reasonable
0.545455	-12.8608	and | morphology.
0.0988201	-12.8609	in | input
0.451613	-12.8609	a | (called
0.205128	-12.861	4.1 | is
0.297297	-12.861	english | translated
0.056057	-12.8611	a | sentence
0.185185	-12.8611	said | the
0.106713	-12.8612	for | document
0.875	-12.8612	translating | german
0.521739	-12.8613	since | represent
0.223077	-12.8613	grammaticality | the
0.578947	-12.8613	systems | 2002).
0.254902	-12.8614	several | such
0.119565	-12.8614	problems | the
0.226562	-12.8614	the rest of | is
0.297297	-12.8617	the | independently.
0.174528	-12.8619	by | system.
0.269663	-12.8623	was | project
0.466667	-12.8623	comparing | see
0.367347	-12.8623	they | what
0.266667	-12.8624	tf-idf | the
0.416667	-12.8624	parsers | parser
0.521739	-12.8624	users | system.
0.275862	-12.8625	alternative | that
0.625	-12.8625	models | optimized
0.875	-12.8625	displayed | user
0.36	-12.8626	we | lexicalized
0.352941	-12.8627	200 | 400
0.175355	-12.8627	with | original
0.0644753	-12.8628	they | a
0.158915	-12.8629	we | observed
0.150877	-12.8629	which | both
0.545455	-12.8629	(section | 4).
0.118143	-12.8633	to | through
0.145631	-12.8634	making | of
0.1	-12.8634	the | makes
0.441176	-12.8634	corpus | four
0.247619	-12.8635	pair | words
0.390244	-12.8636	institute | information
0.147651	-12.8636	the | arc
0.148148	-12.8636	of | edges
0.521739	-12.8637	sentences | marked
0.280488	-12.8637	and | arbitrary
0.369565	-12.8637	1998) | a
0.173516	-12.8638	study | for
0.875	-12.8638	sciences | 90292
0.182741	-12.8639	cost | and
0.390244	-12.864	the | pronouns.
0.625	-12.864	described | (2010).
0.236842	-12.8641	using | following
0.8	-12.8641	information | modules.
0.578947	-12.8642	parser | curran,
0.124413	-12.8642	for | noun
0.13388	-12.8643	we | random
0.451613	-12.8643	infrastructure | and
0.308824	-12.8645	phrases | their
0.154412	-12.8646	from | new
0.386364	-12.8648	to | query,
0.304348	-12.8648	mixed | and
0.212329	-12.8648	for | 20
0.0796545	-12.8649	using | with
0.322581	-12.8649	of | theoretical
0.8	-12.8649	mutual | neighbor
0.283951	-12.8652	will | us
0.14966	-12.8652	finally, | is
0.369565	-12.8652	of | hierarchy.
0.123832	-12.8652	dependencies | the
0.322581	-12.8653	split | two
0.369565	-12.8653	in | rarely
0.578947	-12.8653	and | svm.
0.152482	-12.8655	g | a
0.18	-12.8656	different | used
0.0514676	-12.8656	an | that
0.179104	-12.8657	of | constituent
0.428571	-12.8657	so | need
0.236842	-12.8658	a | index
0.163265	-12.8658	of | fixed
0.345455	-12.8658	modify | and
0.0628743	-12.8658	is | if
0.229508	-12.8659	phase, | the
0.257732	-12.8661	the | wm
0.8	-12.8661	system | detects
0.201258	-12.8662	essentially | the
0.339286	-12.8665	final | all
0.875	-12.8665	functor-argument | phrase
0.377778	-12.8665	a total | were
0.428571	-12.8666	and | business
0.205128	-12.8666	and | analysis.
0.386364	-12.8667	in | stages
0.369565	-12.8668	ng, | and
0.55	-12.8668	table 2 presents | of
0.260417	-12.8669	used | input
0.625	-12.8669	error analysis | error
0.205128	-12.8669	name | as
0.692308	-12.8669	in | indices.
0.428571	-12.867	people | use
0.236842	-12.8671	in | n-best
0.125908	-12.8671	certain | of
0.545455	-12.8671	and | 2006.
0.578947	-12.8671	task | prediction
0.25	-12.8672	created | with
0.428571	-12.8672	sentiment | about
0.137143	-12.8672	and | events
0.18617	-12.8674	proposed | which
0.159533	-12.8675	for | sequence
0.327869	-12.8675	the | parser’s
0.352941	-12.8677	still | more
0.208054	-12.8677	used | target
0.8	-12.8677	runs | time
0.2	-12.8678	we investigate | of
0.333333	-12.8679	and | together.
0.451613	-12.868	created | used
0.578947	-12.868	19 | 20
0.416667	-12.8681	suggests | may
0.390244	-12.8681	is | factor.
0.8	-12.8681	has | senses,
0.521739	-12.8684	extraction | biomedical
0.229508	-12.8684	method | both
0.416667	-12.8685	the | projection.
0.155556	-12.8687	and | weights
0.327869	-12.8688	describe | or
0.5	-12.8688	employing | for
0.247619	-12.8689	of | representative
0.410256	-12.8689	rules | reordering
0.386364	-12.8689	there | reasons
0.451613	-12.8691	contributes | to
0.36	-12.8692	that | opposite
0.0510774	-12.8692	the | approach
0.0822199	-12.8693	the | question
0.195266	-12.8693	training | models
0.625	-12.8694	used | place
0.280488	-12.8694	built | from
0.283951	-12.8695	and | adapted
0.416667	-12.8695	kendall’s | for
0.521739	-12.8695	n-grams | n-gram
0.106457	-12.8696	we | out
0.36	-12.8697	developed | an
0.203822	-12.8697	only | small
0.241071	-12.8698	we | default
0.124122	-12.8698	its | with
0.545455	-12.8698	result | being
0.875	-12.87	wi | wj,
0.148148	-12.87	in | position
0.377778	-12.87	is | computationally
0.128141	-12.8701	a | tag
0.451613	-12.8702	name | entities
0.8	-12.8702	is | complex.
0.322581	-12.8703	in | appeared
0.466667	-12.8703	make | large
0.0441791	-12.8704	the | score | the
0.134454	-12.8704	sentence | or
0.466667	-12.8705	it | possibly
0.136752	-12.8706	ranking | to
0.4	-12.8706	measure | when
0.266667	-12.8707	of | granularity
0.109091	-12.8707	sentences | are
0.289474	-12.8708	methods | they
0.142415	-12.8709	and | perform
0.247619	-12.8711	it was | that
0.36	-12.8712	and | records
0.209459	-12.8712	of | currently
0.187845	-12.8712	variants | and
0.625	-12.8713	lack | annotated
0.875	-12.8713	|d|
0.285714	-12.8714	of | biological
0.416667	-12.8714	the | jeopardy!
0.255102	-12.8714	consider | following
0.8	-12.8714	correlation | opi
0.481481	-12.8715	2 | architecture
0.55	-12.8715	decoding | alignment
0.283951	-12.8716	the | (lines
0.451613	-12.8716	share | they
0.159533	-12.8717	of | applications
0.428571	-12.8718	rank | their
0.0493323	-12.8718	we | all
0.875	-12.8718	performance | sa
0.283951	-12.8719	6.3 | of
0.208054	-12.8721	we can | which
0.229508	-12.8721	step | from
0.55	-12.8721	however, | nature
0.122402	-12.8721	the | roles
0.247619	-12.8724	employed | to
0.3	-12.8724	the | smt.
0.277108	-12.8724	not | because | not
0.170404	-12.8725	with | words.
0.205298	-12.8725	behavior | of
0.31746	-12.8725	translation | 2007).
0.521739	-12.8726	automatic | processing
0.352941	-12.8727	to | areas
0.352941	-12.8727	sets | sentences
0.304348	-12.8727	used | determine
0.428571	-12.8729	in | positively
0.120805	-12.8731	table 4 | the
0.255102	-12.8731	use | available
0.242991	-12.8731	the | proficiency
0.269663	-12.8732	were | human
0.521739	-12.8732	to | efficiency.
0.0890244	-12.8733	the | represent
0.625	-12.8735	perceptron | learning
0.31746	-12.8736	addressed | in
0.466667	-12.8736	participated | task
0.578947	-12.8736	standard | 2003).
0.339286	-12.8737	seem | for
0.451613	-12.8737	output | some
0.242991	-12.8737	tweet | a
0.352941	-12.8738	recently, | have
0.15035	-12.8738	of | complexity
0.293333	-12.8738	constructed | by
0.289474	-12.8739	phenomenon | in
0.308824	-12.874	latent | which
0.875	-12.874	items | chart
0.545455	-12.874	rules, | an
0.5	-12.8741	we | rescoring
0.238938	-12.8742	nouns | are
0.241071	-12.8742	linguistic | or
0.692308	-12.8743	but | equally
0.266667	-12.8744	pairs | same
0.4	-12.8745	w. | and
0.369565	-12.8745	new | current
0.266667	-12.8746	the | person,
0.322581	-12.8746	2010 | for
0.55	-12.8746	group | university
0.8	-12.8746	maximize | objective
0.390244	-12.8748	corpus | linguistic
0.205298	-12.8748	steps | and
0.144695	-12.8748	that | generated
0.3125	-12.875	valuable | for
0.241071	-12.8752	repeatedly | the
0.727273	-12.8753	0.86 | 0.82
0.875	-12.8753	mitchell | lapata | models
0.4375	-12.8754	values, | for
0.8	-12.8754	models | richer
0.451613	-12.8755	knowledge | world
0.727273	-12.8756	guaranteed | optimal
0.727273	-12.8756	strategy | space
0.12156	-12.8758	to | overall
0.521739	-12.8758	test | evaluation.
0.161943	-12.8758	of | used.
0.416667	-12.8759	we | by selecting
0.8	-12.8759	errors | non-native
0.0605381	-12.876	given | a
0.25	-12.876	a | annotator
0.169643	-12.8763	2009) | a
0.0623875	-12.8763	input | the
0.625	-12.8764	models | parsing.
0.236842	-12.8765	[ | [
0.169643	-12.8767	percentage | in
0.0260988	-12.8767	a | with
0.692308	-12.8767	is | fastest
0.8	-12.8767	while | remain
0.191011	-12.877	their | using
0.137026	-12.8771	the | access
0.161943	-12.8771	that | after
0.242991	-12.8771	feature | sentence
0.234783	-12.8773	representative | of
0.104202	-12.8773	for | best
0.521739	-12.8774	evaluation | good
0.428571	-12.8774	a | (not
0.352941	-12.8775	that | coherent
0.0983607	-12.8775	present | a
0.727273	-12.8779	todd | and
0.180905	-12.8779	a | rank
0.134454	-12.8779	on | corpus.
0.875	-12.878	findings, | publication
0.0995406	-12.8781	= | summationdisplay
0.0844298	-12.8781	4.2 | the
0.147157	-12.8783	the | sections
0.269663	-12.8783	sentiment | using
0.14377	-12.8783	goal | is
0.308824	-12.8786	documents | into
0.214286	-12.8787	run | for
0.280488	-12.8788	these | input
0.0492354	-12.879	experiments | the
0.0817805	-12.8791	must | the
0.12	-12.8792	the | classifiers
0.147651	-12.8792	new | by
0.578947	-12.8792	split | three
0.182292	-12.8792	a | unique
0.25	-12.8794	a | namely
0.234783	-12.8794	the | analyzer
0.428571	-12.8794	experiment | used
0.3125	-12.8795	sentences | annotated
0.0959885	-12.8795	is | since
0.308824	-12.8797	and | dialogues
0.197531	-12.8801	second, | is
0.144695	-12.8801	some | from
0.0973451	-12.8803	for | with | with
0.101587	-12.8803	new | for
0.221374	-12.8805	perspective | the
0.0495087	-12.8806	as | by
0.578947	-12.8806	(between | and
0.149826	-12.8807	sections | the
0.875	-12.8807	mining | translations
0.0441176	-12.8808	to be | the
0.55	-12.8808	(sag | al.,
0.36	-12.8809	are | informative
0.0526316	-12.881	the | was | to
0.0825397	-12.8811	where | is | in
0.875	-12.8812	sentence | (2),
0.875	-12.8812	ranking | relevance
0.578947	-12.8812	one | challenges
0.481481	-12.8813	collect | each
0.692308	-12.8813	epoch | is
0.16	-12.8814	pipeline | the
0.55	-12.8814	compare | results | different
0.304348	-12.8816	the | container
0.55	-12.8816	semantic | 1998).
0.3125	-12.8817	the | issue.
0.692308	-12.8817	resulting | form
0.0730897	-12.8818	results | with
0.206667	-12.8818	the | guide
0.352941	-12.8819	automatically | into
0.416667	-12.8819	determine | given
0.218045	-12.882	prediction | for
0.25	-12.882	the | taxonomy
0.451613	-12.882	for | metric,
0.168142	-12.882	we | candidates
0.333333	-12.8822	for | symbols
0.229508	-12.8822	which | related
0.451613	-12.8822	examples | include
0.202532	-12.8824	the | date
0.295775	-12.8825	system | documents
0.155894	-12.8825	a | just
0.197531	-12.8826	word | words | words
0.692308	-12.8827	such that | every
0.692308	-12.8827	when | agenda
0.110899	-12.8827	a | sentence.
0.428571	-12.8829	relations | them.
0.155894	-12.883	only | have
0.16318	-12.883	we | our approach
0.0982143	-12.8831	the | decision
0.293333	-12.8832	terms, | and
0.126866	-12.8833	2003; | et
0.185185	-12.8834	implemented | a
0.464286	-12.8835	concrete | is
0.692308	-12.8835	second, | multiple
0.185792	-12.8835	you | to
0.4	-12.8837	compared | different
0.295775	-12.8838	have | common
0.692308	-12.8838	from | january
0.304348	-12.8839	model | distributions
0.174528	-12.884	f-measure | the
0.8	-12.884	incorporated | model.
0.221374	-12.8842	illustrate | of
0.55	-12.8842	guidelines, | a
0.157088	-12.8842	(i.e., | and
0.1625	-12.8844	these | other
0.625	-12.8844	2nd | july
0.211268	-12.8844	to | traditional
0.692308	-12.8845	capturing | their
0.8	-12.8845	our | ngrams
0.0709828	-12.8847	sentence | is
0.390244	-12.8848	manner, | of
0.0813594	-12.885	hypothesis | the
0.134454	-12.885	in | there are
0.192982	-12.8851	e | f
0.117521	-12.8851	they | or
0.277108	-12.8852	the | accordingly.
0.416667	-12.8853	as | decreases
0.481481	-12.8853	performed | analysis
0.168889	-12.8853	has | number
0.464286	-12.8854	but | slightly
0.31746	-12.8855	our | sets.
0.257732	-12.8856	to | errors,
0.55	-12.8856	for | spanish.
0.377778	-12.8856	g. | and
0.224	-12.8857	knowledge | or
0.152174	-12.8857	the | vectors.
0.0700685	-12.8858	computational linguistics | of
0.727273	-12.8861	franz | and
0.464286	-12.8862	within | text
0.266667	-12.8862	seeds | the
0.125307	-12.8863	and | applied
0.361702	-12.8863	features | help
0.416667	-12.8864	participants, | and
0.211268	-12.8864	acts | a
0.2	-12.8864	our | experiments
0.232759	-12.8865	the | comment
0.428571	-12.8865	are | inside
0.170507	-12.8866	five | in
0.107335	-12.8866	measures | and
0.0931174	-12.8867	figure | shows
0.5	-12.8867	finding | their
0.692308	-12.887	error | incorrect
0.202614	-12.8872	the | compositional
0.0987842	-12.8873	the | prediction
0.4375	-12.8875	various | their
0.625	-12.8876	coefficient | coefficient
0.245283	-12.8876	sentence: | the
0.242991	-12.8877	2.3 | in
0.8	-12.8877	w | word;
0.8	-12.8877	manually | labeled
0.277108	-12.8878	the | match,
0.352941	-12.8879	where | assigned
0.727273	-12.888	stories | shift
0.289474	-12.888	directly | or
0.148789	-12.8881	the | gigaword
0.216418	-12.8881	our | simple
0.198758	-12.8883	regulation | regulation
0.304348	-12.8883	we | draw
0.416667	-12.8885	however, | larger
0.219697	-12.8885	and | others
0.289474	-12.8887	of | conceptual
0.234783	-12.8888	from | simple
0.0597015	-12.8889	3 | and
0.137725	-12.8889	the | approximately
0.152727	-12.8889	of | interest
0.8	-12.889	questions | did
0.727273	-12.8891	knowledge | resolution,
0.137313	-12.8891	if | same
0.211268	-12.8891	the | interaction.
0.578947	-12.8893	name | pairs
0.588235	-12.8894	2007) | 2009).
0.625	-12.8895	outperform | results
0.692308	-12.8895	regulation | binding
0.0829741	-12.8895	a | distribution
0.4	-12.8897	methods, | are
0.266667	-12.8898	of | occurred
0.285714	-12.8898	to | conventional
0.219697	-12.8898	the | sense.
0.361702	-12.8898	the | independent.
0.692308	-12.8899	example | illustrates
0.138138	-12.89	means | is
0.307692	-12.8902	systems | best
0.625	-12.8902	triggers | their
0.875	-12.8902	& | (eds.),
0.8	-12.8902	are | parallel,
0.155303	-12.8903	a | concepts
0.464286	-12.8905	have | clear
0.245283	-12.8905	is | to capture
0.578947	-12.8905	which | independently
0.151079	-12.8906	moreover, | to
0.352941	-12.8906	reduced | by
0.196319	-12.8906	is | expression
0.692308	-12.8906	in | equation,
0.8	-12.8906	shows | advantage
0.10582	-12.8908	the | making
0.118681	-12.891	of | main
0.55	-12.891	number | clusters.
0.875	-12.8911	reversible | attribute-value
0.361702	-12.8911	if | already
0.322034	-12.8913	the | task:
0.258065	-12.8913	each | category
0.31746	-12.8914	(2001) | of
0.118421	-12.8914	word, | the
0.588235	-12.8915	illustrated | table
0.8	-12.8915	log-likelihood | given
0.521739	-12.8917	lessons | and
0.377778	-12.8917	is | integer
0.4375	-12.8918	data | training,
0.5	-12.8918	process | document
0.405405	-12.8919	4.3 | each
0.4	-12.8919	and | probabilities,
0.481481	-12.892	for | f,
0.875	-12.892	a|=i
0.255102	-12.8921	the | smallest
0.230769	-12.8921	challenge | and
0.0334378	-12.8922	this | a | of
0.405405	-12.8923	from | papers
0.361702	-12.8923	note that | any
0.100471	-12.8924	time | to
0.4	-12.8924	event | temporal
0.185792	-12.8924	evaluations | the
0.588235	-12.8924	but | here.
0.727273	-12.8925	some | include:
0.266667	-12.8925	our | results.
0.625	-12.8925	proposed | address
0.141066	-12.8926	of | paraphrase
0.578947	-12.8926	to | vertices.
0.0857474	-12.8926	that | better
0.295775	-12.8927	accuracy | performance
0.481481	-12.8928	system | identifying
0.16318	-12.8928	we | build
0.55	-12.893	propose | technique
0.405405	-12.8931	qg | and
0.270588	-12.8932	evaluation | systems
0.578947	-12.8932	into | groups.
0.0821053	-12.8933	a | algorithm
0.19186	-12.8933	the | updated
0.55	-12.8933	makes | less
0.167401	-12.8934	this | previous
0.31746	-12.8934	handling | and
0.195122	-12.8934	of | noise
0.115546	-12.8935	design | the
0.642857	-12.8935	model | semantics.
0.0602615	-12.8936	and | information
0.255102	-12.8937	of | concepts.
0.0682148	-12.8938	we used | the
0.625	-12.8938	(2009) | problem
0.625	-12.8938	semantic | wordnet.
0.377778	-12.8938	this | repeated
0.171296	-12.8939	in | &
0.416667	-12.8939	in | regions
0.236842	-12.894	algorithm, | a
0.169725	-12.894	an | score
0.135447	-12.8943	the | strings
0.339623	-12.8944	using | toolkit
0.451613	-12.8948	unsupervised | 2010).
0.692308	-12.8949	both | and,
0.333333	-12.8954	however, | need
0.151079	-12.8955	methods | are
0.642857	-12.8955	esl | native
0.307692	-12.8956	parse, | the
0.8	-12.8956	unlike | methods
0.170507	-12.8957	additional | that
0.2	-12.8957	: | →
0.322034	-12.8958	of | debate
0.352941	-12.8959	on | domains.
0.405405	-12.8959	parsing, | with
0.5	-12.896	approach | name
0.266667	-12.8961	and | look
0.224	-12.8963	tagger | a
0.346154	-12.8964	is | expressing
0.205298	-12.8964	in | (as
0.119369	-12.8966	determined | the
0.106195	-12.8967	used | our
0.277108	-12.8967	set | features.
0.145763	-12.8968	models | language
0.267442	-12.897	the | specification
0.727273	-12.897	explore | within
0.588235	-12.897	types | features:
0.416667	-12.8971	powerful | for
0.521739	-12.8971	conventions | to
0.578947	-12.8971	these | labels.
0.134286	-12.8972	the | modifier
0.8	-12.8973	xi | xj
0.8	-12.8973	source-to-target | target-to-source
0.8	-12.8973	co-training | co-training
0.5	-12.8973	at | core
0.218045	-12.8974	context | an
0.3	-12.8975	we call | and
0.390244	-12.8975	higher | better
0.369565	-12.8976	were | more than
0.588235	-12.8976	in | place,
0.346154	-12.8977	a | hidden markov
0.104811	-12.8977	algorithm | a | of
0.107861	-12.8977	is | value
0.727273	-12.8978	& | a.
0.481481	-12.8978	test | wsj
0.380952	-12.8979	sentences | containing
0.0752688	-12.8979	this | was
0.625	-12.898	between | reference.
0.117391	-12.898	measured | the
0.369565	-12.8981	lms | to
0.875	-12.8983	advantage | ability
0.247525	-12.8983	the | phase | a
0.171296	-12.8984	list of | in
0.390244	-12.8984	features | presence
0.451613	-12.8986	that | away
0.464286	-12.8987	distance between | two
0.0841751	-12.8988	therefore | the
0.17734	-12.8988	views | the
0.240741	-12.8988	metrics, | the
0.692308	-12.8989	kernels | structures
0.642857	-12.8989	antecedent | pronoun
0.481481	-12.899	in | precedes
0.203947	-12.899	a | reduction
0.202614	-12.8991	and | quite
0.588235	-12.8991	approach | extraction.
0.3125	-12.8993	crowdsourcing | of
0.122642	-12.8993	action | the
0.625	-12.8993	measured | performance
0.110057	-12.8994	of | containing
0.333333	-12.8994	function | its
0.588235	-12.8994	rte | with
0.232759	-12.8995	finally, | from
0.127226	-12.8995	f-score | the
0.481481	-12.8995	after | step
0.481481	-12.8995	in | presented.
0.333333	-12.8998	instantiated | the
0.361702	-12.8998	using | resources
0.55	-12.9001	model | long-distance
0.0721311	-12.9002	such | is
0.642857	-12.9002	grammatical | linguistic
0.692308	-12.9003	explored | use
0.0532438	-12.9004	of | sentences
0.0918775	-12.9005	with | features
0.3125	-12.9006	lemma | in
0.197531	-12.9006	best | model
0.258065	-12.9006	joint | by
0.481481	-12.9008	datasets | two
0.196319	-12.9008	this | considered
0.380952	-12.9009	we are interested | that
0.240741	-12.9011	for | greater
0.352941	-12.9012	system | f-score
0.390244	-12.9012	for | write
0.625	-12.9012	alexandre | and
0.369565	-12.9012	represented | vector
0.291667	-12.9012	with | unique
0.354167	-12.9014	each | split
0.193939	-12.9015	contain | that
0.225806	-12.9016	and | domains
0.190751	-12.9016	results, | and
0.369565	-12.9016	translated | translation
0.380952	-12.9017	a | correction.
0.234783	-12.9017	with | few
0.15625	-12.9017	users | a
0.0898588	-12.9018	extracted | and
0.369565	-12.9018	english, | are
0.122353	-12.9018	and | 2003;
0.238532	-12.9019	member | the
0.322034	-12.9019	then | best
0.5	-12.902	set | s,
0.405405	-12.9022	the | was chosen
0.578947	-12.9022	0.77 | 0.77
0.3125	-12.9023	are | special
0.451613	-12.9024	people | different
0.55	-12.9024	structure | specific
0.216418	-12.9025	language models | in
0.5	-12.9025	some | them.
0.270588	-12.9026	this | particularly
0.369565	-12.9026	all | information.
0.263736	-12.9027	simple | using
0.327586	-12.9028	all | per
0.369565	-12.9028	even | where
0.107078	-12.9029	in | (e.g.,
0.692308	-12.9032	please | me
0.220472	-12.9033	it | what
0.154135	-12.9034	speech | for
0.0497259	-12.9035	to | their
0.17561	-12.9035	the | constrained
0.30303	-12.9035	for | instances.
0.642857	-12.9036	guided | which
0.5	-12.9036	around | words
0.282051	-12.9037	scenario | a
0.0629723	-12.9038	is | between | and
0.214815	-12.904	the | operator
0.236364	-12.904	pair | two
0.168182	-12.904	in | senses
0.153558	-12.9041	stack | the
0.4375	-12.9041	if | element
0.481481	-12.9041	we can | values
0.588235	-12.9041	(with | (with
0.242991	-12.9041	problem | using
0.55	-12.9042	also | feature.
0.0964392	-12.9042	of | current
0.692308	-12.9043	multiple | combined
0.405405	-12.9044	extended | set
0.0386884	-12.9045	a | that | a
0.173077	-12.9045	the | is:
0.369565	-12.9045	until | or
0.55	-12.9045	shown | perform
0.128272	-12.9046	nature | the
0.625	-12.9048	overall | comparison
0.625	-12.9048	identifying | important
0.114108	-12.9049	that | syntactic
0.184783	-12.9051	dialogues | the
0.20979	-12.9051	that | quite
0.451613	-12.9052	two | same | they
0.174757	-12.9053	and | ones
0.521739	-12.9053	rouge-2 | and
0.263736	-12.9053	the | modules.
0.109827	-12.9055	figure 1: | the
0.19186	-12.9055	sentence, | is
0.183784	-12.9055	of | information,
0.27381	-12.9056	is | yet
0.55	-12.9056	tag | tagging
0.25	-12.9057	describing | and
0.361702	-12.9059	to | your
0.201299	-12.906	when | other
0.226891	-12.906	with | approaches
0.339623	-12.9061	relations | entities
0.0912698	-12.9062	finally, | of
0.5	-12.9062	translator | a
0.0779596	-12.9064	the | including
0.252525	-12.9064	the | token,
0.727273	-12.9065	a | introduced.
0.140625	-12.9065	none | the
0.875	-12.9066	lower | low
0.26087	-12.9066	showing | a
0.133903	-12.9066	to | systems.
0.464286	-12.9067	of | own.
0.270588	-12.9067	model | significant
0.30303	-12.9067	are | partial
0.100478	-12.9068	our | two
0.238532	-12.9069	variance | the
0.380952	-12.9071	utterances | system
0.380952	-12.9071	of | transitions.
0.55	-12.9071	significantly | random
0.380952	-12.9073	was | corpus,
0.226891	-12.9073	0.4 | 0.6
0.109405	-12.9073	that | those
0.128272	-12.9075	the | needed
0.481481	-12.9076	or | methods.
0.354167	-12.9078	these | provided
0.131148	-12.9079	we | against
0.208333	-12.9081	tokens | are
0.588235	-12.9081	this | constitutes
0.263736	-12.9082	obvious | to
0.127937	-12.9083	of | 2007).
0.5	-12.9083	from | literature.
0.727273	-12.9084	predicates | arguments,
0.27381	-12.9085	no | any
0.5	-12.9085	judges, | of
0.405405	-12.9086	in | forms,
0.307692	-12.9087	chris | and
0.226891	-12.9087	stem | the
0.369565	-12.9088	our | filtering
0.195122	-12.909	union | the
0.159184	-12.9091	of | dependencies
0.155642	-12.9091	however, | our
0.424242	-12.9093	to | sentential
0.588235	-12.9093	reader | or
0.275	-12.9094	with | morphological
0.226891	-12.9096	a. | and
0.160494	-12.9096	a | towards
0.159184	-12.9098	while | an
0.0995261	-12.9098	the | five
0.481481	-12.9099	had | score
0.30303	-12.9099	maximal | of
0.727273	-12.91	general | summaries.
0.153558	-12.91	of | product
0.0859835	-12.9101	with | data
0.416667	-12.9102	of | location.
0.222222	-12.9102	by | patterns
0.222222	-12.9103	simultaneously | the
0.252525	-12.9103	is | interest
0.354167	-12.9104	calculate | between
0.202614	-12.9107	that | consistent
0.875	-12.9107	out-of-domain | performance
0.307692	-12.9108	the | earlier.
0.0395424	-12.9108	use | to
0.21875	-12.9109	expectations | the
0.4375	-12.911	accuracy, | with
0.191617	-12.9111	however, | other
0.122596	-12.9113	the | start
0.138889	-12.9113	for | case
0.278481	-12.9113	an | edge
0.55	-12.9114	multiple | including
0.361702	-12.9116	which | decision
0.275	-12.9117	transformation | to
0.588235	-12.9118	reduces | time
0.4375	-12.9119	the | be:
0.30303	-12.9119	which | identify
0.110672	-12.912	threshold | the
0.291667	-12.912	only | found
0.464286	-12.9121	paraphrases | bilingual
0.588235	-12.9121	and | lengths.
0.161157	-12.9124	beyond | of
0.4375	-12.9124	will | techniques
0.369565	-12.9125	have | coverage
0.369565	-12.9125	it | easier
0.316667	-12.9126	unfortunately, | is
0.322034	-12.9127	distribution over | is
0.394737	-12.9128	or | expression
0.307692	-12.9129	with | fixed
0.201299	-12.9129	only | if
0.727273	-12.913	useful | applications,
0.521739	-12.913	several | approaches
0.521739	-12.913	motivation | using
0.875	-12.913	it | conceivable
0.125313	-12.9133	the | has | been
0.727273	-12.9134	proceedings | conference.
0.133523	-12.9134	relative | in
0.642857	-12.9134	decisions | decisions
0.138462	-12.9135	in | tasks
0.190751	-12.9135	we | potential
0.287671	-12.9135	are | real
0.116129	-12.9137	and | among
0.171429	-12.9137	a | disambiguation
0.191617	-12.9138	and | evaluation.
0.182796	-12.9138	this | rule
0.30303	-12.9139	for | sufficient
0.55	-12.914	biggest | is
0.220472	-12.9141	the | continuous
0.625	-12.9141	while | links
0.642857	-12.9141	that | optimizes
0.130081	-12.9142	original | in
0.140575	-12.9142	classifiers | and
0.369565	-12.9142	the | 1986).
0.186441	-12.9143	2007), | al.,
0.202614	-12.9144	experiments | different
0.1875	-12.9144	only | when
0.875	-12.9144	original | creating
0.242718	-12.9145	to | writing
0.220472	-12.9146	i | .
0.179894	-12.9146	produces | of
0.216418	-12.9146	is | resource
0.307692	-12.9147	again, | is
0.327586	-12.9149	the | given.
0.875	-12.9149	outperform | baseline.
0.198718	-12.915	are | words,
0.255319	-12.9151	the | array
0.245098	-12.9152	by | points
0.122892	-12.9156	a | k
0.176768	-12.9156	the | dimension
0.521739	-12.9157	applying | method
0.625	-12.9157	documents | retrieval
0.316667	-12.9157	although | also
0.247525	-12.9157	tried | of
0.390244	-12.9158	of | high,
0.109195	-12.9158	this | task
0.625	-12.9161	an | environment
0.52381	-12.9162	cfg | hpsg
0.875	-12.9163	heads | mentions
0.26087	-12.9165	(12) | the
0.153257	-12.9166	the | encoded
0.133144	-12.9167	section 3 | the
0.10453	-12.9169	different | are
0.165919	-12.917	the | bias
0.140575	-12.917	that | meaning
0.333333	-12.9171	et | 1993).
0.448276	-12.9171	able | improve
0.55	-12.9172	other | languages,
0.150183	-12.9173	: | =
0.361702	-12.9173	each | cluster.
0.270588	-12.9174	under | model
0.0609683	-12.9174	its | and
0.481481	-12.9175	algorithms | large
0.354167	-12.9175	of | world.
0.55	-12.9175	two | sets,
0.354167	-12.9177	for | category,
0.150735	-12.9179	we | directly
0.0941512	-12.918	pattern | the
0.346154	-12.9181	for | conversational
0.448276	-12.9181	deletion | deletion
0.3125	-12.9183	evaluate | models
0.214815	-12.9183	of | size.
0.4375	-12.9184	conversations | in
0.316667	-12.9184	we | γ
0.642857	-12.9185	manual | annotators
0.448276	-12.9185	where | (or
0.1875	-12.9187	hypothesis | for
0.30303	-12.9187	for | attributes
0.115385	-12.9188	choose | the
0.642857	-12.9188	approach | builds
0.191617	-12.9189	the | terminal
0.0533516	-12.9189	the | and | other
0.0830565	-12.9189	the | left
0.240741	-12.9189	segmentation | word
0.5	-12.9189	reviewers | feedback.
0.205479	-12.919	for | experiments,
0.322034	-12.9191	david | of
0.5	-12.9191	we | text-based
0.642857	-12.9192	concept, | concept
0.875	-12.9195	{ | otherwise
0.380952	-12.9198	present | (section
0.267442	-12.9199	have | value
0.0316568	-12.92	only | the
0.228814	-12.92	the | clusters,
0.275	-12.9201	gave | and
0.52381	-12.9201	concern | that
0.247525	-12.9202	demonstrated | in
0.175879	-12.9204	recent | on
0.278481	-12.9204	measuring | in
0.448276	-12.9204	attempting | a
0.111554	-12.9204	we | a | approach
0.140575	-12.9205	we | every
0.875	-12.9205	corpora | 2010.
0.109615	-12.9206	our | at
0.52381	-12.9206	polarity | blog
0.179894	-12.9208	the | limit
0.0815451	-12.9208	are | at
0.692308	-12.9208	say | about
0.147368	-12.9209	partially | the
0.30303	-12.9209	of | phenomena.
0.588235	-12.9209	comparing | those
0.261364	-12.9209	entries | are
0.625	-12.921	(from | (from
0.0261145	-12.921	all | the
0.201299	-12.9211	to | understand
0.275	-12.9212	geometric | the
0.875	-12.9214	|s|
0.225	-12.9215	for | questions
0.52381	-12.9215	similar | ours
0.278481	-12.9216	develop | that
0.30303	-12.9216	average | different
0.0910308	-12.9218	our | data
0.0474466	-12.9218	used | to
0.380952	-12.9219	a | point,
0.448276	-12.9219	alternate | of
0.424242	-12.9224	of | fragments,
0.170616	-12.9224	the | will have
0.192771	-12.9225	in | german
0.236364	-12.9225	a | bayesian
0.271605	-12.9227	3.4 | in
0.588235	-12.9227	replacement | using
0.135294	-12.9228	is | particular
0.283784	-12.9228	a | 1999).
0.145833	-12.9229	the | topic.
0.242718	-12.923	levels | in
0.222222	-12.923	hence, | to
0.118182	-12.9231	gain | the
0.30303	-12.9232	of | library
0.185393	-12.9233	a | highly
0.0925414	-12.9234	defined | a
0.234234	-12.9234	the | u.
0.5	-12.9234	perfect | not
0.137195	-12.9235	objective | to
0.48	-12.9237	δ | δ
0.190476	-12.9237	of | messages
0.183333	-12.9238	partial | in
0.380952	-12.9239	used | ranking
0.106422	-12.9239	candidate | a
0.282051	-12.9241	alignment, | and
0.188571	-12.9241	the | matrices
0.692308	-12.9241	td03 | td04
0.692308	-12.9241	snippets | snippets
0.692308	-12.9241	pitch | pitch
0.171429	-12.9242	systems, | and
0.0489969	-12.9242	set | to
0.875	-12.9242	• | id
0.205479	-12.9244	in | around
0.394737	-12.9244	the | scenes.
0.588235	-12.9246	is | are.
0.157895	-12.9247	objective | a
0.247525	-12.9248	numbers | for
0.354167	-12.9251	proposes | for
0.5	-12.9253	relations | entities.
0.52381	-12.9254	correction | errors
0.642857	-12.9254	workers | tasks
0.202703	-12.9255	(2005) | a
0.642857	-12.9257	weights | sum
0.448276	-12.9257	accidents | health
0.316667	-12.9258	there are two | in
0.372093	-12.9258	effect | performance
0.448276	-12.926	paul | and
0.875	-12.9261	without | precision.
0.234234	-12.9261	4.2 | with
0.108779	-12.9262	compared | and
0.4375	-12.9264	models | detection
0.0975232	-12.9266	we | accuracy
0.311475	-12.9266	does | provide
0.21875	-12.9267	we | lists
0.287671	-12.9267	and | 1996).
0.588235	-12.9268	overall | about
0.55	-12.9269	the | (eng.
0.48	-12.927	deals | of
0.252632	-12.9272	visual | in
0.188235	-12.9273	possible | which
0.354167	-12.9273	research | focused
0.30303	-12.9274	finds | a
0.175	-12.9275	a | e.g.
0.875	-12.9275	as | propagate
0.777778	-12.9276	harnessing | sources
0.190476	-12.9278	before | a
0.0963855	-12.9279	is | lexical
0.52381	-12.9279	constraints | inference
0.777778	-12.928	2. | 1,
0.258065	-12.928	4.1 | our
0.588235	-12.9281	identify | product
0.307692	-12.9283	for | topic.
0.282051	-12.9283	are | independent
0.322034	-12.9284	we examine | in
0.727273	-12.9285	keywords | documents
0.204082	-12.9285	while | use
0.158537	-12.9286	a | predicate
0.1875	-12.9287	six | of
0.127321	-12.9287	of | items
0.0987261	-12.9288	the | vocabulary
0.777778	-12.9288	component, | component
0.278481	-12.9289	a | question.
0.875	-12.9289	ann | mi
0.106814	-12.9291	that | could
0.405405	-12.9291	it | ordering
0.346939	-12.9291	while | even
0.316667	-12.9292	on | classification.
0.264368	-12.9292	was | because
0.642857	-12.9292	dataset | movie
0.173267	-12.9295	have | some
0.207143	-12.9295	explicit | in
0.55	-12.9296	such | framework
0.777778	-12.9297	wpp | target
0.103327	-12.9298	the | scoring
0.258427	-12.9298	in | earlier
0.106814	-12.9299	the | calculated
0.875	-12.9299	exploits | information
0.875	-12.9299	several | exist
0.099359	-12.93	this | from | to
0.247525	-12.93	results | higher
0.48	-12.9301	this | alone
0.48	-12.9301	this | compact
0.255319	-12.9301	relevance | to
0.875	-12.9303	christoph | and
0.234234	-12.9303	a | pronoun
0.48	-12.9304	rules | minimal
0.0769971	-12.9304	and | new
0.372093	-12.9304	level | performance
0.161017	-12.9304	small | for
0.727273	-12.9305	expert | experts
0.777778	-12.9305	trains | classifier
0.777778	-12.9305	biomedical | clinical
0.106422	-12.9305	in this paper, | and
0.188235	-12.9306	and | aspects
0.464286	-12.9307	related | textual
0.777778	-12.9309	kim | kim
0.777778	-12.9309	module, | module
0.777778	-12.9309	dialogue | traum,
0.642857	-12.9309	transitioning | state
0.121655	-12.9311	is | small
0.264368	-12.9311	a | user’s
0.777778	-12.9313	if | again.
0.361702	-12.9314	show that | best
0.394737	-12.9314	that | unrelated
0.221311	-12.9315	roughly | and
0.121951	-12.9315	the | fully
0.4375	-12.9316	intended | which
0.333333	-12.9316	figure | shows | accuracy
0.380952	-12.9317	the | speed.
0.142857	-12.9318	the | λ
0.777778	-12.9318	response | response.
0.183333	-12.9319	matches | and
0.204082	-12.932	of | control
0.448276	-12.932	showed | could
0.52381	-12.9321	when | otherwise
0.464286	-12.9322	p, | is
0.0881612	-12.9322	not | this
0.55	-12.9322	in | media,
0.307692	-12.9323	if | appears
0.311475	-12.9323	have | significant
0.424242	-12.9324	used | calculate
0.405405	-12.9326	represent | same
0.167442	-12.9326	is | always
0.247423	-12.9327	propose | method | the
0.875	-12.9327	|e|
0.464286	-12.9329	document | summary
0.183333	-12.9329	the | probably
0.173267	-12.9332	has | different
0.225	-12.9334	marcu, | and
0.0772994	-12.9334	of | relations
0.55	-12.9334	incorporates | from
0.271605	-12.9335	using | models.
0.588235	-12.9335	largely | our
0.875	-12.9336	full | matrix
0.380952	-12.9337	the | θ,
0.189349	-12.9338	to | types.
0.48	-12.9338	seem | have
0.106227	-12.9338	with | semantic
0.464286	-12.9339	build | sentences
0.18232	-12.9339	a | around
0.333333	-12.9339	intuitively, | that
0.126984	-12.934	the | other.
0.642857	-12.934	constraints | models.
0.424242	-12.9342	each | document,
0.363636	-12.9342	human | scores
0.411765	-12.9342	experimental results, | the
0.258427	-12.9342	the third | is
0.346939	-12.9343	is | synonym
0.194969	-12.9343	the | to.
0.354167	-12.9344	for | articles.
0.198718	-12.9345	of | events.
0.157895	-12.9347	first, | is
0.267442	-12.9349	it | larger
0.134328	-12.935	the | providing
0.394737	-12.935	acl-hlt | or,
0.4375	-12.9351	of | hypotheses,
0.245098	-12.9351	responses | a
0.588235	-12.9351	consistency | annotation
0.15415	-12.9354	thus | is
0.163717	-12.9356	arcs | the
0.333333	-12.9357	correlation | judgments
0.226891	-12.9358	a | substitution
0.236364	-12.9359	aligned | word
0.0150312	-12.9359	the | are
0.354167	-12.9359	input | x
0.110442	-12.936	the | tagger
0.777778	-12.936	union | intersection
0.55	-12.9361	and | 2007.
0.287671	-12.9361	used | automatic
0.372093	-12.9362	of | ma
0.0243176	-12.9363	is | for
0.354167	-12.9363	one | see
0.411765	-12.9364	is | recursively
0.777778	-12.9364	reordering | mt.
0.133929	-12.9364	we | supervised
0.115721	-12.9367	is | specific
0.55	-12.9367	in | 5.3.
0.202703	-12.9368	of | vectors.
0.152091	-12.9369	grammatical | of
0.588235	-12.937	and | loss.
0.124365	-12.937	the | scope
0.727273	-12.9372	1997; | 2003;
0.165919	-12.9374	for each | word
0.04803	-12.9374	the | method
0.30303	-12.9374	any | they
0.363636	-12.9375	the | variant.
0.55	-12.9376	and | compression.
0.238095	-12.9378	score | 1
0.107143	-12.9378	improvements | the
0.48	-12.9379	give | results
0.156	-12.9379	a | nlp
0.380952	-12.938	the | sentence).
0.125984	-12.9382	model | used
0.316667	-12.9384	on | parsing.
0.115217	-12.9384	to | reference
0.55	-12.9385	using | hybrid
0.294118	-12.9386	a | thereby
0.145833	-12.9386	have | semantic
0.0471869	-12.9387	the | on | on
0.875	-12.9388	several | variants
0.144828	-12.939	they | an
0.22314	-12.9391	motivated by | the
0.4375	-12.9391	we define | function
0.291667	-12.9391	with | count
0.405405	-12.9392	features | information.
0.727273	-12.9395	5 results | discussion
0.333333	-12.9395	noted | that
0.146429	-12.9395	adjective | the
0.283784	-12.9395	structure | semantic
0.10118	-12.9395	that | very
0.777778	-12.9398	(balahur | al.,
0.875	-12.9398	have | statistics.
0.207143	-12.9398	produces | and
0.588235	-12.9399	pipeline | target
0.642857	-12.9399	languages | morphology
0.464286	-12.94	variability | for
0.159664	-12.94	+ | where
0.294118	-12.9402	confusion | for
0.777778	-12.9402	2010 | 2011
0.164444	-12.9403	a | markov
0.0518909	-12.9404	the | accuracy
0.114719	-12.9404	parameters | and
0.268293	-12.9405	the | markables
0.264368	-12.9406	that | distinct
0.727273	-12.9407	used | descriptions.
0.346939	-12.9407	performed | two
0.875	-12.9407	corpus, | corpora.
0.588235	-12.9408	answering | question
0.252632	-12.9409	a | page
0.52381	-12.9409	at | starting
0.642857	-12.941	syntactic | rules.
0.21374	-12.9411	partition | the
0.114719	-12.9412	loss | the
0.875	-12.9412	english/hindi | english/arabic
0.875	-12.9412	gameloser | teamfinalscore
0.875	-12.9412	0.74 | 0.8
0.875	-12.9412	25% | 40%
0.875	-12.9412	cross-cultural | persuasion
0.875	-12.9412	83 | 87
0.156627	-12.9413	in | 2010),
0.339623	-12.9413	probabilities | using
0.146953	-12.9413	do | on
0.298507	-12.9414	paraphrases, | the
0.169811	-12.9414	and | queries
0.0572193	-12.9416	approach | to
0.411765	-12.9416	and | me
0.316667	-12.9417	presented | system
0.642857	-12.9417	examples | pool
0.113734	-12.9419	of | local
0.219512	-12.942	the | inserted
0.0830489	-12.9421	the | observed
0.52381	-12.9421	of | locations.
0.333333	-12.9422	a | unannotated
0.100334	-12.9422	mapping | the
0.424242	-12.9423	be | design
0.424242	-12.9423	results | accurate
0.154762	-12.9423	paths | the
0.133136	-12.9423	produce | of
0.0779221	-12.9424	of | verb
0.464286	-12.9425	accuracy | majority
0.48	-12.9426	acknowledgments | grants
0.105072	-12.9426	graph | a
0.146429	-12.9426	scotland, | july
0.52381	-12.9426	loss | use
0.116071	-12.9426	the | child
0.727273	-12.9427	& | p.
0.298507	-12.9427	words | known
0.169903	-12.9429	methodology | of
0.178947	-12.943	of | kind
0.55	-12.943	model | essentially
0.247423	-12.9431	a | prefix
0.101868	-12.9431	a | knowledge
0.0928571	-12.9431	present | to
0.30303	-12.9432	then | following
0.5	-12.9434	der | und
0.777778	-12.9436	using | tool.
0.48	-12.9437	annotators | agreement
0.5	-12.9437	all | positions
0.555556	-12.9438	20% | 100%
0.124682	-12.9438	for each | are
0.261364	-12.9438	moving | the
0.100334	-12.9439	by | their
0.131805	-12.9441	detailed | of
0.642857	-12.9441	(pang | 2002;
0.152091	-12.9442	assignment | the
0.252632	-12.9442	for | evaluating
0.52381	-12.9443	needed | order
0.234234	-12.9444	the | covers
0.48	-12.9445	summary | content
0.642857	-12.9445	main | behind
0.210145	-12.9446	of | students
0.298507	-12.9446	in | constrained
0.235849	-12.9446	using | labeled
0.5	-12.9447	with | modification
0.242718	-12.9448	with | score.
0.363636	-12.9448	achieving | in
0.363636	-12.9448	adding | information
0.384615	-12.9449	all | contained
0.464286	-12.945	a | base,
0.48	-12.945	show | provide
0.157676	-12.9451	label | as
0.384615	-12.9451	results | directly
0.169903	-12.9452	we compute | of
0.208633	-12.9453	however, | two
0.113734	-12.9453	the | performing
0.777778	-12.9453	benefit | larger
0.190476	-12.9453	the | synonyms
0.5	-12.9453	gold | mentions
0.55	-12.9454	by | decoder.
0.411765	-12.9456	these | quite
0.234234	-12.9456	usually | as
0.0754006	-12.9456	for example, | to
0.258427	-12.9459	in | negation
0.464286	-12.946	based on | linear
0.181319	-12.946	graph | with
0.298507	-12.9461	of | predicate-argument
0.152091	-12.9461	procedure | to
0.159664	-12.9462	work | research
0.346939	-12.9462	a | computes
0.48	-12.9463	initial | but
0.55	-12.9463	both | versions
0.0684089	-12.9464	about | of
0.372093	-12.9464	set | unique
0.464286	-12.9465	complex | their
0.197368	-12.9465	generated | from
0.34	-12.9465	to | interpretation.
0.240385	-12.9466	current | or
0.384615	-12.9468	holding | and
0.424242	-12.9469	searn | a
0.252632	-12.9469	the | data;
0.588235	-12.9469	result | lower
0.333333	-12.947	parse | parser
0.25	-12.9471	average | score
0.179348	-12.9471	0.0 | 0.0
0.162996	-12.9471	for | classes
0.186047	-12.9476	the | conversion
0.283784	-12.9476	this | suggests
0.405405	-12.9478	classifiers | feature
0.727273	-12.9479	simulation | real
0.48	-12.9479	when | appear
0.0794148	-12.9479	in | while
0.287671	-12.9479	there | even
0.22314	-12.948	methodology | in
0.327273	-12.948	intelligent | and
0.298507	-12.9481	study | using
0.354167	-12.9481	explanation | is
0.306452	-12.9481	a | large-scale
0.165899	-12.9482	the | distributed
0.777778	-12.9482	generation | through
0.727273	-12.9483	conclusions | further
0.135802	-12.9483	that | good
0.363636	-12.9484	so, | are
0.642857	-12.9484	induced | lexicon
0.140523	-12.9486	also | features
0.327273	-12.9487	the | example).
0.642857	-12.9487	algorithm | iterations.
0.26506	-12.9488	generate | word
0.311475	-12.9488	vectors | each
0.116854	-12.9489	it | one
0.727273	-12.9491	so | simply
0.283784	-12.9491	is | −
0.727273	-12.9495	enforce | constraints
0.114471	-12.9495	the | fixed
0.289855	-12.9496	the | chain.
0.306452	-12.9496	an important | that
0.034067	-12.9496	can be | the
0.162996	-12.9498	were | into
0.287671	-12.9498	number | candidate
0.289855	-12.9499	the | lemma,
0.146429	-12.9499	terms | as
0.354167	-12.95	approach | task.
0.13871	-12.9502	for | including
0.187135	-12.9502	that | assigned
0.464286	-12.9504	while | it has
0.363636	-12.9504	left-hand | of
0.372093	-12.9504	shortest | to
0.384615	-12.9504	srl | semantic
0.777778	-12.9504	f | e.
0.247423	-12.9506	word | senses
0.138264	-12.9506	next | to
0.289855	-12.9507	paradigm | and
0.394737	-12.9507	of | combined.
0.219512	-12.9509	the | presentation
0.0661818	-12.9509	the | list
0.298507	-12.951	for | topic,
0.333333	-12.951	research | learning
0.555556	-12.9511	lemma | its
0.588235	-12.9511	in | antecedent.
0.588235	-12.9511	(clarke | al.,
0.346939	-12.9512	was | manually
0.131965	-12.9512	that | quality
0.130312	-12.9512	and | translations
0.103757	-12.9514	the | discussed
0.588235	-12.9514	1. | 2.
0.261364	-12.9514	algorithm, | in
0.261364	-12.9514	set | e
0.0871212	-12.9516	this | different
0.555556	-12.9517	later, | to
0.149254	-12.9519	that | must
0.5	-12.9519	readymades | the
0.333333	-12.9521	is | acoustic
0.165138	-12.9522	then | all
0.354167	-12.9522	only | sentences.
0.642857	-12.9522	bayesian | (goldwater
0.196078	-12.9523	heuristic | and
0.588235	-12.9524	vectors | represent
0.120482	-12.9525	our | into
0.114224	-12.9525	the | computing
0.372093	-12.9525	2010; | 2010)
0.217742	-12.9525	annotations | by
0.52381	-12.9527	cumulative | for
0.311475	-12.9527	consists | all
0.22807	-12.9529	style | and
0.346939	-12.9529	10 | were
0.777778	-12.9529	scales | data
0.642857	-12.9529	analysis | representations
0.26506	-12.9531	annotation, | and
0.181319	-12.9532	highest | for
0.464286	-12.9534	example | 2,
0.186047	-12.9534	and | online
0.588235	-12.9534	double | double
0.448276	-12.9535	five | different
0.433333	-12.9536	and | pipeline.
0.139159	-12.9537	∈ | =
0.187135	-12.9538	are | up
0.727273	-12.9539	dialogue | multimodal
0.15415	-12.9539	is | known
0.242424	-12.9541	which | current
0.242424	-12.9541	in | later
0.0714894	-12.9543	have | with
0.276316	-12.9543	(zhang | et
0.167442	-12.9543	to | tasks.
0.588235	-12.9544	solution | following
0.131965	-12.9545	prediction | of
0.143836	-12.9547	is | several
0.411765	-12.9547	sometimes | with
0.268293	-12.9548	approaches | using
0.0570652	-12.955	the | entity
0.134557	-12.955	used | translation
0.0674754	-12.955	for | but
0.354167	-12.9551	additional | into
0.411765	-12.9551	the | robot.
0.642857	-12.9551	ranges | 1
0.394737	-12.9551	there | potential
0.411765	-12.9553	dashed line | the
0.116071	-12.9553	for | topic
0.149254	-12.9556	and | constraints
0.372093	-12.9556	resource | language
0.52381	-12.9556	further | specific
0.177083	-12.9556	scenario | the
0.258427	-12.9556	great | to
0.48	-12.9558	bad | good
0.289855	-12.9558	helped | the
0.384615	-12.9559	a | service
0.111814	-12.9559	here | to
0.226087	-12.9559	encode | of
0.0883117	-12.956	of | previous
0.394737	-12.9561	hit | a
0.152941	-12.9562	to | potential
0.205674	-12.9563	statistically significant | the
0.175258	-12.9565	and | applications
0.158798	-12.9565	another | for
0.133929	-12.9565	other | which
0.188235	-12.9566	etc. | to
0.727273	-12.9567	news | time
0.247423	-12.9567	a | size.
0.363636	-12.9568	three | them
0.642857	-12.9568	charniak | stanford
0.271605	-12.9569	now | by
0.192547	-12.9569	is | manually
0.588235	-12.9569	year | language
0.261364	-12.9569	are | limited
0.555556	-12.957	ranked | order
0.588235	-12.9573	from | center
0.363636	-12.9574	rules | without
0.346939	-12.9575	best | reported
0.247423	-12.9578	line | with
0.333333	-12.9578	6 | 11
0.48	-12.9579	tuning | data
0.109756	-12.9579	annotation | for
0.448276	-12.9582	opinion | expansion
0.202797	-12.9582	in | lexicon.
0.0492929	-12.9583	that | an
0.196078	-12.9583	try | of
0.230088	-12.9584	and | cues
0.271605	-12.9584	the | relaxation
0.233645	-12.9585	decomposition | a
0.096124	-12.9585	of | corpora
0.5	-12.9585	since | complexity
0.268293	-12.9586	information, | is
0.363636	-12.9586	we use | linear
0.311475	-12.9587	of | configurations
0.424242	-12.9588	since | during
0.217742	-12.9588	found | were
0.5	-12.959	4.1 | three
0.555556	-12.9591	where | latter
0.224138	-12.9591	to | user’s
0.133739	-12.9592	the | window
0.642857	-12.9593	combine | form
0.258427	-12.9593	analogous | the
0.101724	-12.9594	of | data,
0.22807	-12.9594	methods | such
0.777778	-12.9594	n-grams, | n-grams
0.129213	-12.9595	training | using
0.100855	-12.9598	further | and
0.216	-12.9598	graph-based | and
0.327273	-12.9599	of | neighbors
0.087969	-12.9599	(e.g., | and
0.233645	-12.9601	this | times
0.136223	-12.9601	annotated | for
0.154472	-12.9601	a | paraphrase
0.355556	-12.9602	these | languages.
0.588235	-12.9602	ir | ir
0.283784	-12.9602	3.1 | model
0.727273	-12.9603	could | get
0.152344	-12.9604	the | above,
0.271605	-12.9605	particular | which
0.588235	-12.9605	quadratic | in
0.588235	-12.9605	joint | multiple
0.177419	-12.9605	above, | of
0.411765	-12.9606	used | define
0.727273	-12.9607	these | conversational
0.433333	-12.9608	our | system’s
0.0758157	-12.9608	available | the
0.171569	-12.961	as | sequence
0.283784	-12.961	formally, | =
0.126685	-12.9612	a | category
0.165899	-12.9612	the | format
0.182857	-12.9613	lm | a
0.283784	-12.9613	words | form
0.321429	-12.9614	gratefully acknowledge | the
0.52381	-12.9614	step | compute
0.448276	-12.9614	chodorow, | and
0.588235	-12.9615	after | selected
0.588235	-12.9615	the | rare.
0.167464	-12.9616	frames | the
0.192547	-12.9616	the | citation
0.411765	-12.9617	of | phenomena,
0.333333	-12.9617	has | type
0.172589	-12.9618	with | labeled
0.5	-12.9618	models | full
1	-12.9619	$ | morgen
1	-12.9619	salim | zhu.
1	-12.9619	adaptive | (ast)
1	-12.9619	csidm | csidm-200804
1	-12.9619	turney | littman
1	-12.9619	laboratory | cambridge,
1	-12.9619	no. | administered
1	-12.9619	del | natural,
1	-12.9619	angry | strained
1	-12.9619	(reichart | rappoport,
1	-12.9619	l-bfgs | nocedal,
1	-12.9619	00 | (test)
1	-12.9619	(bannard | brockett,
1	-12.9619	wordnet-affect: | extension
1	-12.9619	furniture | people
1	-12.9619	mc30 | rg65
1	-12.9619	chemical | covalent
1	-12.9619	(grishman | sundheim,
1	-12.9619	procesamiento | lenguaje
1	-12.9619	(ide | romary,
1	-12.9619	ptm+mdl | ve+mdl
1	-12.9619	dreyer | eisner
1	-12.9619	europarl, | commentary,
1	-12.9619	motorbike | shiny
1	-12.9619	d., | e.,
1	-12.9619	tsvetkov | wintner
1	-12.9619	<arg> | based_on
1	-12.9619	miller | charles
1	-12.9619	unzipped | gb
1	-12.9619	<hasprop> | villeroy_and_boch
1	-12.9619	esuli | sebastiani.
1	-12.9619	hypernym | hypernym
1	-12.9619	vorbei | früh
1	-12.9619	e|f | sym.
1	-12.9619	tiger | (brants
1	-12.9619	(andrew | gao,
1	-12.9619	policies | odni
1	-12.9619	[], | s2
1	-12.9619	conll-2011 | (pradhan
1	-12.9619	sind | miteinander
1	-12.9619	allege | aired.
1	-12.9619	maximize | w.r.t.
1	-12.9619	(grosz | sidner,
1	-12.9619	prosecutors | molested
1	-12.9619	twly | altwqye
1	-12.9619	nizar | columbia
1	-12.9619	scales | happy
1	-12.9619	leifern | ?
1	-12.9619	peking | 100871,
1	-12.9619	2010a; | 2010b).
1	-12.9619	carrillo | albornoz
1	-12.9619	(bunescu | mooney,
0.52381	-12.9619	entailment | predicates
0.272727	-12.9622	in | extraction,
0.52381	-12.9622	which | 2).
0.588235	-12.9622	translation | context-free
0.857143	-12.9624	(poon | vanderwende,
0.857143	-12.9624	chains | chains
0.857143	-12.9624	resnik, | chiang
0.857143	-12.9624	# | acc.
0.857143	-12.9624	right-hand | right-hand
0.857143	-12.9624	public | elderly
0.164384	-12.9624	parse | with
0.34	-12.9624	english | words.
0.448276	-12.9624	part, | to
0.16087	-12.9625	of | paraphrases
0.555556	-12.9625	had | effect
0.433333	-12.9625	use | hypothesis
0.327273	-12.9626	click | and
0.327273	-12.9626	solely | of
0.315789	-12.9626	content | each
0.315789	-12.9626	project | with
0.727273	-12.9627	dependent | head
0.180328	-12.9627	the | reach
0.212121	-12.9627	translations, | the
0.152344	-12.9627	the | 0.
0.0227607	-12.9627	the | of | by
0.219512	-12.9627	to | topics.
0.0899183	-12.9628	we | using | in
0.857143	-12.9629	counters | stream
0.857143	-12.9629	position | government,
0.857143	-12.9629	achieve | goal.
0.857143	-12.9629	increasing | size.
0.857143	-12.9629	solving | problems.
0.857143	-12.9629	taken | consideration,
0.186047	-12.9629	a | part-of-speech
0.0947368	-12.963	grammar | and
0.060219	-12.963	a | of | words
0.271605	-12.963	to | category.
0.202797	-12.963	over | data
0.192547	-12.9631	the | participant
0.149254	-12.9632	the | history
0.311475	-12.9632	recent work | to
0.424242	-12.9633	can be | existing
0.857143	-12.9633	backchannels | tts
0.857143	-12.9633	many | spending
0.857143	-12.9633	indicators’ | allophonic
0.727273	-12.9635	& | 2010)
0.327273	-12.9635	perform | training
0.333333	-12.9635	which | appropriate
0.6	-12.9635	(lee | al.,
0.182857	-12.9638	results | these
0.857143	-12.9638	string-to-tree | source-side
0.6	-12.9638	shift-reduce | parsing
0.5	-12.9638	• | e.g.
0.128134	-12.9639	of | vocabulary
0.26506	-12.9639	section 4 | in
0.204225	-12.9639	communication | and
0.219512	-12.9639	that | choice
0.289855	-12.964	algorithm 1 | for
0.777778	-12.9641	abstract | detect
0.588235	-12.9641	an | bias
0.0645161	-12.9641	classifier | the
0.261905	-12.9642	consider | two
0.394737	-12.9642	are | others.
0.857143	-12.9643	paraphrases | callison-burch
0.48	-12.9643	(rosti | al.,
0.136508	-12.9643	(for | the
0.222222	-12.9644	80% | the
0.666667	-12.9645	ill-formed | standard
0.411765	-12.9645	these | binary
0.34	-12.9647	given | previous
0.642857	-12.9647	mert | (och,
0.433333	-12.9647	occurrences | corpus
0.201389	-12.9648	word | sense
0.217742	-12.9648	a | mixture
0.666667	-12.9649	file | contains
0.152941	-12.9649	we | interested
0.179348	-12.965	first | was
0.144366	-12.9651	to | maximum
0.0945946	-12.9652	for | learning
0.666667	-12.9652	corpus | full-text
0.52381	-12.9652	build | graph
0.272727	-12.9654	same | also
0.375	-12.9654	the | translated,
0.5	-12.9654	low, | is
0.231481	-12.9656	improved | for
0.666667	-12.9656	deep | analysis
0.857143	-12.9657	martha | linguistics
0.461538	-12.9657	local | over
0.122137	-12.9657	we | examples
0.0735695	-12.9658	system | on
0.4	-12.9659	from | hypotheses
0.276316	-12.9659	this | long
0.0838253	-12.966	the | rate
0.244898	-12.9661	pairs | two
0.857143	-12.9662	intended message | salient
0.555556	-12.9662	intuitively, | model
0.448276	-12.9662	over | state-of-the-art
0.130435	-12.9663	the | attributes
0.346939	-12.9663	a | paraphrases.
0.777778	-12.9663	consider | segment
0.727273	-12.9664	integrated into | system
0.190184	-12.9664	mert | the
0.666667	-12.9664	as | observed,
0.588235	-12.9664	all | collected
0.235849	-12.9665	are | forms
0.372093	-12.9665	modelling | of
0.130058	-12.9666	is | simple
0.0325044	-12.9666	an | a
0.384615	-12.9666	experience | a
0.857143	-12.9667	related | “batter”
0.857143	-12.9667	either | or,
0.363636	-12.9667	0 | 1000
0.555556	-12.9668	performance | increases
0.208955	-12.9668	holds | the
0.588235	-12.9671	exist | this
0.777778	-12.9672	pθ(y|x)
0.0868455	-12.9672	defined | and
0.642857	-12.9672	models | competitive
0.178378	-12.9672	the | opposite
0.0217357	-12.9673	we | a | of
0.394737	-12.9675	when | just
0.727273	-12.9676	t-de | t-it
0.34	-12.9676	as | later
0.214286	-12.9676	robot | the
0.091954	-12.9677	on | but
0.120988	-12.9678	indicating | the
0.461538	-12.9678	slight | for
0.14652	-12.9678	we | by | each
0.289855	-12.968	useful | or
0.161435	-12.9681	can | some
0.48	-12.9681	being | also
0.116438	-12.9681	it | only
0.6	-12.9682	expects | to
0.461538	-12.9683	evaluate | impact
0.375	-12.9683	traditionally | of
0.272727	-12.9684	a | predicts
0.168269	-12.9685	out | with
0.141414	-12.9686	who | in
0.555556	-12.9687	meanings | their
0.25	-12.9687	however, | would
0.727273	-12.9688	ordering | binomial
0.122137	-12.9688	smaller | the
0.276316	-12.9688	what | this
0.137821	-12.9689	that | multiple
0.107422	-12.969	on | sentences
0.363636	-12.969	later | by
0.171717	-12.969	consistently | the
0.148148	-12.969	in | ranking
0.857143	-12.9691	local | long-distance
0.166667	-12.9691	one | use
0.311475	-12.9691	recent | 2010).
0.394737	-12.9691	the | returned.
0.252747	-12.9693	a | matched
0.276316	-12.9693	we | annotate
0.201389	-12.9693	and | run
0.448276	-12.9694	model | morphology
0.4	-12.9695	even | human
0.857143	-12.9696	conjunction | preposition
0.372093	-12.9696	was | four
0.231481	-12.9698	distributional | for
0.26506	-12.97	we | explain
0.276316	-12.9701	of | arguments,
0.0379839	-12.9701	using | of
0.306452	-12.9704	section 2. | and
0.142857	-12.9704	we find | of
0.52381	-12.9704	obviously, | is
0.174359	-12.9706	the | salient
0.219512	-12.9706	tag | as
0.315789	-12.9707	the | eventually
0.411765	-12.9707	which | degree
0.172589	-12.9709	they | used
0.448276	-12.9709	reliability | agreement
0.0875817	-12.971	do not | of
0.178378	-12.971	we | distributions
0.261905	-12.9712	have | linguistic
0.181818	-12.9712	output | by
0.4	-12.9713	with | internal
0.186747	-12.9713	of | relation.
0.301587	-12.9714	6 | evaluation
0.52381	-12.9716	four | most
0.0758017	-12.9717	this paper | of
0.242424	-12.9718	use | train
0.48	-12.9718	as | immediate
0.355556	-12.9719	the | case).
0.311475	-12.9719	using | derived
0.5	-12.9719	agreement | higher
0.433333	-12.9719	constitute | for
0.727273	-12.9721	decade, | has
0.145985	-12.9721	questions | and
0.666667	-12.9723	equation | product
0.289855	-12.9726	using | n-gram
0.4	-12.9726	to | with,
0.642857	-12.9726	iteratively | each
0.171717	-12.9727	n-gram | for
0.588235	-12.9727	demonstrated | performance
0.28	-12.9729	the | network,
0.375	-12.9729	accuracy | percentage
0.252747	-12.973	foreign | and
0.124011	-12.973	first, we | the
0.333333	-12.973	infinite | of
0.210526	-12.973	if | could
0.555556	-12.9731	disambiguation | gene
0.165877	-12.9733	in | vocabulary
0.0685805	-12.9733	the | across
0.121827	-12.9733	is | problem
0.0759494	-12.9734	the | four
0.375	-12.9734	barzilay | and
0.161435	-12.9735	only | used
0.173469	-12.9735	all | sentences
0.666667	-12.9735	we | 12.
0.384615	-12.9735	by | (2009).
0.113537	-12.9735	to | automatic
0.311475	-12.9735	these | indicate
0.170854	-12.9735	segmented | the
0.233645	-12.9737	an | space
0.384615	-12.9737	if | want
0.588235	-12.9737	of | new.
0.857143	-12.9739	dimensionality | reduced
0.28	-12.9739	note | when
0.085396	-12.9741	part of | the
0.107004	-12.9741	to | meaning
0.255556	-12.9741	features | previous
0.5	-12.9741	while | performs
0.289855	-12.9742	user, | the
0.108434	-12.9742	process | to
0.108	-12.9742	to | particular
0.202797	-12.9742	design | in
0.555556	-12.9743	in | cognates
0.6	-12.9743	allowing | tree
0.857143	-12.9744	experiments | xeon
0.642857	-12.9744	assumed | all
0.727273	-12.9745	section | treebank.
0.272727	-12.9747	theme | theme
0.306452	-12.9748	(snover | al.,
0.14	-12.9748	are | specific
0.52381	-12.9748	and | (2000).
0.857143	-12.9749	fitness | reference
0.555556	-12.9749	features | detect
0.26506	-12.975	middle | of
0.14386	-12.975	in | lexicon
0.10087	-12.975	the | trigger
0.190184	-12.9751	results for | are
0.48	-12.9751	since | binary
0.0817253	-12.9751	this | only
0.394737	-12.9751	from | brown
0.119512	-12.9751	thus | in
0.857143	-12.9753	sentences | leary’s
0.411765	-12.9753	using | discriminative
0.375	-12.9753	of | capabilities
0.0515837	-12.9753	we | features
0.14386	-12.9755	a | decision
0.642857	-12.9755	2005), | syntactic
0.461538	-12.9756	tweets | sentiment
0.6	-12.9756	added | additional
0.149425	-12.9758	of | functions
0.4	-12.9758	of | grammars,
0.0966719	-12.9758	1) | the
0.0884718	-12.9761	of | across
0.5	-12.9761	need | annotated
0.857143	-12.9763	english, | german,
0.384615	-12.9765	simplification | sentence
0.363636	-12.9767	detailed | on
0.106796	-12.9767	on | set.
0.2	-12.9768	students | and
0.4	-12.9771	the | witness
0.34	-12.9771	logarithm | the
0.0339247	-12.9771	are | for
0.666667	-12.9772	mention | mention.
0.276316	-12.9772	and | varying
0.777778	-12.9772	formal | expressions
0.857143	-12.9773	arabic | diacritics
0.034358	-12.9773	the | of | words
0.588235	-12.9773	outperforms | approaches
0.34	-12.9775	in | interaction.
0.5	-12.9775	funded by | through
0.384615	-12.9776	alternative | such
0.6	-12.9776	aware of | work
0.6	-12.9776	goal | predict
0.375	-12.9776	agreement | human
0.555556	-12.9777	manually | reference
0.321429	-12.9779	rule-based | for
0.384615	-12.978	strategy | using
0.588235	-12.978	that, | do
0.0552947	-12.9781	were | in
0.296875	-12.9781	in | brown
0.0685805	-12.9782	of | document
0.333333	-12.9783	at | local
0.588235	-12.9783	denotes | over
0.384615	-12.9784	improving | translation
0.104089	-12.9785	lm | the
0.777778	-12.9785	until | point
0.727273	-12.9786	previously | types
0.0671815	-12.9788	the | meaning | the
0.48	-12.9788	specifically, | following
0.0899032	-12.9788	values | and
0.139535	-12.9788	2.3 | the
0.285714	-12.9789	the | conditions,
0.315789	-12.9789	rules | side
0.141379	-12.9789	category | a
0.207407	-12.979	improving | and
0.555556	-12.979	top | baseline
0.555556	-12.979	method | disambiguation
0.19863	-12.979	sample | for
0.588235	-12.979	trigger | protein
0.857143	-12.9792	column | bold.
0.109731	-12.9792	and | show
0.28	-12.9793	growth | the
0.727273	-12.9794	code, | a
0.242424	-12.9794	the | aligning
0.777778	-12.9794	knowledge | crucial
0.0712424	-12.9795	algorithm | a
0.857143	-12.9797	overcome | limitations
0.201389	-12.9798	and | until
0.158798	-12.9798	algorithm | each
0.24	-12.98	system | text
0.6	-12.98	we address | issue
0.857143	-12.9802	(moore | 2010).
0.642857	-12.9802	is | animal
0.132931	-12.9803	variants | the
0.0909091	-12.9804	appear | the
0.411765	-12.9804	mixture | each
0.384615	-12.9804	could | better
0.384615	-12.9804	for | ≥
0.0244273	-12.9805	a | that | the
0.666667	-12.9806	classifying | sentiment
0.102888	-12.9806	for | et al.
0.857143	-12.9807	our | extractors
0.727273	-12.9807	such | essential
0.777778	-12.9807	context | neighboring
0.285714	-12.9808	rest | is | follows.
0.162896	-12.9809	is | tag
0.666667	-12.981	thus, | further
0.0777202	-12.981	and | performance
0.224138	-12.9811	decoding | with
0.306452	-12.9811	we | ten
0.0915594	-12.9813	a | node
0.169154	-12.9814	is | factor
0.15261	-12.9815	words | sentence
0.224138	-12.9815	this | annotations
0.433333	-12.9815	is | hypernym
0.229358	-12.9817	black | the
0.182857	-12.9819	after | from
0.207407	-12.9819	the | settings.
0.347826	-12.982	validate | of
0.14717	-12.982	techniques | in
0.111828	-12.982	abstract | for
0.208955	-12.982	to | external
0.857143	-12.9821	values | [0,1]
0.666667	-12.9821	hashtags | linguistic
0.224138	-12.9821	2.3 | a
0.231481	-12.9823	and | relationships
0.117371	-12.9824	accuracy | on
0.6	-12.9824	however, | solution
0.310345	-12.9826	model | information.
0.4	-12.9828	equivalents | in
0.857143	-12.9831	(zhang | 2006)
0.0991597	-12.9831	on | one
0.143369	-12.9831	the | reliable
0.355556	-12.9834	syntactic | rules
0.075217	-12.9834	number of | for
0.6	-12.9834	collected | pairs
0.289855	-12.9835	both | single
0.857143	-12.9836	carry | meaning.
0.666667	-12.9836	methods | expected
0.666667	-12.9836	implementation | space
0.123037	-12.9836	and | text.
0.375	-12.9837	permits | to
0.227273	-12.9838	this | just
0.24	-12.9838	constrained | of
0.642857	-12.9839	form | complete
0.727273	-12.984	non-terminal | non-terminal
0.0929204	-12.984	instead of | the
0.0942249	-12.9841	the | framework
0.347826	-12.9842	present | analysis
0.310345	-12.9842	are | exactly
0.289855	-12.9842	wil | the
0.375	-12.9842	select | sentences
0.642857	-12.9842	accuracies | data.
0.642857	-12.9842	property | one
0.4	-12.9843	systems | evaluated
0.247312	-12.9843	from | 2007).
0.173684	-12.9843	the | 2001).
0.419355	-12.9843	on | ...
0.727273	-12.9844	able | reach
0.17	-12.9844	that | learn
0.306452	-12.9844	cell | a
0.296875	-12.9844	if | <
0.48	-12.9846	fixed | set
0.176471	-12.9846	corpus | corpus.
0.642857	-12.9846	and | equal.
0.433333	-12.9847	and | writing.
0.433333	-12.9847	competitive | on
0.666667	-12.9848	1in | we
0.384615	-12.9848	achieve | results
0.235294	-12.9849	operations | of
0.0807175	-12.985	is | some
0.363636	-12.985	and | tetreault
0.0911681	-12.9851	the | mention
0.164319	-12.9851	set | set.
0.135647	-12.9852	stage | the
0.355556	-12.9854	thank | his
0.261905	-12.9854	achieved | by
0.34	-12.9854	and | error.
0.13253	-12.9854	used | features
0.237624	-12.9855	a | missing
0.333333	-12.9855	that | content.
0.355556	-12.9856	of | sentiments
0.252747	-12.9857	a | events.
0.34	-12.9858	hebrew | and
0.666667	-12.9859	arguments | identified
0.48	-12.9859	into | language,
0.26506	-12.986	there are | other
0.5	-12.986	retrieved | for
0.0872396	-12.9862	mention | the
0.252747	-12.9863	typical | is
0.276316	-12.9863	of | sequences.
0.195946	-12.9864	axioms | the
0.48	-12.9865	a | distribution:
0.6	-12.9865	in | inventory.
0.269231	-12.9868	stream | lm
0.255814	-12.9868	already | as
0.0899032	-12.9868	about | in
0.727273	-12.9869	mapping | mentions
0.26506	-12.9869	pass | a
0.777778	-12.9869	perplexity | models.
0.419355	-12.9869	approach | grammar
0.184524	-12.987	that | constraints
0.144928	-12.987	we | clustering
0.1	-12.9871	finally, | to
0.147727	-12.9871	used | test
0.347826	-12.9872	while | general
0.100173	-12.9873	our | only
0.48	-12.9873	elementary | nodes
0.165877	-12.9873	there | only
0.448276	-12.9874	better | models.
0.25	-12.9875	is | symbol
0.26506	-12.9875	a | code
0.247312	-12.9875	find | between
0.34	-12.9875	usability | of
0.642857	-12.9875	by | (2001).
0.355556	-12.9876	of | interactions.
0.52381	-12.9876	em | all
0.727273	-12.9877	c. | p.
0.188679	-12.9877	the | applies
0.48	-12.9878	three | matching
0.197279	-12.9879	that | match
0.255814	-12.988	with | five
0.347826	-12.9881	this | pipeline
0.192308	-12.9881	effects | and
0.15678	-12.9881	is | prior
0.127841	-12.9882	and | considered
0.411765	-12.9882	with | datasets.
0.168317	-12.9884	our | human
0.727273	-12.9885	tend | longer
0.555556	-12.9885	depth | tree
0.146617	-12.9885	a | via
0.433333	-12.9886	analysis | text.
0.355556	-12.9888	distinct | that
0.285714	-12.9889	results | tables
0.666667	-12.9889	national | computing
0.24	-12.9889	also | context
0.448276	-12.9889	in | collections.
0.555556	-12.9891	tree | tree,
0.777778	-12.9891	following | corpus:
0.461538	-12.9892	evaluation, | system
0.326923	-12.9893	of | baselines.
0.727273	-12.9894	despite | resources
0.180791	-12.9895	for | experiments.
0.301587	-12.9896	were | four
0.34	-12.9897	the | implied
0.301587	-12.9898	similar | also
0.6	-12.9899	list | word pairs
0.28169	-12.9899	system | systems.
0.107356	-12.99	(in | the
0.34	-12.99	for | classifier,
0.777778	-12.99	process | involves
0.642857	-12.9901	however | large
0.34	-12.9902	however, | make
0.857143	-12.9904	were | ].
0.857143	-12.9904	task | marsi,
0.375	-12.9904	experiment, | used
0.0549223	-12.9905	an | with
0.229358	-12.9906	of | narrative
0.48	-12.9908	these | become
0.433333	-12.9908	according | frequency
0.321429	-12.9911	furthermore, | or
0.202899	-12.9911	these | would
0.388889	-12.9912	has | years
0.255814	-12.9913	this | having
0.195946	-12.9913	grammatical | in
0.52381	-12.9915	fortunately, | and
0.384615	-12.9916	also | individual
0.6	-12.9916	rising | to
0.433333	-12.9916	the | long-forms
0.148289	-12.9917	made | in
0.207407	-12.9918	is | actual
0.205882	-12.9919	to | participants
0.347826	-12.9919	priors | and
0.15261	-12.9919	that | common
0.244681	-12.9919	binary | are
0.173684	-12.9919	the | others.
0.13354	-12.992	correction | the
0.321429	-12.992	2004; | 2009;
0.6	-12.992	which | states.
0.117506	-12.9921	discussion | of
0.123684	-12.9922	pairs | are
0.6	-12.9923	fit | data.
0.310345	-12.9924	an | answer
0.125	-12.9925	token | a
0.6	-12.9926	paired | test
0.132132	-12.9927	this | about
0.727273	-12.9927	r2 | r4
0.727273	-12.9927	framework | combined
0.666667	-12.9927	speakers | responses
0.0598396	-12.9927	the | sense
0.642857	-12.9927	hypotheses | systems
0.0929535	-12.9928	and | 2005;
0.555556	-12.9929	by | neutral
0.118812	-12.9929	portion | the
0.261905	-12.993	and | involving
0.347826	-12.9931	we | (we
0.115741	-12.9931	independent | the
0.0970395	-12.9932	large | in
0.244681	-12.9933	these | sets
0.244681	-12.9933	worst | the
0.137705	-12.9933	the | six
0.10757	-12.9934	then | =
0.5	-12.9934	in | works.
0.666667	-12.9935	parallel | (bannard
0.117506	-12.9935	we compute | the
0.218487	-12.9935	a | tree,
0.130178	-12.9937	were | with | and
0.202899	-12.9937	this | larger
0.388889	-12.9939	in | live
0.555556	-12.9939	computes | between
0.301587	-12.994	funded | grant
0.6	-12.994	overall | among
0.5	-12.994	plausible | for
0.261905	-12.9941	(in | as
0.34	-12.9943	only | labels
0.6	-12.9944	given | generates
0.555556	-12.9945	r. | d.
0.777778	-12.9945	bleu | en-de
0.0790043	-12.9946	in | data.
0.321429	-12.9946	improve | over
0.433333	-12.9948	that | aggregate
0.121134	-12.9949	to | models.
0.355556	-12.9949	and | grammars.
0.777778	-12.9949	compressed | compression
0.272727	-12.995	one | sentence.
0.48	-12.9952	study | english
0.182353	-12.9952	off | the
0.642857	-12.9952	once | corpus.
0.2	-12.9952	if | should
0.132716	-12.9952	computation | the
0.857143	-12.9953	bansal | (2010)
0.857143	-12.9953	results | promising,
0.777778	-12.9954	highly | highly
0.176796	-12.9956	is | rank
0.135048	-12.9956	just | of
0.186747	-12.9957	modified | and
0.34	-12.9957	it was | as
0.292308	-12.9957	cannot | it
0.857143	-12.9958	applied | structures.
0.347826	-12.9959	threshold | which
0.150198	-12.9959	is | process
0.326923	-12.996	texts | source
0.419355	-12.996	the | increased.
0.555556	-12.9961	substitutions | in
0.296875	-12.9962	genia | et
0.285714	-12.9965	problems, | and
0.229358	-12.9965	figure 4 | of
0.214876	-12.9965	20% | the
0.229358	-12.9966	lattice | a
0.5	-12.9966	requires | syntactic
0.0874172	-12.9967	proposed | to
0.777778	-12.9967	sender | recipient
0.857143	-12.9968	just | name
0.419355	-12.9968	good | human
0.326923	-12.9969	description | an
0.109278	-12.997	the | improves
0.252874	-12.997	s, | a
0.247312	-12.997	by | matrix
0.292308	-12.997	to | representations.
0.305085	-12.9971	of | first-order
0.25	-12.9971	increased | in
0.333333	-12.9971	model | θ
0.388889	-12.9972	further, | that
0.857143	-12.9973	recal | precision
0.34	-12.9973	a | dimensional
0.433333	-12.9973	investigation | that
0.191083	-12.9973	this | result
0.25	-12.9974	and | ones.
0.074215	-12.9975	as | more
0.777778	-12.9976	detecting | sentiment
0.321429	-12.9977	work | nlp
0.857143	-12.9978	provides | prediction
0.111354	-12.9978	the | seed
0.526316	-12.9979	with | probability,
0.172775	-12.9979	recently | and
0.261905	-12.998	the | substring
0.0971993	-12.998	the | confidence
0.223214	-12.998	a | t,
0.321429	-12.9981	equations | and
0.6	-12.9982	raw | text
0.642857	-12.9982	design | day
0.642857	-12.9982	(s | (vp
0.0812357	-12.9983	following | a
0.461538	-12.9984	experiments | run
0.225225	-12.9986	experiments | performance
0.355556	-12.9986	with | abstract
0.555556	-12.9987	parsing | parser.
0.292308	-12.9987	evaluation | against
0.857143	-12.9988	solve | problems.
0.109705	-12.9988	general | in
0.258824	-12.9988	used | improve
0.148438	-12.9989	select | to
0.216667	-12.9989	our | precision
0.365854	-12.9989	an | entailment
0.555556	-12.999	as | (zhang
0.13354	-12.9991	datasets | the
0.48	-12.9991	text | network
0.207692	-12.9991	2.3 | and
0.2	-12.9991	in | root
0.164319	-12.9991	latter | of
0.165854	-12.9992	is | users
0.555556	-12.9993	this | correlates
0.555556	-12.9993	pruning | used
0.384615	-12.9993	leaves | in
0.340426	-12.9994	these | directly
0.252874	-12.9994	the | with,
0.375	-12.9994	resource | which
0.375	-12.9994	a | unit.
0.666667	-12.9996	with | chosen.
0.191083	-12.9996	away | the
0.857143	-12.9998	pc(¯t|γ,
0.526316	-12.9998	main | being
0.269231	-12.9998	for | usage
0.0911722	-12.9998	this paper | to
0.5	-12.9998	with | authors.
0.140893	-12.9999	alternative | of
0.461538	-13	in | 9,
0.157205	-13.0001	subject | a
0.235294	-13.0001	from | treebank
0.117788	-13.0002	containing | of
0.301587	-13.0002	relation, | a
0.555556	-13.0003	but | hard
0.11639	-13.0003	a | random
0.174603	-13.0003	appears | in
0.301587	-13.0005	and | jointly
0.107071	-13.0006	another | is | the
0.0434068	-13.0006	the | of | their
0.207692	-13.0007	to | labels.
0.0401685	-13.0007	if | is | the
0.201439	-13.0009	that | require
0.315789	-13.0009	evaluation | metric
0.15859	-13.0011	interaction | and
0.365854	-13.0012	4.2 | two
0.347826	-13.0013	corpus | automatically
0.237624	-13.0013	by | local
0.320755	-13.0013	a | dependencies.
0.258824	-13.0014	approach | large
0.0910405	-13.0017	they | that
0.16	-13.0017	fit | the
0.176796	-13.0018	the | filtered
0.052381	-13.0018	the | function
0.375	-13.0018	the | decreases.
0.419355	-13.0018	3 | without
0.252874	-13.0019	ends | the
0.555556	-13.0019	5 conclusion | approach
0.174863	-13.0021	we | basic
0.777778	-13.0021	sentence | paraphrased
0.857143	-13.0022	figure 1: | simple
0.188679	-13.0022	viterbi | the
0.223214	-13.0022	it | still
0.189873	-13.0023	could | more
0.142857	-13.0024	sentence | word
0.108163	-13.0024	a | might
0.0282887	-13.0024	also | the
0.365854	-13.0025	data | gigaword
0.285714	-13.0026	bootstrapping | for
0.48	-13.0027	user: | confirmed
0.0261899	-13.0027	the | are | in
0.340426	-13.0029	the | tool.
0.252874	-13.0029	the | rate.
0.216667	-13.003	5.3 | to
0.461538	-13.0032	tg | =
0.214876	-13.0032	models | over
0.134185	-13.0032	an | corpus
0.141343	-13.0033	symbol | the
0.315789	-13.0033	however | that
0.272727	-13.0033	queries | are
0.305085	-13.0034	big | and
0.185629	-13.0034	the | morpheme
0.242105	-13.0035	describe | used
0.419355	-13.0035	genre | a
0.5	-13.0035	a | aproach
0.28169	-13.0035	introducing | in
0.20438	-13.0036	involve | and
0.108163	-13.0036	word | it
0.857143	-13.0037	hierarchy | generation
0.857143	-13.0037	full | unknown
0.6	-13.0037	are | meaning,
0.326923	-13.0037	obtains | of
0.258824	-13.0038	department | usa
0.433333	-13.0038	has | it.
0.340426	-13.0041	findings | that
0.433333	-13.0041	may | correct
0.355556	-13.0042	or | simply
0.14786	-13.0042	(3) | is
0.0973597	-13.0042	original | and
0.340426	-13.0043	in | forest.
0.2625	-13.0043	a | dimensions
0.114416	-13.0043	on | methods
0.301587	-13.0044	f-measure | for
0.6	-13.0044	our | stateof-the-art
0.0736544	-13.0045	the | assigned
0.150794	-13.0046	feedback | the
0.162037	-13.0047	limited | to
0.127119	-13.0048	the | reflect
0.152263	-13.0048	the | tree,
0.310345	-13.0049	or | another
0.301587	-13.0051	a | accuracy,
0.526316	-13.0052	with | heads
0.365854	-13.0053	contrasting | and
0.375	-13.0053	we | corrections
0.223214	-13.0054	that | greater
0.152263	-13.0054	in | (i.e.,
0.194631	-13.0055	assigned to | of
0.857143	-13.0057	character | processed,
0.166667	-13.0057	tool | and
0.419355	-13.0057	values | computed
0.138514	-13.006	the | fit
0.197279	-13.006	gold-standard | the
0.111842	-13.0061	the | reading
0.177778	-13.0061	reduce | to
0.857143	-13.0062	degrade | performance
0.355556	-13.0062	the | owl
0.235294	-13.0062	that | proper
0.25	-13.0063	and | terms.
0.285714	-13.0064	in | collection.
0.167488	-13.0064	documents | are
0.5	-13.0064	similarity scores | in
0.207692	-13.0065	for | relatively
0.157895	-13.0065	description | in
0.340426	-13.0067	paired | a
0.165049	-13.0068	translation | into
0.192053	-13.0069	and | constituent
0.777778	-13.0071	inference | variables.
0.209302	-13.0072	given | probability
0.213115	-13.0072	a | instances.
0.273973	-13.0073	number | classes
0.444444	-13.0073	system | predicted
0.25	-13.0076	b. | and
0.15102	-13.0077	first | all
0.230769	-13.0079	graphical | and
0.5	-13.0079	bidirectional | to
0.192053	-13.008	collect | of
0.777778	-13.008	split | parts.
0.195946	-13.008	in | response
0.172775	-13.0081	to | (or
0.355556	-13.0081	we | leverage
0.0275596	-13.0083	the | of | or
0.347826	-13.0084	has | question
0.12828	-13.0084	is | gold
0.6	-13.0086	w/o | w/
0.0637079	-13.0086	we | number of
0.285714	-13.0087	pronouns | to
0.365854	-13.0087	also | content
0.2625	-13.0088	system | out
0.223214	-13.0088	how | system
0.666667	-13.0089	report | results.
0.444444	-13.0089	also | making
0.555556	-13.009	subtree | by
0.244681	-13.009	the | correct,
0.127907	-13.0091	are | 1.
0.419355	-13.0091	is | reaches
0.0262461	-13.0091	one | the
0.857143	-13.0092	easy | implement.
0.310345	-13.0097	substantial | for
0.347826	-13.0098	thank | providing
0.152263	-13.0099	2006) | and
0.111842	-13.01	the | slightly
0.146067	-13.0101	to | point
0.478261	-13.0102	languages, | english
0.138983	-13.0102	constraints | a
0.777778	-13.0102	segmentor | using
0.235294	-13.0103	computes | of
0.340426	-13.0104	the | dump
0.0661538	-13.0104	algorithm | and
0.116114	-13.0104	in | processing
0.255814	-13.0104	in | years,
0.419355	-13.0104	words | various
0.193333	-13.0105	are | local
0.193333	-13.0105	to | actions
0.252874	-13.0107	the | parts.
0.355556	-13.0108	compared | same
0.666667	-13.0109	b. | a.
0.180233	-13.0109	the | desired
0.375	-13.0109	scored | for
0.292308	-13.0109	using | probabilities
0.6	-13.011	expressions | events
0.189873	-13.0111	the | covering
0.188679	-13.0111	language model | for
0.143911	-13.0112	and | learned
0.355556	-13.0112	it | component
0.555556	-13.0112	2005; | 2007)
0.213115	-13.0112	is | scoring
0.292308	-13.0112	by | sum
0.461538	-13.0113	category | its
0.25	-13.0114	in | reverse
0.244681	-13.0114	other | methods
0.6	-13.0114	2.1 | analysis
0.5	-13.0114	that | idiosyncratic
0.0189618	-13.0115	the | of | that
0.333333	-13.0116	corpus: | and
0.134615	-13.0116	(2009) | and
0.25	-13.0117	the | played
0.6	-13.0117	nullnullnull | nullnullnullnullnull
0.162037	-13.0117	level | for
0.292308	-13.0117	not | therefore
0.433333	-13.0119	also | appropriate
0.6	-13.0121	pruning | only
0.6	-13.0121	ontology | types
0.857143	-13.0122	edit | edit distance
0.526316	-13.0123	edge | vertex
0.305085	-13.0123	we | reasons
0.0433344	-13.0123	that | can
0.0230964	-13.0123	the | of | on
0.247312	-13.0126	scores | two
0.857143	-13.0127	?c a | .
0.857143	-13.0127	algorithm | similarities
0.247312	-13.0129	whether | same
0.28169	-13.0129	is | specified
0.388889	-13.013	constructed | using
0.0901288	-13.0131	the | utterances
0.526316	-13.0132	induced | senses
0.269231	-13.0132	extract | using
0.0554047	-13.0132	one | is | the
0.228571	-13.0132	the | answers.
0.305085	-13.0134	and | merging
0.5	-13.0134	semantic | predicates.
0.216667	-13.0135	developing | and
0.100358	-13.0135	predicted | the
0.666667	-13.0136	encode | knowledge
0.123967	-13.0136	is | domain
0.047505	-13.0139	and | have
0.235294	-13.014	helps | to
0.206107	-13.0141	that | true
0.478261	-13.0141	opinions | aspects
0.857143	-13.0142	42 | 45
0.857143	-13.0142	has | terms,
0.326923	-13.0142	genres | and
0.120823	-13.0143	processing | of
0.5	-13.0143	while | individual
0.214876	-13.0144	english | other
0.478261	-13.0144	a | period.
0.148438	-13.0145	the | frames
0.223214	-13.0145	a | arc
0.6	-13.0145	including | 2003),
0.296875	-13.0145	annotations | each
0.214876	-13.0147	properties | that
0.265823	-13.0147	the | tweet.
0.0825243	-13.0148	parameter | the
0.555556	-13.0148	23 | test
0.184049	-13.015	a | 100
0.0624571	-13.015	for | translation
0.155172	-13.015	the | f-measure
0.213115	-13.0151	it | correct
0.357143	-13.0152	given | small
0.146718	-13.0152	use | corpus
0.444444	-13.0153	word | otherwise
0.197183	-13.0156	policy | a
0.310345	-13.0157	introduce | model
0.347826	-13.0159	is | deterministic
0.326923	-13.0159	because | pairs
0.2	-13.0159	might | or
0.287879	-13.016	examples | used
0.333333	-13.0161	ece | the
0.857143	-13.0162	supports | development
0.209302	-13.0162	internal | of
0.227273	-13.0162	hierarchical | with
0.170984	-13.0163	control | of
0.388889	-13.0164	a | validity
0.104046	-13.0164	(i.e., | the
0.4	-13.0165	a | reasonably
0.526316	-13.0166	one-to-one | between
0.214876	-13.0166	of | ambiguity
0.365854	-13.0166	constructed | set
0.102421	-13.0167	examples | in
0.357143	-13.0167	language | lm
0.6	-13.0167	both | identical
0.14786	-13.0168	that, | and
0.188679	-13.0168	the | the system’s
0.28169	-13.0168	the | classifier’s
0.5	-13.0169	these | names
0.40625	-13.017	extracting | comparable
0.214876	-13.017	using | new
0.162679	-13.0172	label | with
0.235294	-13.0172	this | always
0.0603725	-13.0173	the | level
0.6	-13.0174	analysis | discussion
0.526316	-13.0175	models | variable
0.164251	-13.0175	use | learning
0.666667	-13.0175	document | indicator
0.138889	-13.0177	(in | of
0.100179	-13.0177	the | that they
0.526316	-13.0178	while | indicates
0.0790646	-13.0178	is | source
0.333333	-13.0178	interestingly, | in
0.666667	-13.0179	6.3 | results
0.101476	-13.0179	search | a
0.230769	-13.0179	is | core
0.216667	-13.018	the | operators
0.444444	-13.018	these | actual
0.526316	-13.0181	type | nodes
0.6	-13.0181	efforts | information
0.857143	-13.0182	compose | relations.
0.0758197	-13.0183	the | and | 2009).
0.777778	-13.0184	that | gaps.
0.252874	-13.0186	this | gives
0.181287	-13.0186	generated | with
0.857143	-13.0187	matrix | matrix,
0.146154	-13.0187	positive | negative | and
0.206107	-13.0188	future | will
0.107438	-13.0188	a | size
0.201439	-13.0189	creating | of
0.444444	-13.019	part-of-speech | syntactic
0.197183	-13.0192	the | spanning
0.0950081	-13.0193	first | with
0.173913	-13.0193	to | actual
0.179191	-13.0193	selected | from
0.375	-13.0193	itself, | and
0.340426	-13.0194	the | sizes.
0.225225	-13.0196	syntactic | dependency
0.283582	-13.0197	further, | in
0.204545	-13.0198	active learning | the
0.174863	-13.0199	gold standard | and
0.228571	-13.0199	keeping | the
0.478261	-13.02	easily | using
0.277778	-13.02	machine | human
0.170103	-13.0202	2.2 | a
0.333333	-13.0202	instead | only
0.777778	-13.0203	sampling, | each
0.296875	-13.0203	we propose | using
0.186335	-13.0204	a | designed
0.146718	-13.0204	for | technology
0.314815	-13.0205	a | plausible
0.237113	-13.0205	were | data.
0.2	-13.0205	an | without
0.105368	-13.0206	such | by
0.175824	-13.0207	below, | the
0.216667	-13.0208	of | event.
0.265823	-13.0208	‘the | the
0.666667	-13.021	active | research
0.666667	-13.021	paradise | dialogue
0.461538	-13.021	acld | to
0.175824	-13.0211	having | a
0.0843829	-13.0212	a | corpus.
0.40625	-13.0213	by | finally
0.172973	-13.0215	this | what
0.104651	-13.0215	a | 2010).
0.296875	-13.0215	(2011) | a
0.5	-13.0216	using | google
0.0976027	-13.0217	the | e.g.,
0.223214	-13.0218	hypothesis, | the
0.444444	-13.0219	candidate | highest
0.217391	-13.022	the | right.
0.555556	-13.022	building | parallel
0.239583	-13.0221	have | since
0.211382	-13.0222	as | next
0.142857	-13.0222	in | 2006).
0.857143	-13.0223	(in | sense
0.555556	-13.0226	wsd | word
0.186335	-13.0226	and | helpful
0.857143	-13.0228	american | arabic.
0.217391	-13.0228	fall | the
0.6	-13.023	because | want
0.347826	-13.0231	lda | model
0.388889	-13.0231	global | which
0.226415	-13.0232	the | query,
0.461538	-13.0232	meteor | bleu
0.857143	-13.0233	s., | h.,
0.242105	-13.0234	in | step.
0.378378	-13.0235	they | various
0.388889	-13.0235	overall | score
0.444444	-13.0235	which | importance
0.0660964	-13.0235	we | sentences
0.165049	-13.0236	known | a
0.357143	-13.0236	this | occurs
0.388889	-13.0237	6we | the
0.340426	-13.0238	output | human
0.777778	-13.0239	native | make
0.777778	-13.0239	role | role.
0.223214	-13.0239	of | dictionary.
0.181287	-13.024	another | with
0.184049	-13.0241	normalized | and
0.666667	-13.0242	generated | generation
0.125714	-13.0242	in | overall
0.857143	-13.0243	allows | share
0.305085	-13.0243	bullet | the
0.110151	-13.0243	measure | a
0.0796359	-13.0244	a | method
0.777778	-13.0244	baseline | meteor
0.333333	-13.0245	the | definitions.
0.27027	-13.0246	copy | the
0.357143	-13.0247	speaker | or
0.857143	-13.0248	ability | complex
0.107216	-13.0248	training | data
0.777778	-13.0248	among | ratio
0.145594	-13.0248	useful | in
0.256098	-13.0249	this | represents
0.184049	-13.0251	outputs | and
0.857143	-13.0253	euromatrixplus | under
0.130699	-13.0254	for | we used | the
0.100733	-13.0255	the | product
0.526316	-13.0256	mining | entities
0.4	-13.0256	abstract | 2011
0.197183	-13.0258	is | short
0.5	-13.0258	three | further
0.461538	-13.0259	standard | model,
0.287879	-13.0259	all | sentences.
0.277778	-13.026	learners | to
0.305085	-13.0261	less | less
0.444444	-13.0261	training | discriminative
0.204545	-13.0262	major | is
0.184049	-13.0263	distinct | and
0.265823	-13.0264	and | 3).
0.128358	-13.0264	any | with
0.526316	-13.0265	text | sentences,
0.555556	-13.0266	using | thesaurus
0.555556	-13.0266	varying | different
0.150407	-13.0266	define | for
0.101103	-13.0266	the | answer
0.857143	-13.0268	search-based | prediction
0.5625	-13.0268	stress | stress
0.239583	-13.0269	classifier | each
0.252874	-13.0269	total | were
0.6	-13.0269	follow | where
0.0910448	-13.027	the | simply
0.0860927	-13.027	if | not
0.5625	-13.0271	search | key
0.419355	-13.0271	user | text
0.305085	-13.0272	only | annotated
0.244444	-13.0272	through | which
0.40625	-13.0273	of | symbols.
0.154506	-13.0273	et | 2003).
0.6	-13.0273	which | part.
0.165829	-13.0274	of | equivalent
0.237113	-13.0274	translation | languages
0.113426	-13.0275	• | with
0.5	-13.0275	symbol, | is
0.149194	-13.0276	that | larger
0.0265576	-13.0276	- | - | -
0.478261	-13.0277	of | null2
0.154506	-13.0277	that | a given
0.193333	-13.0279	of | automated
0.320755	-13.028	for | targets
0.165829	-13.028	specific | that
0.173913	-13.0281	vector | =
0.333333	-13.0281	on | knowledge.
0.777778	-13.0281	proceeds | two
0.777778	-13.0281	analyzing | structure
0.172973	-13.0281	therefore | for
0.143911	-13.0282	probabilistic | and
0.0984456	-13.0283	the | article
0.209677	-13.0284	for | summaries
0.209677	-13.0284	we | minimum
0.164251	-13.0285	called | and
0.777778	-13.0285	n | (n
0.239583	-13.0287	validity | the
0.244444	-13.0287	track | and
0.252874	-13.0289	these | standard
0.333333	-13.0289	transport | transport
0.777778	-13.029	indeed, | only
0.259259	-13.029	is | function,
0.526316	-13.0293	has | history
0.0582524	-13.0293	the | annotated
0.857143	-13.0294	showing | current
0.142336	-13.0295	3.2 | in
0.116505	-13.0295	provides | and
0.444444	-13.0295	of | persons
0.189542	-13.0295	to | reduce
0.174863	-13.0296	and | utterances
0.171123	-13.0296	we | mention
0.156951	-13.0296	increase | in
0.166667	-13.0298	the | title
0.226415	-13.0298	inputs | the
0.167513	-13.0299	the | off
0.478261	-13.0299	a | granularity.
0.777778	-13.0299	em | bayesian
0.388889	-13.03	but | difference
0.168367	-13.0301	dimension | the
0.6	-13.0301	aligner | alignment
0.4	-13.0302	experiments | results.
0.388889	-13.0304	investigate | use of
0.777778	-13.0304	difficult | estimate
0.186335	-13.0305	yet | of
0.153846	-13.0307	he | to
0.0426155	-13.0307	first | of
0.5625	-13.0308	on | transcriptions
0.340426	-13.0309	uses | its
0.0826141	-13.031	use | from
0.0875	-13.031	and | 2007;
0.223214	-13.031	for | information.
0.326923	-13.031	in | terms,
0.172973	-13.0311	of | analysis.
0.25	-13.0312	and | algorithms.
0.6	-13.0312	structure | june
0.388889	-13.0314	3 | overall
0.221239	-13.0315	of | relational
0.419355	-13.0316	not | necessary
0.155556	-13.0317	the | relevance
0.666667	-13.0317	4.2 | corpora
0.7	-13.0317	syntactically | but
0.365854	-13.0317	documents | text
0.295082	-13.0318	is | to.
0.857143	-13.0319	scale | prediction
0.234694	-13.0319	and | relation.
0.555556	-13.0319	output: | 1:
0.175141	-13.032	6 | our
0.666667	-13.0321	given | sequence.
0.419355	-13.0321	present | automatically
0.167513	-13.0322	that | entities
0.305085	-13.0323	of | (one
0.857143	-13.0324	p(z|d)
0.136986	-13.0324	= | (1)
0.111359	-13.0325	use | or
0.0776805	-13.0326	used | with
0.172043	-13.0326	contains | from
0.175141	-13.0327	grammars | and
0.777778	-13.0327	semantic | 1990)
0.777778	-13.0327	npk | its
0.305085	-13.0328	if there | no
0.122616	-13.0329	algorithm | with
0.156951	-13.0329	for | corpus,
0.444444	-13.0329	we | for.
0.5	-13.0329	complete | system
0.857143	-13.033	sense | (yarowsky,
0.320755	-13.0331	represents | word
0.184049	-13.0332	of | six
0.142857	-13.0332	since | this
0.105477	-13.0335	random | and
0.365854	-13.0335	tasks | document
0.295082	-13.0336	of | 2010,
0.165	-13.0336	factor | to
0.265823	-13.0337	accuracy | but
0.0974359	-13.0338	need | to
0.192053	-13.0338	the | point.
0.419355	-13.0341	k-means | for
0.0551009	-13.0342	present | the
0.461538	-13.0344	improves | both
0.27027	-13.0344	by | path
0.5	-13.0344	when | occurs
0.0969388	-13.0345	all | are | of
0.172973	-13.0346	reduced | of
0.526316	-13.0347	β | set
0.555556	-13.0348	in | column,
0.136519	-13.0348	a | hypothesis
0.666667	-13.0349	studied | 2002;
0.5625	-13.0351	vbp | jj
0.165829	-13.0351	is | just
0.252874	-13.0352	supervised | training
0.221239	-13.0353	it | multiple
0.242105	-13.0354	model, | an
0.348837	-13.0354	easier | more
0.857143	-13.0355	learn | unannotated
0.388889	-13.0355	size | corpus.
0.164179	-13.0355	exception | the
0.347826	-13.0357	that | considerably
0.217391	-13.0358	largely | of
0.234694	-13.0358	is | (a
0.555556	-13.0358	most | properties
0.461538	-13.0358	it’s | is
0.27027	-13.0358	abstract | have
0.478261	-13.0359	“a | is
0.139785	-13.036	to | next
0.234694	-13.0362	besides | and
0.428571	-13.0362	are | classes,
0.6	-13.0362	like | just
0.109649	-13.0364	determine | of
0.247191	-13.0364	rather, | the
0.242105	-13.0365	question | information
0.0969388	-13.0366	used | which
0.292308	-13.0366	a | si
0.347826	-13.0367	of | dominant
0.0786517	-13.0367	using | from
0.211382	-13.0368	bilingual | are
0.190789	-13.0369	may | other
0.348837	-13.0369	hash | is
0.178161	-13.0371	restricted | of
0.232323	-13.0371	which | user
0.419355	-13.0371	list | given
0.0857909	-13.0371	in | tree
0.348837	-13.0371	window | words
0.215517	-13.0372	different | models
0.320755	-13.0373	significantly | both
0.6	-13.0373	utility | this
0.0685859	-13.0375	algorithm | to
0.13355	-13.0375	to | inference
0.0603675	-13.0375	the | state
0.0703473	-13.0375	annotation | and
0.857143	-13.0376	extending | existing
0.378378	-13.0377	constraints | used
0.5625	-13.0378	as | problematic
0.295082	-13.0378	this | different from
0.239583	-13.0379	challenges | to
0.198582	-13.0379	a | proportion
0.305085	-13.0379	hovy, | and
0.126471	-13.038	single | for
0.0103235	-13.0381	the | is | the
0.478261	-13.0382	two | identical
0.0687876	-13.0383	* | *
0.5625	-13.0385	problem | joint
0.2625	-13.0385	model | unsupervised
0.857143	-13.0386	goals | used
0.133987	-13.0386	classification | is
0.314815	-13.0387	their | (e.g.,
0.133987	-13.0387	detection | a
0.6	-13.0387	are | discarded.
0.139785	-13.0388	most | or
0.5625	-13.0389	also | characteristics
0.5	-13.0389	annotations | automatically
0.0432871	-13.039	based on | the
0.526316	-13.0391	for | unconstrained
0.217391	-13.0391	we | hypotheses
0.6	-13.0391	systems | typically
0.244444	-13.0392	(this | the
0.114558	-13.0392	the | factors
0.0621075	-13.0392	one | for
0.7	-13.0393	parameters | control
0.7	-13.0393	evaluation | wsi
0.158371	-13.0393	for | 1,
0.209677	-13.0394	sufficient | of
0.461538	-13.0394	three | final
0.461538	-13.0394	1991), | and
0.40625	-13.0395	even | single
0.0911854	-13.0396	two | which
0.2625	-13.0396	of | sciences
0.295082	-13.0397	however, | important
0.122283	-13.0397	problem, | the
0.259259	-13.0398	rates | and
0.444444	-13.0399	1000 | test
0.14741	-13.0399	evaluation | which
0.461538	-13.04	as | complement
0.156951	-13.04	results | models
0.7	-13.0401	institute | stuttgart
0.0563218	-13.0402	we | a | which
0.14741	-13.0402	cell | the
0.213115	-13.0403	model: | the
0.388889	-13.0404	2003) | used
0.444444	-13.0404	β | with
0.7	-13.0405	under | grammar
0.185897	-13.0406	generic | and
0.857143	-13.0407	specificity | summary
0.204545	-13.0407	close | to
0.189542	-13.0407	the | pass
0.5	-13.0407	as | irrelevant
0.666667	-13.0409	with | lrscore
0.6	-13.0409	before, | are
0.136519	-13.0411	(2009) | to
0.247191	-13.0411	an | under
0.777778	-13.0411	synthesis | which
0.5625	-13.0412	speaking, | is
0.444444	-13.0412	research | about
0.266667	-13.0413	and | scientific
0.143396	-13.0413	would | this
0.239583	-13.0414	thus | not
0.253012	-13.0414	to | (with
0.119792	-13.0415	results | system
0.141818	-13.0415	discuss | and
0.206349	-13.0416	6.2 | of
0.461538	-13.0416	over | data,
0.777778	-13.0416	“give | largest
0.6	-13.0416	(finkelstein | al.,
0.230769	-13.0417	negation | and
0.206349	-13.0418	we | efficient
0.180723	-13.0418	we | data from
0.228571	-13.0418	nature of | and
0.419355	-13.0419	are | separately
0.279412	-13.042	a | phrasal
0.333333	-13.042	all | appear
0.7	-13.0422	task | predefined
0.7	-13.0422	parser, | parser,
0.615385	-13.0423	phoneme | phoneme
0.0920245	-13.0424	make | to
0.777778	-13.0425	features | koo
0.348837	-13.0425	that | specifically
0.5625	-13.0426	although | relatively
0.615385	-13.0426	through | information.
0.348837	-13.0429	experimental | proposed
0.166667	-13.043	lines | the
0.241758	-13.043	to | classify
0.3	-13.0431	protein | event
0.857143	-13.0432	p(e2|e1)
0.526316	-13.0432	2000 | 4000
0.178161	-13.0432	r1 | r2
0.224299	-13.0432	we | distinguish
0.133987	-13.0433	as | 1.
0.326531	-13.0433	indeed | for
0.27027	-13.0433	the | thing
0.134426	-13.0433	of | vs.
0.188312	-13.0434	updated | the
0.241758	-13.0436	a | graph,
0.666667	-13.0437	evaluation | classification.
0.0843061	-13.0438	it | more
0.444444	-13.0439	also | rank
0.5	-13.044	with | probabilities.
0.143939	-13.0441	the | readability
0.320755	-13.0441	defined | over
0.40625	-13.0441	2.2 | model
0.428571	-13.0442	need | find
0.174157	-13.0443	in | vs.
0.256098	-13.0444	this | partial
0.777778	-13.0444	collocations | reordering
0.193103	-13.0444	this | provides
0.259259	-13.0445	offers | of
0.378378	-13.045	is | risk
0.256098	-13.045	the | flow
0.27027	-13.0451	in | fine-grained
0.378378	-13.0452	often | some
0.228571	-13.0452	2. | an
0.309091	-13.0452	we | (using
0.259259	-13.0453	with | structures
0.478261	-13.0454	although | improvement
0.27027	-13.0454	a | antecedent
0.097561	-13.0455	to | documents
0.0790378	-13.0455	a | about
0.5625	-13.0456	of | gender.
0.6	-13.0456	resulting | very
0.0866575	-13.0458	the | just
0.777778	-13.0458	update | updated
0.777778	-13.0458	factors | human
0.6	-13.0459	but | scale
0.188312	-13.046	frame | a
0.5625	-13.046	are | below:
0.5625	-13.046	with | effort.
0.114558	-13.046	several | for
0.104208	-13.0461	however, | on
0.0528724	-13.0461	found | the
0.179641	-13.0461	because | more
0.5	-13.0461	size | even
0.23	-13.0462	the | (2007).
0.143939	-13.0464	the | confusion
0.428571	-13.0465	itg | a
0.40625	-13.0466	pairs | relation
0.333333	-13.0466	in | health
0.615385	-13.0467	step | direction
0.419355	-13.0467	subject, | a
0.320755	-13.0468	shorter | and
0.163366	-13.0469	dependency | which
0.6	-13.047	multiple | provide
0.777778	-13.0472	university | groningen
0.253012	-13.0474	by | european
0.393939	-13.0476	a | weights,
0.444444	-13.0476	instantiations | and
0.388889	-13.0477	or | just
0.5625	-13.0477	errors | annotation
0.179641	-13.0477	can | even
0.263158	-13.0479	a | scenario
0.5	-13.0479	from | 5,
0.5625	-13.048	list | manually
0.5625	-13.048	allow | multiple
0.419355	-13.048	sample | sentences
0.275362	-13.0481	use | representation
0.167539	-13.0481	section 3 | and
0.283582	-13.0483	thre | the
0.165829	-13.0483	from | documents
0.153509	-13.0483	(2008) | and
0.857143	-13.0484	usc | institute
0.419355	-13.0485	between | arguments.
0.5	-13.0485	extracting | opinion
0.393939	-13.0486	intuition | if
0.378378	-13.0486	selection | best
0.0499319	-13.0486	approach | and
0.340426	-13.0487	for example, | query
0.5625	-13.0487	are | similar,
0.5625	-13.0487	involves | sentence
0.136519	-13.0487	very | that
0.201493	-13.0487	in | design
0.326531	-13.0487	list, | a
0.114558	-13.0488	of | coreference
0.222222	-13.0489	table | shows | each
0.388889	-13.0491	wizard | in
0.203008	-13.0491	the | word’s
0.259259	-13.0491	where | possible
0.5	-13.0491	algorithm | hierarchical
0.5	-13.0491	required to | this
0.458333	-13.0492	similar | related
0.290323	-13.0492	conversational | in
0.0813107	-13.0492	to | better
0.6	-13.0492	problems | text
0.108696	-13.0493	automatically | a
0.615385	-13.0493	previous | approaches.
0.419355	-13.0493	alternatives | that
0.857143	-13.0494	14: | return
0.111364	-13.0494	next, | the
0.253012	-13.0494	therefore | from
0.5	-13.0494	that | lost
0.777778	-13.0495	pairs | close
0.0227273	-13.0495	are | in
0.177143	-13.0496	and | parts
0.6	-13.0499	laboratory | information
0.378378	-13.0502	is | handling
0.3	-13.0502	previously | by
0.388889	-13.0503	marginals | the
0.23913	-13.0503	on | whole
0.428571	-13.0504	others | or
0.365854	-13.0504	for | reason
0.108932	-13.0504	statistical | to
0.190476	-13.0504	approximate | and
0.121622	-13.0504	the | equivalent
0.857143	-13.0505	iteratively | labeled
0.263158	-13.0505	given | context
0.326531	-13.0505	the | dialects
0.357143	-13.0506	simple | over
0.365854	-13.0506	include: | and
0.7	-13.0507	support | annotation
0.182927	-13.0508	they | one
0.168421	-13.0508	the | journal
0.263158	-13.0508	we | alignments.
0.215517	-13.0508	of | entity.
0.378378	-13.0509	difference, | the
0.176136	-13.0509	filtering | of
0.326531	-13.0509	methods | proposed
0.857143	-13.051	relevant | research.
0.180723	-13.051	3 | system
0.777778	-13.051	method | identification
0.6	-13.051	relevant | (e.g.,
0.458333	-13.0511	classes | syntactic
0.204724	-13.0512	comparison | our
0.615385	-13.0512	devtest | devtest
0.666667	-13.0513	between | two.
0.25	-13.0513	that | changes
0.615385	-13.0516	effect | increasing
0.170213	-13.0517	from | context
0.155556	-13.0518	of | characters
0.263158	-13.0518	(8) | is
0.777778	-13.0519	expansion | text
0.156682	-13.0519	effort | the
0.444444	-13.052	review | sentiment
0.151899	-13.052	roughly | the
0.172222	-13.0521	this approach | a
0.666667	-13.0522	greedy | this
0.201493	-13.0522	the | agreement.
0.226415	-13.0523	particular | or
0.168421	-13.0523	largely | the
0.184713	-13.0523	coefficients | the
0.526316	-13.0524	various | linguistic
0.131833	-13.0524	development | for
0.232323	-13.0524	translation | mt
0.378378	-13.0525	polarity | opinion
0.0608519	-13.0525	a | model | a
0.253012	-13.0525	of | pp
0.857143	-13.0526	|x)
0.666667	-13.0526	make | clear
0.130573	-13.0526	a | cluster
0.348837	-13.0526	in | (2008).
0.458333	-13.0528	probability | estimate
0.0557452	-13.0528	to | these
0.234694	-13.0528	corpus | news
0.184713	-13.0528	with | 10
0.0846561	-13.0529	that | any
0.333333	-13.0529	of | electronic
0.279412	-13.053	training, | test
0.357143	-13.053	translated | french
0.857143	-13.0531	nonetheless, | only
0.857143	-13.0531	need | traditional
0.180723	-13.0532	and | de
0.0639579	-13.0532	the | .
0.777778	-13.0533	algorithms | improve
0.283582	-13.0535	table 4 | results
0.129231	-13.0536	the | graphs
0.368421	-13.0537	were | against
0.368421	-13.0537	average | three
0.666667	-13.0538	becomes: | =
0.206349	-13.0538	objective | that
0.215517	-13.0538	metric, | the
0.419355	-13.0538	our | namely
0.184713	-13.0539	and | german
0.666667	-13.0542	where | incoming
0.1261	-13.0542	also | using
0.0447214	-13.0542	2 | of
0.615385	-13.0542	performance | increased
0.172222	-13.0543	analysis | from
0.777778	-13.0543	synchronous | s
0.6	-13.0543	+ | s.t.
0.5	-13.0543	a | prior,
0.320755	-13.0544	tradeoff | and
0.159624	-13.0544	a | appropriate
0.388889	-13.0546	(koo | al.,
0.857143	-13.0547	acm | knowledge
0.5	-13.0549	model | preceding
0.526316	-13.055	these | categories.
0.103113	-13.055	by | into
0.193103	-13.0551	phrases | from
0.857143	-13.0552	word–colour | by
0.5625	-13.0552	all | segmentations
0.777778	-13.0552	scheme | schemes
0.378378	-13.0553	that | overlapping
0.27027	-13.0553	such | should
0.181818	-13.0554	generalization | the
0.237113	-13.0555	documents | set
0.259259	-13.0555	sequences | as
0.145669	-13.0556	help | and
0.237113	-13.0556	and | approximately
0.326531	-13.0556	are | equivalent
0.27027	-13.0556	an | considered
0.7	-13.0558	regression | regularization
0.161765	-13.0558	we | wordnet
0.142322	-13.0561	labeling | and
0.857143	-13.0562	linguistic | summaries
0.777778	-13.0562	using | operator
0.6	-13.0562	events | corpus.
0.162562	-13.0564	linguistic | is
0.263158	-13.0564	how | given
0.3	-13.0566	quality | using
0.263158	-13.0566	benefits | of
0.393939	-13.0567	peter | and
0.154867	-13.0567	w | =
0.154867	-13.0567	results | were
0.104628	-13.0567	in | system.
0.857143	-13.0568	since | days
0.857143	-13.0568	(female) | .
0.320755	-13.0569	the | mert.
0.6	-13.0569	english, | languages
0.348837	-13.0569	context-free | is
0.444444	-13.0571	work | generating
0.0998152	-13.0571	even | a
0.106472	-13.0572	our experiments | the
0.615385	-13.0572	possible | choose
0.615385	-13.0572	predicting | constituents
0.419355	-13.0572	the | built.
0.303571	-13.0572	several | models
0.198529	-13.0575	study | by
0.147541	-13.0575	the | appearing
0.25	-13.0576	10 | 70
0.40625	-13.0577	the | us.
0.215517	-13.0577	this | funded by
0.340909	-13.0578	recognition, | and
0.340426	-13.058	future work | our
0.393939	-13.0581	in | 4th
0.314815	-13.0581	can | ranking
0.213675	-13.0581	three | different
0.526316	-13.0582	be | unless
0.340426	-13.0582	performances | and
0.226415	-13.0582	of | subsequent
0.148148	-13.0582	the | gap
0.615385	-13.0583	related | topic.
0.615385	-13.0583	idea | combine
0.139286	-13.0583	the | occurring
0.458333	-13.0584	size | beam
0.224299	-13.0585	therefore, | are
0.168421	-13.0585	joint | for
0.244444	-13.0586	take | from
0.666667	-13.0587	as | msummationdisplay
0.159624	-13.0587	the | article.
0.193103	-13.0588	are | difficult
0.25	-13.0589	precise | of
0.368421	-13.0589	can | accurate
0.0609589	-13.059	the | available
0.0832266	-13.059	specific | of
0.777778	-13.059	29 | 29
0.777778	-13.059	‘the | criterion’
0.777778	-13.059	hierarchy | agents
0.259259	-13.059	factored | of
0.6	-13.0591	percent | were
0.104628	-13.0591	a | length
0.348837	-13.0592	to | responses.
0.12	-13.0593	furthermore, | of
0.5625	-13.0593	useful | during
0.253012	-13.0593	correlation | between
0.190476	-13.0594	captured | the
0.666667	-13.0595	tag | tags.
0.193103	-13.0595	for | subset
0.7	-13.0596	forms | basis
0.162562	-13.0596	for | v
0.295082	-13.0596	use | during
0.357143	-13.0597	analysis | found
0.0848806	-13.0598	the | wikipedia
0.171271	-13.0598	the | generalized
0.283582	-13.0599	rules | parse
0.275362	-13.06	term, | the
0.333333	-13.06	task | any
0.393939	-13.0601	of | subject,
0.428571	-13.0601	set | symbols
0.2	-13.0601	we | to compute
0.357143	-13.0602	the | layer.
0.27027	-13.0602	a | threshold.
0.256098	-13.0603	and | validation
0.190476	-13.0603	the | meaningful
0.458333	-13.0604	often | even
0.173184	-13.0606	which | syntactic
0.151515	-13.0607	to | ones
0.156682	-13.0607	2000; | et
0.478261	-13.0608	structures | trees
0.326531	-13.0608	2004; | 2008;
0.148148	-13.0608	then, | of
0.857143	-13.0609	indicates | chosen
0.393939	-13.061	there | distinct
0.303571	-13.0611	seen | training
0.283582	-13.0613	including | features
0.187919	-13.0613	the | are:
0.326531	-13.0614	many | pairs
0.357143	-13.0615	in | literal
0.428571	-13.0617	workshop, | 24
0.333333	-13.0617	linear | vector
0.0808203	-13.0618	inference | the
0.283582	-13.062	acknowledgements | research | the
0.208	-13.0621	n-grams | a
0.117647	-13.0622	standard | is
0.20155	-13.0622	as | full
0.146939	-13.0622	considering | and
0.666667	-13.0623	current | preceding
0.666667	-13.0623	posterior | model.
0.0963855	-13.0624	mentions | the
0.6	-13.0624	extracted | them.
0.203125	-13.0624	remaining | in
0.3	-13.0625	we found | our
0.666667	-13.0627	denote | variables
0.32	-13.0628	information | web
0.40625	-13.0629	by | program.
0.162562	-13.0629	of | active
0.5	-13.0629	various | tasks
0.314815	-13.063	grounding | to
0.0980392	-13.0631	these | which
0.174157	-13.0631	instances | by
0.444444	-13.0631	from | titles
0.393939	-13.0632	learned | corpus
0.357143	-13.0632	and | responses.
0.27027	-13.0632	our | improved
0.295082	-13.0634	section | concludes the
0.340909	-13.0634	that | complementary
0.283582	-13.0635	and | agree
0.18239	-13.0635	also | these
0.155251	-13.0635	the | composition
0.227723	-13.0635	by | particular
0.109865	-13.0636	represented | a
0.314815	-13.0637	5 | algorithm
0.0695574	-13.0637	in | 2
0.295082	-13.0638	similarly, | be
0.137324	-13.0638	we | comparable
0.206349	-13.0639	semantic | relations
0.615385	-13.064	within | considered
0.227723	-13.064	in | representation.
0.247059	-13.064	best | over
0.857143	-13.0641	agents | their
0.5	-13.0641	correct | output
0.0913313	-13.0642	of | discourse
0.40625	-13.0642	graph | relations
0.428571	-13.0643	linguistic | text
0.428571	-13.0643	corpus | study
0.198529	-13.0643	events | are
0.185897	-13.0645	by | parser
0.208	-13.0646	version | with
0.0962199	-13.0646	form | a
0.25974	-13.0647	this | too
0.478261	-13.0647	a | strings.
0.00851994	-13.0647	the | and | the
0.615385	-13.0648	productdisplay | (1)
0.40625	-13.0649	thus, | semantic
0.32	-13.0649	perspective | in
0.5	-13.065	is | subsequence
0.25	-13.0651	this | evidence
0.174157	-13.0651	that | include
0.0752916	-13.0652	the | the | language.
0.211864	-13.0652	implementation | for
0.326531	-13.0652	collection | each
0.348837	-13.0652	exchange | and
0.220183	-13.0653	on | translation.
0.185897	-13.0654	the | concepts.
0.220183	-13.0655	initially | the
0.263158	-13.0656	consider | sentence
0.615385	-13.0659	acoustic | word
0.0689046	-13.066	the | times
0.0778781	-13.0661	2. | of
0.32	-13.0661	0 | 300
0.340909	-13.0661	learned | using
0.244444	-13.0662	count | number
0.40625	-13.0662	have | direct
0.0746888	-13.0662	scores | and
0.526316	-13.0663	(2006) | approach
0.234694	-13.0663	bigram | for
0.666667	-13.0664	removed | sentences
0.125	-13.0664	optimization | the
0.141264	-13.0665	when | word
0.526316	-13.0666	’s | .
0.5625	-13.0666	vector | elements
0.146341	-13.0666	optimal | of
0.428571	-13.0667	magnitude | than
0.18239	-13.0667	vector | each
0.348837	-13.0667	19 | for
0.161765	-13.0668	to compare | of
0.666667	-13.0669	were | corrected
0.241758	-13.0669	in | respective
0.615385	-13.067	taking | product
0.18239	-13.067	the | emotional
0.348837	-13.067	using | pairs.
0.393939	-13.0671	extract | phrases
0.6	-13.0672	scanning | scanning
0.6	-13.0672	null(null) | null(null)
0.368421	-13.0672	score | n-gram
0.244444	-13.0673	in | writing
0.256098	-13.0673	as | sum
0.2	-13.0673	languages, | a
0.138686	-13.0674	for | processing
0.139194	-13.0675	so | are
0.40625	-13.0677	in | u.s.
0.857143	-13.0678	track | conll
0.7	-13.0678	changes | time
0.22549	-13.0679	by | under grant
0.666667	-13.0681	frequency | inverse
0.368421	-13.0681	not | answer
0.7	-13.0682	in | respects.
0.177515	-13.0682	strong | to
0.143411	-13.0682	the | drawn
0.164103	-13.0684	of | item
0.253012	-13.0686	to | users.
0.357143	-13.0688	every | this
0.857143	-13.0689	german | (dfki)
0.140221	-13.0689	1 | i
0.0309114	-13.069	the | corpus
0.160976	-13.069	for | topics
0.615385	-13.0693	therefore, | need
0.32	-13.0693	it | distinguish
0.478261	-13.0694	verify | our
0.7	-13.0695	thus, | estimate
0.158654	-13.0696	the | segmented
0.203125	-13.0696	the | ranks
0.458333	-13.0697	much | difficult
0.12766	-13.0697	way | is
0.615385	-13.0697	component | independent
0.103175	-13.0698	used | or
0.220183	-13.0699	method | sentences
0.192857	-13.0701	comparison | with
0.40625	-13.0702	subsequently | to
0.0894188	-13.0702	the | derived
0.357143	-13.0703	example | argument
0.175439	-13.0705	regularization | the
0.227723	-13.0707	3.5 | the
0.444444	-13.0708	as | targeted
0.216216	-13.0709	time | or
0.198529	-13.0709	itself | of
0.326531	-13.0709	event | argument
0.857143	-13.071	measures | assess
0.378378	-13.0711	sequentially | the
0.244444	-13.0711	is | update
0.138182	-13.0711	understanding | of
0.110599	-13.0711	by | over
0.210084	-13.0711	to | variation
0.171271	-13.0712	as | often
0.112981	-13.0712	a | must
0.247059	-13.0713	the | states.
0.275362	-13.0714	10 | 100
0.176471	-13.0714	identifies | the
0.857143	-13.0715	resource | multilingual
0.0386648	-13.0715	of | at
0.171271	-13.0717	an | probability
0.232323	-13.0717	for | 2011).
0.145749	-13.0718	bleu | with
0.309091	-13.0718	feature, | in
0.40625	-13.0719	cluster | clustering
0.141264	-13.0719	head | is
0.184713	-13.0719	exploit | to
0.458333	-13.072	hypothesis | h
0.428571	-13.072	table 1 | some
0.857143	-13.0721	introduce | external
0.7	-13.0721	can | :=
0.120548	-13.0721	an | semantic
0.526316	-13.0722	random | sequence
0.104938	-13.0722	similar | for
0.0795724	-13.0722	semantic | for
0.142857	-13.0723	less than | the
0.154545	-13.0725	(in | and
0.857143	-13.0726	matches | them
0.169399	-13.0726	4, | and
0.615385	-13.0727	baseline | head
0.19708	-13.0727	is | loss
0.118421	-13.0729	left | of
0.105485	-13.073	instead | a
0.333333	-13.073	they used | to
0.857143	-13.0731	define | arguments.
0.857143	-13.0731	syntactic | 2002)
0.526316	-13.0731	standard | metrics
0.0884956	-13.0731	users | the
0.208333	-13.0731	our | instead
0.22549	-13.0732	to | bias
0.478261	-13.0732	ces | the
0.130159	-13.0732	control | the
0.368421	-13.0732	unfortunately, | are
0.165803	-13.0734	one | same
0.666667	-13.0735	performed | selection
0.857143	-13.0736	acquired | corpora
0.23913	-13.0736	of | example.
0.340909	-13.0737	food | the
0.333333	-13.0738	iteratively | a
0.15942	-13.0739	6.2 | the
0.0648734	-13.0739	that | same
0.0754516	-13.0741	• | in
0.25641	-13.0741	accuracy, | to
0.13289	-13.0741	ways | the
0.348837	-13.0741	2011; | al.,
0.244186	-13.0742	a | cause
0.666667	-13.0743	our | segmentation.
0.393939	-13.0744	often | about
0.106157	-13.0745	the | world
0.203125	-13.0747	primarily | the
0.7	-13.0748	set | pubmed
0.25	-13.075	a | structures.
0.368421	-13.075	words | seed
0.0978261	-13.0753	we | classification
0.526316	-13.0754	over | just
0.358974	-13.0754	these | entire
0.615385	-13.0754	by | generated.
0.5625	-13.0756	find | top
0.263158	-13.0757	presented | approach
0.0958188	-13.0758	then | as
0.192857	-13.0758	in | fully
0.348837	-13.0758	the | count.
0.358974	-13.0759	that | belief
0.413793	-13.076	2) | using
0.7	-13.0761	uses | pivot
0.7	-13.0761	frame | fes
0.192857	-13.0762	we | combine
0.857143	-13.0763	composition | meaning
0.244186	-13.0763	draw | and
0.0896657	-13.0764	single | and
0.326531	-13.0764	the | conjunct
0.348837	-13.0765	in | highlight
0.0980036	-13.0766	the | word,
0.160976	-13.0766	to | true
0.152174	-13.0766	only | were
0.40625	-13.0767	a | british
0.177515	-13.0767	to | response
0.247059	-13.0767	that | interaction
0.167568	-13.0768	is | optimal
0.267606	-13.0768	the | reranker
0.5625	-13.077	requirements | that
0.13245	-13.077	2.2 | of
0.0794781	-13.0771	ranking | the
0.25641	-13.0771	more than | than
0.150215	-13.0771	here, | and
0.368421	-13.0771	with | method,
0.309091	-13.0772	code | is
0.428571	-13.0773	evaluated | both
0.526316	-13.0774	learn | patterns
0.285714	-13.0774	additionally, | that
0.144578	-13.0774	follow | and
0.0739979	-13.0775	set of | is
0.113527	-13.0775	answer | the
0.326531	-13.0776	synthetic | and
0.0564904	-13.0777	a | was
0.152174	-13.0777	we | containing
0.7	-13.0778	parsers | collins,
0.413793	-13.0778	also | annotators
0.204724	-13.0778	and | domains.
0.303571	-13.0778	discourse | relation
0.340909	-13.0778	of | cohn
0.142857	-13.0779	implemented | and
0.666667	-13.078	compared | approaches.
0.253012	-13.078	read | and
0.210084	-13.078	columns | and
0.234043	-13.0781	above, | is
0.271429	-13.0781	complex | or
0.358974	-13.0783	using | accuracy.
0.204724	-13.0784	inverse | the
0.333333	-13.0784	the | (2005).
0.857143	-13.0785	given | string,
0.857143	-13.0785	heads | syntactic
0.378378	-13.0786	approach | meaning
0.266667	-13.0787	error | errors
0.192857	-13.0787	performing | in
0.103239	-13.0787	the | smt
0.348837	-13.0787	are | sufficient
0.0868946	-13.0788	speech | and
0.0900763	-13.0788	the | predicate
0.127726	-13.0789	is | sentiment
0.160976	-13.0789	for | final
0.615385	-13.0789	with | pos,
0.0856354	-13.079	local | the
0.163265	-13.079	a | logical
0.104822	-13.079	of | i.e.
0.295082	-13.0791	we assume | each
0.22549	-13.0792	and | noise
0.179012	-13.0792	detecting | and
0.263158	-13.0792	it | under
0.285714	-13.0793	critical | of
0.17033	-13.0794	thus | as
0.857143	-13.0795	then | re-scoring
0.298246	-13.0796	we | list.
0.666667	-13.0797	bi | bi
0.666667	-13.0797	hello | querrien
0.666667	-13.0797	ρ | ρ
0.131579	-13.0797	model | both
0.149573	-13.0798	is | text.
0.241758	-13.0798	the | 1996).
0.267606	-13.0798	that | widely
0.171271	-13.08	whenever | the
0.10793	-13.0801	result | a
0.17341	-13.0801	the | bottom
0.298246	-13.0801	sibling | of
0.458333	-13.0802	unigram, | and
0.150215	-13.0804	a | comparison
0.348837	-13.0804	data | tasks
0.234043	-13.0805	include | from
0.857143	-13.0806	capture | characteristics
0.183544	-13.0808	the | paired
0.358974	-13.0808	impact | different
0.7	-13.0809	scales | number
0.857143	-13.0811	family | problems
0.857143	-13.0811	labeling | domain
0.285714	-13.0811	a | subtree
0.0785463	-13.0813	prior | the
0.7	-13.0813	modification | algorithm
0.177914	-13.0814	the | classes,
0.333333	-13.0815	a | m,
0.136364	-13.0816	base | and
0.13245	-13.0816	the | predicting
0.218182	-13.0817	these | context
0.393939	-13.082	that | clues
0.857143	-13.0822	different | ii)
0.105708	-13.0822	the | f-score
0.0603981	-13.0822	precision | the
0.478261	-13.0822	these | potentially
0.309091	-13.0822	analysis | opinion
0.0814536	-13.0824	also | on
0.214286	-13.0824	approach | text
0.0942761	-13.0826	recent | and
0.5625	-13.0827	words; | that
0.5625	-13.0827	the | presentation.
0.18239	-13.0827	this | step
0.25641	-13.0829	lee, | and
0.526316	-13.083	because | instead
0.0323471	-13.0831	a | an
0.147541	-13.0831	type | that
0.348837	-13.0832	m, | of
0.164103	-13.0833	of | indicating
0.478261	-13.0834	dots | are
0.5	-13.0834	types | general
0.313725	-13.0835	creative | a
0.478261	-13.0837	an | unseen
0.20155	-13.0837	information | different
0.106754	-13.0837	we | candidate
0.326531	-13.0837	one | expect
0.340909	-13.0837	by | considering
0.7	-13.0839	see | benefit
0.526316	-13.084	is | misspelled
0.368421	-13.084	(papineni | 2002),
0.0951557	-13.084	sentences, | the
0.5625	-13.0841	if | small,
0.0646178	-13.0842	recall | the
0.368421	-13.0842	for | them,
0.106754	-13.0842	of | social
0.393939	-13.0843	define | :
0.271429	-13.0844	usually | by
0.0764967	-13.0844	the | aligned
0.309091	-13.0844	to | possible,
0.526316	-13.0846	something | not
0.5	-13.0846	(2011), | a
0.174419	-13.0847	the | determines
0.615385	-13.0847	using | domain-specific
0.0888554	-13.0847	the | mean
0.458333	-13.0848	model | tuned
0.223301	-13.0848	all | input
0.125	-13.0848	indicate | of
0.271429	-13.0849	of | moves
0.0686275	-13.0849	the | every
0.18239	-13.0849	have been | on
0.333333	-13.0849	has | correlation
0.20155	-13.0849	features | learning
0.0557185	-13.085	computational | of
0.5625	-13.0851	but | complexity
0.0941176	-13.0851	does not | to
0.368421	-13.0851	that | naturally
0.146341	-13.0853	and | hierarchical
0.313725	-13.0853	monolingual | language
0.211864	-13.0855	a | denoted
0.19697	-13.0856	method | translation
0.191489	-13.0857	dataset | with
0.5625	-13.0858	are | different,
0.25641	-13.0858	thus, | an
0.121813	-13.0859	we | through
0.309091	-13.0859	a | middle
0.275362	-13.086	model | supervised
0.167568	-13.0861	we | manual
0.244186	-13.0861	are | evaluated
0.358974	-13.0862	tree | trees.
0.358974	-13.0862	by | expert
0.134483	-13.0862	the | constant
0.204918	-13.0862	3.4 | of
0.236559	-13.0865	our | four
0.378378	-13.0865	square6 | boxshadowdwn
0.0541295	-13.0865	and | use
0.183007	-13.0865	loss | in
0.458333	-13.0866	our method | only
0.0307997	-13.0867	for | with
0.183007	-13.0868	their | but
0.148936	-13.0868	and | means
0.478261	-13.087	then | back
0.478261	-13.087	(landauer | al.,
0.285714	-13.0872	unsupervised | using
0.085554	-13.0872	the | of | languages
0.32	-13.0872	operator | to
0.210084	-13.0872	of | specified
0.0884558	-13.0873	a | candidate
0.20155	-13.0873	f-measure | and
0.0686887	-13.0874	of | corpus.
0.326531	-13.0874	and | interactive
0.214286	-13.0875	however, | no
0.358974	-13.0876	we | configurations
0.5	-13.0878	form | pairs
0.5	-13.0878	adds | new
0.285714	-13.0879	provided | which
0.529412	-13.088	(one | one
0.529412	-13.088	are | network.
0.115979	-13.0881	relevant | and
0.117188	-13.0882	observed | and
0.101562	-13.0883	this | was | and
0.132653	-13.0885	of | corpora.
0.526316	-13.0886	from | 2005
0.0549199	-13.0886	the | previous
0.857143	-13.0887	motivated | part
0.180124	-13.0887	they are | for
0.382353	-13.0887	all | occurring
0.187919	-13.0887	the | double
0.112745	-13.0888	local | of
0.526316	-13.0889	due | constraints
0.340909	-13.0889	answer | it
0.100193	-13.089	is | topic
0.529412	-13.089	works | well
0.236559	-13.0891	came | the
0.478261	-13.0891	constructed | corpus
0.0714985	-13.0893	content | the
0.615385	-13.0893	extract | relevant
0.247059	-13.0896	type | its
0.104384	-13.0898	approaches | in
0.263158	-13.0899	acronym | the
0.142857	-13.09	parent | the
0.7	-13.0901	classifiers | predict
0.326531	-13.0901	found | best
0.298246	-13.0901	height | the
0.0933333	-13.0902	joint | of
0.368421	-13.0902	crucially | the
0.5	-13.0903	our | submission
0.180124	-13.0905	with | content
0.393939	-13.0907	actually | than
0.263158	-13.0907	obtained | an
0.857143	-13.0908	automated | identification
0.234043	-13.0908	these | specific
0.117647	-13.0908	what | a
0.382353	-13.0908	of | subjects.
0.458333	-13.0909	as | performance,
0.478261	-13.0909	both | match
0.28125	-13.0909	opinions | in
0.25	-13.0911	and | datasets.
0.428571	-13.0911	however, | increase
0.5625	-13.0912	up | certain
0.175439	-13.0912	the | perplexity
0.615385	-13.0912	experiments | evaluating
0.136201	-13.0912	et al., | 2008).
0.195489	-13.0913	incremental | and
0.075575	-13.0914	i | a
0.313725	-13.0914	therefore, | our
0.393939	-13.0915	read | with
0.458333	-13.0915	moreover, | method
0.526316	-13.0916	sentences | grammatical
0.298246	-13.0916	has | good
0.393939	-13.0917	for | calculation
0.529412	-13.0917	a | (np)
0.529412	-13.0917	whether | what
0.221154	-13.0917	reference | from
0.298246	-13.0917	dialogue | utterances
0.478261	-13.0918	classified | their
0.303571	-13.0918	is | theory
0.241379	-13.0919	mention, | the
0.263158	-13.0919	sentence | aligned
0.358974	-13.0919	are | captured
0.170455	-13.092	system | both
0.529412	-13.0921	lillian | and
0.478261	-13.0921	then | combination
0.184211	-13.0922	to | direct
0.333333	-13.0922	the | (sec.
0.5	-13.0922	some | had
0.7	-13.0923	rule | applied.
0.22549	-13.0924	as | last
0.529412	-13.0924	training | tl
0.458333	-13.0926	among | top
0.229167	-13.0926	used | classifier
0.25	-13.0927	of | communicative
0.210084	-13.0928	with | common
0.5625	-13.0929	trees | root
0.333333	-13.0929	partly | by
0.206612	-13.0929	• | all
0.857143	-13.093	supported | fellowship.
0.857143	-13.093	computation | context.
0.478261	-13.093	in | deployed
0.267606	-13.0931	problem | given
0.267606	-13.0931	potential | by
0.0798526	-13.0932	for | models
0.313725	-13.0932	paradigm | a
0.478261	-13.0933	0.04 | 0.02
0.298246	-13.0933	pi | a
0.23913	-13.0933	(marcus | al.,
0.102	-13.0934	and | 2004;
0.231579	-13.0935	punctuation | in
0.7	-13.0936	other | besides
0.175758	-13.0936	is | information.
0.0795181	-13.0937	we also | to
0.133106	-13.0937	that | part
0.368421	-13.0937	quality | their
0.303571	-13.0939	baseline | proposed
0.7	-13.0941	performs | performs
0.139098	-13.0941	is | combination
0.529412	-13.0941	conforms | the
0.111374	-13.0942	frame | the
0.0737279	-13.0942	we | english
0.276923	-13.0942	get | on
0.106987	-13.0942	high | in
0.253165	-13.0943	without | which
0.25641	-13.0944	pronouns | in
0.0876672	-13.0944	the | correctly
0.244186	-13.0944	strings | a
0.7	-13.0945	(at | one
0.303571	-13.0946	sense | one
0.5625	-13.0947	directly | features.
0.413793	-13.0947	as | blog
0.382353	-13.0947	it | never
0.115385	-13.0948	results | test
0.615385	-13.0948	proposes | new
0.413793	-13.095	microsoft | redmond,
0.114504	-13.095	the | update
0.0981481	-13.0952	the | distributions | the
0.190141	-13.0952	language, | a
0.247059	-13.0952	it | perform
0.247059	-13.0952	informative | and
0.17341	-13.0953	a | computer
0.263889	-13.0954	hybrid | in
0.183007	-13.0955	this | inference
0.293103	-13.0957	predicate-argument | and
0.293103	-13.0957	this | unique
0.5	-13.0957	its | i.e.,
0.092562	-13.0958	particular | of
0.203252	-13.0958	is | problem.
0.275362	-13.0959	the | phrasebased
0.75	-13.0959	speakers’ | proficiency
0.75	-13.0959	bracketleftbigg | bracketrightbigg
0.75	-13.0959	penn | cpos
0.298246	-13.0959	the | down.
0.208333	-13.096	most | work
0.263889	-13.0961	and | infer
0.15493	-13.0962	of | introduced
0.333333	-13.0962	all | together
0.0790419	-13.0963	of | length
0.20155	-13.0963	like | which
0.75	-13.0964	probabilistic | (hofmann,
0.75	-13.0964	biological | molecular
0.190141	-13.0965	first, we | a
0.227723	-13.0966	exactly | in
0.75	-13.0968	computation | push
0.290323	-13.0969	set | available
0.181818	-13.0969	(a | and
0.108352	-13.097	such | not
0.529412	-13.0971	language-independent | for
0.5625	-13.0972	restrict | set
0.7	-13.0972	define | various
0.75	-13.0973	intended | convey
0.144	-13.0973	for | and | data.
0.298246	-13.0974	resource, | the
0.0967742	-13.0975	the | e.g.
0.44	-13.0975	our | ner
0.112745	-13.0975	is | its | is
0.0970696	-13.0975	we | see
0.0656692	-13.0976	of | systems
0.5	-13.0976	fixed | it
0.210526	-13.0977	thus, | can
0.75	-13.0978	thai | medicine
0.127796	-13.0978	prior | a
0.857143	-13.0979	outperforms | outperforms
0.615385	-13.0979	these | contribute
0.368421	-13.0979	the | heuristic,
0.7	-13.0981	instance | instance.
0.139098	-13.0981	the | rules,
0.0914928	-13.0981	in | type
0.35	-13.0983	in | life
0.144	-13.0984	the | derivations
0.7	-13.0985	tool | developed
0.106291	-13.0985	following | we
0.5	-13.0985	for | concerns
0.0685302	-13.0986	it | be | to
0.223301	-13.0987	denoting | the
0.75	-13.0988	coerced | cm
0.236559	-13.0989	gains | in
0.166667	-13.0989	employ | and
0.12462	-13.0989	were | which
0.210526	-13.0989	the | twice
0.7	-13.099	problem, | optimal
0.130293	-13.0991	however, | can
0.47619	-13.0991	required | information
0.428571	-13.0992	dataset | available
0.5625	-13.0994	segment | between
0.35	-13.0994	s | uniondbl
0.303571	-13.0995	the | xinhua
0.413793	-13.0997	for | hebrew
0.44	-13.0998	instead, | more
0.529412	-13.0998	a | solves
0.5	-13.0998	for | bigrams,
0.5	-13.0998	poses | for
0.121739	-13.0999	fully | the
0.204918	-13.0999	suggests | for
0.241379	-13.1	giving | of
0.0945017	-13.1003	there are | to
0.234043	-13.1004	a | precision.
0.234043	-13.1004	a | body
0.271429	-13.1004	found | have
0.0684685	-13.1004	et al., | for
0.413793	-13.1005	rules | generated
0.195489	-13.1006	that | semantics
0.413793	-13.1007	algorithm | e
0.203252	-13.1007	the | completed
0.7	-13.1008	arguments | relation.
0.276923	-13.1008	senses | words
0.5	-13.1008	identification | features
0.5	-13.1008	their | values.
0.271429	-13.1009	the | business
0.135714	-13.1009	are | very
0.615385	-13.101	context | e.g.,
0.47619	-13.101	encourages | of
0.529412	-13.1011	if | preferences
0.7	-13.1012	find | summaries
0.18543	-13.1012	actually | in
0.44	-13.1012	y | (x
0.333333	-13.1013	both | annotation
0.201613	-13.1014	yes | yes
0.615385	-13.1014	coherence | discourse
0.615385	-13.1014	process | illustrated
0.0924092	-13.1014	corpus | with
0.340909	-13.1014	is | stable
0.181818	-13.1014	the | non-terminal
0.190141	-13.1015	1, | that
0.303571	-13.1016	text | target
0.149123	-13.1017	and | nouns
0.340909	-13.1017	performance | significantly
0.144	-13.1017	3 | an
0.615385	-13.1018	our | faster.
0.263889	-13.1018	texts, | to
0.428571	-13.102	a | male
0.131313	-13.1021	second | is | of
0.0865801	-13.1021	in this paper, | of
0.114213	-13.1021	we | up
0.0810127	-13.1021	and | where | is
0.615385	-13.1022	represent | among
0.241379	-13.1023	are | short
0.174699	-13.1024	em | and
0.529412	-13.1025	external | used
0.615385	-13.1026	achieve | level
0.253165	-13.1027	contains | word
0.857143	-13.1028	vectors | vectors,
0.303571	-13.1028	hindi | for
0.5625	-13.103	different | and/or
0.0466975	-13.103	3 | the | of
0.340909	-13.103	of | fragments.
0.216216	-13.1031	proposed | translation
0.191489	-13.1031	nps | the
0.223301	-13.1032	of | applications.
0.25974	-13.1033	also | evaluation
0.052973	-13.1034	in | system
0.615385	-13.1034	times | much
0.75	-13.1035	dependency parsing | parsing.
0.143426	-13.1036	log | of
0.47619	-13.1037	manual | annotations
0.0436377	-13.1038	we use | the
0.1875	-13.1038	heuristic | to
0.857143	-13.1039	statistics | features.
0.25	-13.1039	they | them
0.428571	-13.1039	one | weight
0.5625	-13.104	using | bootstrap
0.285714	-13.1041	and | vary
0.121739	-13.1045	the | objects
0.285714	-13.1046	which | include
0.109557	-13.1046	the | interpretation
0.25	-13.1047	normalization | to
0.276923	-13.1047	the | sums
0.326087	-13.1048	texts | language
0.35	-13.1048	learned | model
0.158416	-13.1048	the | of | unit
0.231579	-13.1048	of | speech.
0.102041	-13.1049	top | and
0.32	-13.1049	strategy, | to
0.11007	-13.105	systems | on
0.119114	-13.1051	task | from
0.186207	-13.1052	on | examples
0.371429	-13.1052	the | option.
0.214286	-13.1052	and | (figure
0.5	-13.1052	positional | and
0.276923	-13.1054	at | state
0.340909	-13.1054	when | additional
0.5625	-13.1055	algorithm | minimum
0.75	-13.1055	involves | documents
0.303571	-13.1055	that | approximate
0.857143	-13.1056	following | (2008),
0.857143	-13.1056	gender | number,
0.174699	-13.1056	and | dependencies
0.333333	-13.1056	the | consistent,
0.167568	-13.1057	explanation | the
0.7	-13.1057	other | measures.
0.179487	-13.1058	we | contextual
0.340909	-13.1058	rated | and
0.19403	-13.1058	the | kb
0.382353	-13.1059	used | 2003)
0.303571	-13.1059	these | manually
0.857143	-13.1061	ask | questions
0.857143	-13.1061	terms | 100
0.7	-13.1061	2.2 | proposed
0.5625	-13.1062	is | compositional.
0.326087	-13.1062	they | linguistic
0.181818	-13.1062	procedure | for
0.125382	-13.1063	taking | of
0.210526	-13.1063	and | datasets
0.0960854	-13.1064	and | sets
0.44	-13.1065	morphological | only
0.4	-13.1065	co-occurrence | corpus
0.206612	-13.1065	an | way
0.307692	-13.1066	the | bank
0.161616	-13.1066	of | syntax
0.223301	-13.1067	(blei | al.,
0.175758	-13.1067	speakers | and
0.5	-13.1068	see | effect
0.1875	-13.1069	to address | a
0.75	-13.1069	we write | as
0.428571	-13.1069	pair | generate
0.615385	-13.1069	know | you
0.340909	-13.1069	propose | approach to
0.216216	-13.1071	for | implementation
0.128617	-13.1071	of | grammatical
0.857143	-13.1072	algorithm | identifies
0.857143	-13.1072	of | harvests
0.0266667	-13.1072	the | two
0.5	-13.1072	vb | vb
0.5625	-13.1073	user | response
0.125382	-13.1073	we | include
0.114213	-13.1073	the | assumption
0.615385	-13.1073	probability, | probability
0.142857	-13.1073	dependencies | and
0.298246	-13.1073	when | output
0.131313	-13.1074	in | assigned
0.210526	-13.1074	each | specific
0.393939	-13.1075	remains | be
0.0335366	-13.1076	is | from
0.172414	-13.1077	after | by
0.5625	-13.1077	from | woz
0.615385	-13.1077	different | classification.
0.857143	-13.1078	association | incorrect
0.75	-13.1079	identification | mwe
0.35	-13.1079	is | small,
0.298246	-13.1079	also | example
0.298246	-13.1079	7.2 | the
0.307692	-13.108	in | authors | do
0.5625	-13.108	patrick | and
0.15566	-13.108	that | translated
0.223301	-13.108	issue | is
0.615385	-13.1081	algorithm | correctly
0.28125	-13.1081	used, | a
0.0688073	-13.1081	a | would
0.5	-13.1081	of | rolesets
0.226804	-13.1082	looks | the
0.0427562	-13.1082	in | language
0.175758	-13.1082	develop | and
0.458333	-13.1084	less | because
0.5625	-13.1084	with | boundaries.
0.5625	-13.1084	lead | results.
0.413793	-13.1084	converting | a
0.0689972	-13.1085	of | methods
0.119444	-13.1085	it | information
0.253165	-13.1085	using | input
0.151111	-13.1085	are | form
0.413793	-13.1087	improvement | due
0.298246	-13.1087	supertagger | the
0.615385	-13.1089	is | implement,
0.371429	-13.109	can | space.
0.413793	-13.1092	this | production
0.188811	-13.1092	that | shared
0.7	-13.1093	benefited | from
0.75	-13.1093	tagging | resource
0.35	-13.1093	can be | across
0.857143	-13.1094	4.3 | final
0.076659	-13.1094	the | texts
0.5	-13.1094	pipeline | word
0.219048	-13.1095	to | return
0.151786	-13.1095	coefficient | the
0.458333	-13.1096	in | spaces.
0.288136	-13.1096	is | example.
0.216981	-13.1097	overall | by
0.5625	-13.1098	however | different
0.0981132	-13.1098	with | feature
0.285714	-13.11	in | transitions
0.368421	-13.11	which | give
0.260274	-13.1101	11 | for
0.358974	-13.1102	activity | and
0.129568	-13.1102	name | a
0.271429	-13.1103	a | tokens.
0.19697	-13.1103	from | output
0.340909	-13.1104	take | their
0.857143	-13.1105	similar | body
0.263889	-13.1105	yet, | the
0.25641	-13.1106	of | titles
0.192593	-13.1106	existing | as
0.285714	-13.1107	marked | are
0.529412	-13.1107	constraints | form
0.358974	-13.1107	the | composition.
0.5	-13.1107	process | base
0.0781065	-13.1108	vectors | the
0.0518868	-13.1108	and | between | and
0.214286	-13.1108	to | power
0.204918	-13.1108	the | curves
0.25	-13.1109	candidate | it
0.28125	-13.1109	finally, | results
0.271429	-13.111	matches | for
0.529412	-13.111	contains | tokens
0.105495	-13.111	use | data
0.7	-13.1111	past | 2009).
0.44	-13.1113	# | /
0.615385	-13.1113	standard | representation
0.172619	-13.1114	low | for
0.413793	-13.1114	using | constructed
0.25	-13.1114	to | subsequent
0.47619	-13.1114	and | 20.
0.0528322	-13.1115	the | alignment
0.140625	-13.1116	3) | the
0.104925	-13.1116	also | an
0.0982987	-13.1117	variables | the
0.142857	-13.1118	induced | the
0.340909	-13.1119	matrix | where
0.123494	-13.112	are | because
0.064257	-13.112	system | for
0.7	-13.112	led | development
0.28125	-13.112	the | statistic
0.12069	-13.112	existing | to
0.857143	-13.1122	because | calculation
0.164021	-13.1123	assume | are
0.313725	-13.1123	such | much
0.307692	-13.1125	clark, | and
0.115486	-13.1126	we | statistical
0.35	-13.1126	but | labeled
0.5	-13.1126	restricted | only
0.16	-13.1127	show that | that
0.105495	-13.1127	into | by
0.276923	-13.1127	each | calculated
0.263889	-13.1127	defined | section
0.382353	-13.1128	huge | in
0.0932203	-13.1129	is | relation
0.529412	-13.113	s | c,
0.428571	-13.113	added | feature
0.161458	-13.113	setting, | the
0.285714	-13.1131	align | in
0.293103	-13.1131	with | 2004).
0.267606	-13.1131	of | authors.
0.162304	-13.1133	report | on
0.35	-13.1133	poor | for
0.117486	-13.1133	a | aligned
0.168539	-13.1136	differences | are
0.371429	-13.1136	r1 | r3
0.529412	-13.1137	state | agenda
0.19697	-13.114	1999; | al.,
0.393939	-13.1141	next | these
0.428571	-13.1141	the | saved
0.307692	-13.1143	in | mentions.
0.529412	-13.1144	vary | different
0.212389	-13.1144	for | discussed
0.176829	-13.1145	combining | a
0.126183	-13.1145	sentence | an
0.458333	-13.1146	for | disambiguation,
0.0605413	-13.1146	compared | the
0.052861	-13.1148	in | sentence | the
0.11138	-13.1148	both | and | as
0.126183	-13.1149	back | the
0.206897	-13.1149	concept | as
0.5	-13.1149	text | various
0.101833	-13.115	used for | and
0.0900474	-13.1151	a | annotated
0.7	-13.1152	translation | augmentation
0.267606	-13.1152	for | valuable
0.5	-13.1152	0.79 | 0.79
0.219048	-13.1154	a | model | trained
0.175758	-13.1154	we use | features
0.272727	-13.1155	consists | sentences
0.428571	-13.1155	while | (e.g.
0.358974	-13.1155	events | other
0.22449	-13.1155	initial | are
0.134276	-13.1156	that | annotated
0.25641	-13.1157	strong | that
0.615385	-13.1157	(mcdonald | 2006)
0.333333	-13.1157	use | graph
0.47619	-13.1157	extensions | are
0.219048	-13.1158	to | learning.
0.326087	-13.1159	study | learning
0.253165	-13.116	becomes | and
0.47619	-13.116	type | token
0.110048	-13.116	from | source
0.857143	-13.1161	idiap | martigny,
0.857143	-13.1161	toward | enemy
0.857143	-13.1161	a114 | ∈{σ
0.857143	-13.1161	mvm | dpmm,
0.857143	-13.1161	tutor | student’s
0.219048	-13.1161	detecting | a
0.130872	-13.1161	we | our | using
0.413793	-13.1162	task | core
0.229167	-13.1162	using | function
0.229167	-13.1162	to | parameters.
0.0988593	-13.1163	good | of
0.5625	-13.1164	entities | text,
0.529412	-13.1165	vertex | its
0.615385	-13.1165	using | reviews.
0.253165	-13.1165	plan | our
0.32	-13.1165	require | which
0.5	-13.1165	english-french | and
0.75	-13.1166	produce | summaries
0.4	-13.1166	given | collection
0.195489	-13.1167	maps | the
0.5625	-13.1167	four | them
0.131944	-13.1167	into | each
0.4	-13.1168	recorded | of
0.7	-13.117	bag-of-words | feature
0.203252	-13.117	when | performance
0.75	-13.1171	students | tutors
0.75	-13.1171	hierarchical | constrained
0.192593	-13.1172	that | concept
0.413793	-13.1173	gratefully | support
0.35	-13.1174	an | ontology
0.529412	-13.1175	6 | paper.
0.571429	-13.1175	segment | speech
0.114583	-13.1176	and | per
0.44	-13.1176	that | frequency.
0.293103	-13.1177	1 | 1,
0.142857	-13.1178	implementation | and
0.571429	-13.1179	expressions | considered
0.138996	-13.118	general | that
0.615385	-13.1181	used | speech.
0.5	-13.1181	feature | mert
0.529412	-13.1182	comparison, | same
0.11138	-13.1183	quite | the
0.201613	-13.1184	engineering | of
0.134752	-13.1184	decide | the
0.13806	-13.1185	of | (i.e.
0.246914	-13.1185	the | blue
0.47619	-13.1185	that | in,
0.140625	-13.1186	as | many
0.177215	-13.1186	and | independent
0.109524	-13.1186	is | weight
0.147186	-13.1187	represents | in
0.06171	-13.1188	an | by
0.5625	-13.1189	proposed method | using
0.371429	-13.1189	features | global
0.212389	-13.1189	identify | by
0.35	-13.119	a | entry.
0.413793	-13.1191	since | annotated
0.5	-13.1191	11 | 18
0.44	-13.1193	for | adjectives,
0.571429	-13.1193	furthermore, | additional
0.75	-13.1195	2 | },
0.12	-13.1197	relatively | of
0.7	-13.1197	training | detector
0.11194	-13.1197	search | for
0.340909	-13.1197	words | unknown
0.125786	-13.1198	scores | are
0.15493	-13.1198	annotated | with
0.121037	-13.12	into | which
0.4	-13.12	as follows: | words
0.571429	-13.1201	chart | input
0.177215	-13.1202	a | way.
0.136531	-13.1202	two | they
0.28125	-13.1203	might | different
0.253165	-13.1203	components, | the
0.256757	-13.1205	more | evaluation
0.7	-13.1206	methods | finding
0.144628	-13.1206	the | suitable
0.141176	-13.1206	of | building
0.326087	-13.1208	a | generalize
0.413793	-13.121	its | context.
0.413793	-13.121	high | scores
0.326087	-13.121	switchboard | of
0.341463	-13.121	apply | algorithm
0.28125	-13.1211	on | surface
0.210526	-13.1211	students | in
0.100402	-13.1212	node | a
0.47619	-13.1212	require | large
0.0702439	-13.1213	these | with
0.136029	-13.1213	to | position
0.529412	-13.1213	zero | all
0.333333	-13.1213	et | 2011;
0.298246	-13.1214	is | probability.
0.19685	-13.1215	verbal | and
0.35	-13.1215	example | type
0.288136	-13.1215	to | communicative
0.301887	-13.1216	results | supervised
0.28125	-13.1216	finally, | all
0.340909	-13.1217	the | class-instance
0.0929054	-13.1219	conclusions | the
0.260274	-13.1219	inferences | the
0.208696	-13.122	percent | the
0.75	-13.122	follow | general
0.139535	-13.1223	of | time.
0.241379	-13.1224	were | the | corpus.
0.25	-13.1224	a | forward
0.10643	-13.1225	tweets | the
0.7	-13.1225	currently, | system
0.615385	-13.1225	while | view
0.253165	-13.1225	to | senses.
0.256757	-13.1225	(2006) | on
0.235955	-13.1226	coherence | of
0.263889	-13.1226	enough | be
0.179487	-13.1227	the | strength
0.301887	-13.1228	(2001) | and
0.7	-13.1229	features | characterize
0.107551	-13.123	of | (the
0.165746	-13.123	can | many
0.214953	-13.1231	is | constraint
0.0412475	-13.1232	algorithm | the
0.44	-13.1233	set | examples,
0.44	-13.1233	predict | will
0.615385	-13.1233	because | ways
0.253165	-13.1233	smith, | and
0.0710711	-13.1234	the | n-gram
0.75	-13.1234	table 4: | (in
0.341463	-13.1234	the | collected.
0.571429	-13.1234	id | ge
0.256757	-13.1235	of | insertion
0.0349366	-13.1235	a | that | to
0.44	-13.1236	method | average
0.138462	-13.1237	we | parameter
0.7	-13.1238	respect, | to
0.75	-13.1239	to produce | final
0.341463	-13.1239	all | presented
0.146552	-13.124	line | in
0.371429	-13.124	and | bring
0.22449	-13.124	the | exchange
0.293103	-13.124	frame | semantic
0.458333	-13.1242	utterance | but
0.243902	-13.1244	and | decisions
0.140078	-13.1245	a | matching
0.132404	-13.1245	scores | as
0.615385	-13.1245	mixture | multiple
0.175758	-13.1245	mi | the
0.0893417	-13.1246	statistical | of
0.115486	-13.1246	this work | by
0.571429	-13.1246	analysis | nlp
0.0516333	-13.1247	to | semantic
0.5	-13.1247	verb | infinitive
0.123123	-13.1248	for | relations
0.44	-13.1248	other | nlp
0.75	-13.1249	consisting | pairs,
0.4	-13.1249	translate | translation
0.571429	-13.1249	quantify | between
0.260274	-13.1249	where | input
0.216981	-13.125	since | system
0.166667	-13.1252	the | differences between
0.111386	-13.1253	for | corresponding
0.283333	-13.1253	in | decrease
0.0735608	-13.1254	the | topics
0.28125	-13.1255	an | generation
0.28125	-13.1255	in | phase
0.413793	-13.1256	as | language model.
0.167598	-13.1257	sentences, | a
0.19403	-13.1257	we | seed
0.123123	-13.1258	similar | as
0.529412	-13.1258	is | operate
0.0397269	-13.1259	with | as
0.243902	-13.126	whether | was
0.0831063	-13.126	created | the
0.226804	-13.1262	concepts | are
0.571429	-13.1264	(boyer, | et
0.272727	-13.1265	answer | as
0.144628	-13.1265	the | interface
0.319149	-13.1265	a | statement
0.458333	-13.1266	are | stored
0.636364	-13.1266	prlg | hmm
0.158416	-13.1266	training, | and
0.115903	-13.1267	after | we
0.75	-13.1269	experiment | instances
0.238636	-13.1269	to | grammar.
0.313725	-13.127	ours | and
0.147826	-13.1273	and | before
0.5	-13.1273	until | new
0.636364	-13.1274	besides | other
0.301887	-13.1274	given | d
0.7	-13.1275	clustering | features.
0.7	-13.1275	evidence | lexical
0.413793	-13.1275	that | status
0.268657	-13.1275	each | word,
0.571429	-13.1276	ge | mooney
0.293103	-13.1276	tied | the
0.5625	-13.1277	dataset | many
0.214953	-13.1277	conll | and
0.75	-13.1278	represents | towards
0.148472	-13.128	for | (e.g.
0.191176	-13.1281	that | accurate
0.413793	-13.1281	wsj | training
0.0636071	-13.1281	can | or
0.47619	-13.1281	relation | mentions
0.529412	-13.1282	instances, | training
0.326087	-13.1282	to | parses.
0.636364	-13.1283	expected | evidences
0.341463	-13.1283	point | all
0.167598	-13.1287	features | dependency
0.326087	-13.1287	for | wsd
0.172619	-13.1288	one | their
0.44	-13.1288	length | sentence.
0.75	-13.1288	north | state
0.268657	-13.1288	hence | for
0.529412	-13.1289	include | type
0.136029	-13.1291	j | i
0.5625	-13.1292	phylogenetic | of
0.205128	-13.1292	is | segment
0.293103	-13.1292	since | new
0.7	-13.1293	dictionaries | lexical
0.210526	-13.1294	mixture | and
0.195312	-13.1295	corpus, | with
0.636364	-13.1295	7th | on
0.293103	-13.1295	components | are
0.256757	-13.1296	with | numbers
0.75	-13.1298	algorithm | merges
0.25	-13.1298	thus, | which
0.571429	-13.1298	described | type
0.636364	-13.1299	sentences, | labeled
0.28125	-13.1299	to | nouns.
0.47619	-13.13	draws | and
0.47619	-13.13	size | unique
0.0966543	-13.1302	noun | in
0.145923	-13.1303	the | in | domain.
0.75	-13.1303	convert | format
0.147826	-13.1304	has | other
0.458333	-13.1305	if | antecedent
0.161458	-13.1305	baseline | using
0.382353	-13.1305	with | β
0.107551	-13.1306	we show | of
0.333333	-13.1307	we | syntax-based
0.293103	-13.1307	will | possible
0.307692	-13.1308	using | algorithms
0.172619	-13.1309	sl | the
0.120344	-13.1309	the | 2).
0.276923	-13.1309	metric, | of
0.172619	-13.131	to | understanding
0.4	-13.131	languages | parallel
0.0833333	-13.131	these | not
0.382353	-13.131	as | target.
0.102725	-13.1311	is | corresponding
0.35	-13.1311	contain | about
0.176101	-13.1313	however, | do
0.571429	-13.1313	markov | minimal
0.263889	-13.1313	study | this
0.458333	-13.1314	titles | with
0.128289	-13.1314	currently | the
0.44	-13.1314	only | (or
0.256757	-13.1314	each | randomly
0.067325	-13.1315	phrases | the
0.333333	-13.1316	thresholds | and
0.0676417	-13.1316	we | that | not
0.19403	-13.1316	sentence | was
0.44	-13.1317	translation | japanese
0.214953	-13.1318	many | words
0.256757	-13.1319	which | among
0.253333	-13.1319	5 conclusions | of
0.529412	-13.132	indirect | or
0.4	-13.132	on | semeval
0.132404	-13.132	a | opinion
0.5	-13.1322	classifiers | perform
0.5	-13.1322	teachers | to
0.382353	-13.1323	an | unit
0.123839	-13.1324	is | fact
0.243902	-13.1324	node, | of
0.571429	-13.1324	applied | combination
0.4	-13.1325	fundamental | for
0.423077	-13.1325	2 | review
0.301887	-13.1325	in | incorrectly
0.0996016	-13.1326	selected | and
0.168539	-13.1326	that | increase
0.615385	-13.1326	graph | predicates
0.148472	-13.1327	dataset | for
0.301887	-13.1327	pearson’s | and
0.125392	-13.1327	generation | a
0.106818	-13.1327	different | from
0.75	-13.1328	compound | compound.
0.571429	-13.1328	this | coarse-grained
0.47619	-13.1328	users | tweets
0.7	-13.133	name, | name
0.529412	-13.1331	results | held-out
0.307692	-13.1332	system | outputs
0.571429	-13.1332	program, | a
0.184932	-13.1333	exploit | and
0.4	-13.1333	subjectivity | that
0.529412	-13.1334	and | 2005):
0.7	-13.1335	learning | vectors
0.276923	-13.1335	a | request
0.156098	-13.1337	acknowledgments | work | by
0.0962963	-13.1337	the | equal
0.12	-13.134	pairs, | the
0.205128	-13.134	spaces | the
0.458333	-13.1341	background | related
0.326087	-13.1342	know | or
0.0406158	-13.1342	have | in
0.156863	-13.1343	vertices | the
0.5	-13.1345	with | bleu.
0.166667	-13.1346	in | application
0.382353	-13.1346	rules | aligned
0.235955	-13.1346	model | trained | a
0.134545	-13.1347	the | spelling
0.458333	-13.1347	procedure, | in
0.0401671	-13.1348	to | more
0.268657	-13.1348	identifying | with
0.219048	-13.1349	have | relation
0.7	-13.1349	by | marks.
0.7	-13.1349	b, | d
0.35	-13.135	whether | any
0.0837989	-13.135	best | in
0.47619	-13.135	throughout | with
0.156863	-13.1351	2.2 | and
0.7	-13.1353	annotated | tweets
0.120235	-13.1353	material | the
0.235955	-13.1353	flat | and
0.238636	-13.1353	to | arguments.
0.4	-13.1354	errors | english
0.371429	-13.1354	and | remain
0.240964	-13.1354	above | by
0.15311	-13.1355	extracted | are
0.170588	-13.1355	works | a
0.75	-13.1357	large | sets,
0.0897436	-13.1357	model | word
0.0687382	-13.1358	the | distance
0.458333	-13.1359	probability | equation
0.219048	-13.1359	two | systems
0.175	-13.1359	after | on
0.229167	-13.1359	this | segmentation
0.229167	-13.1359	dialog | and
0.4	-13.136	about | specific
0.176101	-13.136	4.3 | in
0.125	-13.136	have | one
0.44	-13.1361	therefore, | should
0.382353	-13.1361	parent | its
0.636364	-13.1362	handwritten | learned
0.529412	-13.1362	resulting | resulting
0.25	-13.1364	to | efficiently
0.0634538	-13.1364	as | it
0.307692	-13.1365	used | supervised
0.307692	-13.1365	minimum | for
0.0751121	-13.1365	of | structure
0.0246407	-13.1366	the | is | for
0.146552	-13.1368	function, | the
0.272727	-13.1368	which | complex
0.256757	-13.1368	the | collaboration
0.159794	-13.137	the | decrease
0.206897	-13.1371	for | usually
0.75	-13.1372	(c) | semantic
0.105145	-13.1372	selection | to
0.188406	-13.1375	the | category,
0.44	-13.1375	for | descriptions.
0.636364	-13.1375	implication | that
0.326087	-13.1375	a | group.
0.111111	-13.1375	a | labels
0.313725	-13.1375	the | file.
0.214953	-13.1378	is | almost
0.111386	-13.1379	majority | the
0.333333	-13.138	eliminate | of
0.7	-13.1381	model | infers
0.296296	-13.1382	numerous | and
0.387097	-13.1383	summary | document
0.636364	-13.1383	t. | a.
0.0196018	-13.1383	the | which | the
0.073913	-13.1384	the | version | the
0.201681	-13.1384	is | description
0.235955	-13.1384	for | 2004).
0.169591	-13.1387	f | e
0.615385	-13.1387	creation | annotated
0.164835	-13.1387	studied | and
0.238636	-13.1387	in | samples
0.206897	-13.1387	the | connecting
0.44	-13.139	an | location
0.193798	-13.139	the | matrix.
0.1417	-13.139	morphological | in
0.333333	-13.1391	word | resulting
0.0752566	-13.1392	example | in
0.18	-13.1392	and | >
0.319149	-13.1393	the | reliable.
0.25	-13.1394	this | interesting
0.47619	-13.1394	a | policy.
0.47619	-13.1394	convention | to
0.615385	-13.1395	empirically | different
0.5	-13.1395	fl | @
0.636364	-13.1396	references | m.
0.293103	-13.1396	using | per
0.107477	-13.1397	the | around
0.123839	-13.1398	the | had | and
0.341463	-13.1398	chooses | and
0.130137	-13.14	in | 2007).
0.636364	-13.14	from | history.
0.16763	-13.14	the | stack.
0.571429	-13.14	exclude | are
0.22449	-13.1401	results | in | 5.
0.272727	-13.1402	specify | in
0.326087	-13.1403	value | given
0.18	-13.1403	finally, | our
0.0659722	-13.1403	have been | the
0.216981	-13.1403	gi | patterns
0.7	-13.1404	time | actions
0.571429	-13.1404	newswire | corpus
0.44	-13.1407	kappa | kappa
0.177215	-13.1407	in | understanding
0.423077	-13.1407	for | cut
0.571429	-13.1408	an | whenever
0.0689655	-13.1408	of | which is
0.240964	-13.141	promising | to
0.615385	-13.1411	assign | weight
0.615385	-13.1411	ministry | education
0.615385	-13.1411	a | tuples.
0.212963	-13.1411	of | tweets.
0.228261	-13.1412	given | 1.
0.0636215	-13.1412	the | side
0.198413	-13.1414	for | 2,
0.4	-13.1415	with | segments
0.222222	-13.1417	dictionary | from
0.75	-13.1417	document | nd
0.272727	-13.1417	training | parser
0.361111	-13.1418	perhaps | are
0.18705	-13.1419	each | set.
0.131034	-13.142	is | up
0.233333	-13.1421	also | human
0.4	-13.1421	a | part.
0.423077	-13.1421	translation | penalty
0.272727	-13.1422	to | immediately
0.288136	-13.1422	might | but
0.278689	-13.1423	paper | structured
0.0786517	-13.1424	added | the
0.0866261	-13.1426	two | from
0.127451	-13.1427	and | performance.
0.413793	-13.1428	punctuation | or
0.0656871	-13.1429	correct | the
0.102345	-13.1429	evaluated | of
0.301887	-13.1429	cases, | not
0.185714	-13.1429	jointly | the
0.278689	-13.143	when | size
0.0745273	-13.1431	goal | the
0.183673	-13.1431	methods | which
0.5	-13.1431	annotators | them
0.142276	-13.1433	tweets | and
0.571429	-13.1434	represent | binary
0.0536952	-13.1434	extracted | the
0.361111	-13.1437	et | 1993),
0.7	-13.1437	official | conll
0.7	-13.1437	be | manually.
0.636364	-13.1438	our | implements
0.333333	-13.144	70 | (%)
0.293103	-13.1441	of | instruction
0.5	-13.1441	adj | adj
0.195312	-13.1442	and | reviews
1	-13.1442	rep | easy–to–read
1	-13.1442	bullets | bullets
1	-13.1442	east | west
1	-13.1442	histograms | formulation
1	-13.1442	agent | rewarded
1	-13.1442	sf_n | wsf
1	-13.1442	states, | actions,
1	-13.1442	orders | magnitude.
1	-13.1442	acl | workshop,
1	-13.1442	australia, | mexico,
1	-13.1442	pesos, | kroner,
1	-13.1442	green | manning
1	-13.1442	arc-standard | arc-eager
1	-13.1442	views | findings are
1	-13.1442	nlp | uned
1	-13.1442	mohammad | turney,
1	-13.1442	intraand | distance
1	-13.1442	prednisone, | effexor,
1	-13.1442	gambell | yang
1	-13.1442	policies, | defense.
1	-13.1442	(top | vp
1	-13.1442	(johansson | 2010b),
1	-13.1442	partial | undefined
1	-13.1442	directed | sentences)
1	-13.1442	ohio | columbus,
1	-13.1442	yen, | pesetas,
1	-13.1442	128 | 1024
1	-13.1442	yellow | pink
1	-13.1442	surdeanu, | marquez
1	-13.1442	(%) | 1.00
1	-13.1442	dark | light
1	-13.1442	ha, | walis,
1	-13.1442	federico, | zens,
1	-13.1442	hoang, | shen,
1	-13.1442	θq | kb
1	-13.1442	guthrie | hepple
1	-13.1442	hp03 | hp04
1	-13.1442	languages: | english.
1	-13.1442	slt | berkeley,
1	-13.1442	ur | lr
1	-13.1442	monologue | dialogue.
1	-13.1442	φcomm | →utt0
1	-13.1442	(post | gildea,
1	-13.1442	(lefever | hoste,
1	-13.1442	past | barn
1	-13.1442	luyckx | daelemans,
1	-13.1442	cow | fork.
1	-13.1442	c: | o:
1	-13.1442	-4 | -1
1	-13.1442	hyp: | corr:
1	-13.1442	indian | india
1	-13.1442	pain | neck
1	-13.1442	ruthenium, | lead,...
1	-13.1442	(p), | (r),
1	-13.1442	magnesium, | sulfur,
1	-13.1442	t., | k.
1	-13.1442	α1 | α2
1	-13.1442	identity | }.
1	-13.1442	correspondence | (scl)
1	-13.1442	[λ1|i], | )⇒(
1	-13.1442	official | iarpa,
1	-13.1442	universitat | val`encia
1	-13.1442	google | award.
1	-13.1442	f | -score
1	-13.1442	dollars, | escudos,
1	-13.1442	trigrams | trigrams
1	-13.1442	pushed | serbian
1	-13.1442	frank, | pfahringer,
1	-13.1442	pos-tags, | pos-tags
1	-13.1442	setting | mt05
1	-13.1442	oxford | oxford
1	-13.1442	r-prec | r@5
1	-13.1442	reinforcement | mappings.
1	-13.1442	averaged | inter-
1	-13.1442	satta | peserico
1	-13.1442	rice, | grapes,...
1	-13.1442	(cormode | muthukrishnan,
1	-13.1442	emoticons | hashtags
1	-13.1442	palladium, | borium,
1	-13.1442	rivaling | methods.
1	-13.1442	nerba | nerdic
1	-13.1442	pr | [−1,
0.529412	-13.1443	however, | well.
0.2	-13.1443	from | analysis
0.191176	-13.1444	denotes | in
0.260274	-13.1445	f-score | on
0.0990099	-13.1445	this | us | the
0.7	-13.1446	which | precision,
0.387097	-13.1447	prediction | features
0.75	-13.1447	expected | find
0.833333	-13.1447	europarl | (koehn, 2005),
0.833333	-13.1447	composing | primitives.
0.833333	-13.1447	received: | revised
0.833333	-13.1447	bleu, | ter.
0.833333	-13.1447	p(c|e).
0.833333	-13.1447	tesla-m | tesla-m
0.833333	-13.1447	quantities | symbol.
0.833333	-13.1447	co-occurrence | context-vector
0.833333	-13.1447	practice | 2001;
0.833333	-13.1447	home | john
0.833333	-13.1447	calls | calls
0.833333	-13.1447	development, | engineering
0.0841655	-13.1448	rule | a
0.156566	-13.1448	supervised | for
0.529412	-13.145	learns | data
0.230769	-13.1451	performance | baseline
0.111392	-13.1452	from | system
0.140562	-13.1452	path | and
0.326087	-13.1453	exploited | in
0.833333	-13.1453	spelling | correction,
0.833333	-13.1453	s., | j.
0.833333	-13.1453	statistical | (smt).
0.833333	-13.1453	mrc | database
0.833333	-13.1453	community | communities
0.833333	-13.1453	views, | expressed
0.833333	-13.1453	2004 | 2005
0.833333	-13.1453	ci | cj.
0.833333	-13.1453	uses | texttiling
0.454545	-13.1454	binary | over
0.233333	-13.1454	a | step,
0.529412	-13.1454	method | evaluating
0.103448	-13.1455	that | high
0.168605	-13.1455	performance | performance | the
0.5	-13.1455	large | vocabulary
0.238095	-13.1456	also | many
0.0796915	-13.1456	binary | the
0.240964	-13.1458	of | paraphrases.
0.833333	-13.1458	human | instructors
0.833333	-13.1458	concepts, | concepts.
0.833333	-13.1458	augmentative | communication
0.833333	-13.1458	people | disabilities
0.833333	-13.1458	(wiebe | 2005).
0.833333	-13.1458	# % | *.
0.833333	-13.1458	alleviates | problem
0.127451	-13.1458	measure | is
0.0957643	-13.1459	of | processing
0.35	-13.146	use | technique
0.178808	-13.1462	not | much
0.454545	-13.1464	explore | improve
0.156566	-13.1464	phrase | translation
0.636364	-13.1464	values | k,
0.833333	-13.1464	database | english.
0.833333	-13.1464	denis | (2007),
0.833333	-13.1464	2009 | 2010.
0.264706	-13.1464	system | them
0.0535406	-13.1465	the | sentence.
0.454545	-13.1467	the | extensions.
0.166667	-13.1467	features: | of
0.75	-13.1467	similar | applies
0.75	-13.1467	extractions | or
0.132143	-13.1468	above | a
0.636364	-13.1468	sufficient | property
0.571429	-13.1468	by | translators
0.192308	-13.1469	used | new
0.423077	-13.1469	perplexity | models
0.833333	-13.1469	joint | pos-tagger
0.371429	-13.147	of | edge.
0.75	-13.1472	as | by-product
0.75	-13.1472	however, | guarantee
0.18	-13.1472	interesting | is
0.164835	-13.1472	mean | for
0.278689	-13.1475	demonstrate | can
0.44	-13.1475	and | calculated.
0.529412	-13.1475	each | instance.
0.183673	-13.1475	while | they
0.833333	-13.1475	equivalence | co-reference
0.833333	-13.1475	numerical | category
0.833333	-13.1475	dr. | her
0.833333	-13.1475	integration | wsd
0.0872894	-13.1476	form | and
0.636364	-13.1477	for | one-to-one
0.35	-13.1478	in | phone
0.7	-13.1479	english | urdu
0.529412	-13.1479	metaphor | a
0.4	-13.1479	given | very
0.319149	-13.148	constructed | each
0.0762911	-13.148	of | without
0.571429	-13.148	it | function.
0.636364	-13.1481	these | problems.
0.833333	-13.1481	ordered | ordered
0.5	-13.1481	philip | and
0.75	-13.1482	t-test | significantly
0.154589	-13.1482	that | during
0.423077	-13.1483	(12) | =
0.454545	-13.1485	fixed | this
0.25	-13.1485	trained | news
0.44	-13.1486	addition | both
0.18	-13.1486	is | finite
0.333333	-13.1486	pos | tagger
0.833333	-13.1486	haghighi | (2010)
0.75	-13.1487	head | dependents
0.25	-13.1487	years | to
0.243902	-13.1488	classifier | used
0.423077	-13.1488	vocabulary | corpus
0.571429	-13.1488	rwth | translation
0.529412	-13.1489	shown | 2007).
0.615385	-13.1489	models | actions
0.47619	-13.1489	restricted | set
0.301887	-13.149	the | best.
0.107728	-13.149	goal | a
0.382353	-13.1491	and | usefulness
0.145923	-13.1492	model | semantic
0.833333	-13.1492	smoothing | background
0.163043	-13.1492	relatively | in
0.7	-13.1493	- | cloud
0.636364	-13.1494	perform | multiple
0.0415793	-13.1494	the | context
0.288136	-13.1495	are | 2002).
0.5	-13.1495	identical | different
0.75	-13.1497	unigram | tb
0.125806	-13.1497	of | document.
0.833333	-13.1497	ravi | knight
0.833333	-13.1497	edit | deletion,
0.833333	-13.1497	funded | (7th
0.150943	-13.1498	“the | the
0.44	-13.1498	the | gene/protein
0.382353	-13.1498	methods | context
0.139442	-13.1499	second, | a
0.260274	-13.1499	the | effort.
0.529412	-13.15	given | written
0.301887	-13.15	predicting | on
0.170588	-13.1501	this | measure
0.129252	-13.1502	we | that | these
0.833333	-13.1503	svr-based | components
0.833333	-13.1503	head | dependent,
0.833333	-13.1503	has | advantages:
0.833333	-13.1503	librarian | specific
0.833333	-13.1503	povo | italy
0.246914	-13.1504	proposed | two
0.283333	-13.1506	we | 3).
0.253333	-13.1506	in | files
0.7	-13.1507	r, | all
0.7	-13.1507	corpus | consistency
0.571429	-13.1507	n | nsummationdisplay
0.158163	-13.1508	and | summaries
0.361111	-13.1509	will | negative
0.238095	-13.1509	which | applied
0.833333	-13.1509	marina | ca,
0.833333	-13.1509	haitian | english.
0.16129	-13.1509	for | sentence,
0.152381	-13.151	variant | the
0.20339	-13.151	classifier | an
0.75	-13.1512	only | worst
0.4	-13.1512	in | f-score.
0.103226	-13.1512	an | language
0.333333	-13.1513	on | learning.
0.4	-13.1514	allowed | or
0.127517	-13.1515	for | state
0.47619	-13.1515	and | republic
0.7	-13.1516	choosing | probability
0.7	-13.1516	where | changes
0.44	-13.1516	previous | approaches
0.173913	-13.1516	nonterminals | the
0.16129	-13.1517	and | get
0.529412	-13.1518	an | proficiency
0.571429	-13.1518	constraints, | each
0.5	-13.1518	parameters | k
0.5	-13.1518	learning | 2002).
0.833333	-13.152	post-hoc | showed
0.833333	-13.152	threshold | 0.5.
0.833333	-13.152	short | 140
0.189781	-13.1521	and | 1).
0.4	-13.1522	results | filtering
0.615385	-13.1522	knowledge about | between
0.571429	-13.1522	clustering | group
0.296296	-13.1524	they | single
0.130742	-13.1524	= | ∈
0.0897227	-13.1524	points | the
0.387097	-13.1525	linear interpolation | and
0.833333	-13.1525	transcripts | spoken
0.272727	-13.1526	results | approaches
0.0542959	-13.1526	the | words.
0.288136	-13.1526	a | synonym
0.75	-13.1527	for | german→english
0.127946	-13.1527	are | them
0.256757	-13.1529	evaluation | sentences
0.341463	-13.1531	looks | a
0.833333	-13.1531	wordnet, | only
0.217822	-13.1532	that | surface
0.75	-13.1532	polarity | polarities
0.134831	-13.1533	translation | target
0.278689	-13.1534	w, | and
0.423077	-13.1534	although | them
0.529412	-13.1535	to | (riloff
0.0277086	-13.1535	system | the
0.636364	-13.1536	one | learnability
0.209091	-13.1536	reasoning | the
0.0872274	-13.1536	type | in
0.0767386	-13.1537	shows | for
0.307692	-13.1537	in | extra
0.44	-13.1537	only | change
0.205128	-13.1537	reveals | the
0.175	-13.1538	the | significant.
0.75	-13.1538	semantic | verbs.
0.326087	-13.1538	score | pair
0.228261	-13.1539	missing | in
0.615385	-13.1539	despite | accuracy
0.35	-13.1539	propose | syntactic
0.278689	-13.154	unsupervised | methods
0.201681	-13.154	subjectivity | and
0.141129	-13.154	gives | and
0.296296	-13.154	simply | each
0.636364	-13.1541	used | fbis
0.833333	-13.1542	has been | army
0.361111	-13.1543	functional | are
0.75	-13.1543	expectation-maximization | (dempster
0.144068	-13.1544	this | best
0.112861	-13.1545	the | incorrect
0.571429	-13.1545	than | grammatical
0.288136	-13.1546	entities, | a
0.089141	-13.1546	the | potential
0.130282	-13.1547	high | for
0.222222	-13.1548	between | those
0.0804749	-13.1548	only | with
0.7	-13.1549	sorted | order
0.636364	-13.1549	and | campaigns
0.636364	-13.1549	unigram | models.
0.319149	-13.1549	between | words,
0.164835	-13.155	in | typically
0.207207	-13.155	synchronous | and
0.185714	-13.155	40 | 60
0.152381	-13.1551	3.3 | and
0.44	-13.1551	parses | gold
0.44	-13.1551	submissions | and
0.5	-13.1552	reliability | measure
0.529412	-13.1553	if | linked
0.135849	-13.1553	this | three
0.184397	-13.1554	probability | over
0.0475531	-13.1554	however, | the | of
0.333333	-13.1554	the | heldout
0.106729	-13.1556	we | found
0.296296	-13.1556	in | realistic
0.153846	-13.1557	are | high
0.454545	-13.1558	were | ambiguous
0.361111	-13.1558	variant | which
0.636364	-13.1558	are | specific,
0.636364	-13.1558	features | optimized
0.14346	-13.1558	approximation | the
0.7	-13.1559	using | switchboard
0.423077	-13.1559	event | extracted
0.105991	-13.1559	value | in
0.15736	-13.1559	m | n
0.833333	-13.1559	huffman | grid.
0.333333	-13.156	for | context-free
0.0982318	-13.156	all | words
0.47619	-13.156	medlda | for
0.571429	-13.1561	algorithm | returns
0.387097	-13.1562	an example | where
0.19685	-13.1562	that | global
0.222222	-13.1563	collected | for
0.75	-13.1563	extractive | paraphrasing
0.22449	-13.1563	a | pairwise
0.183099	-13.1564	provides | that
0.15311	-13.1565	this, | of
0.240964	-13.1565	an | lexicon
0.19685	-13.1565	straightforward | of
0.833333	-13.1565	constraints | constraints,
0.0434448	-13.1566	is | word
0.107399	-13.1566	tested | the
0.198347	-13.1566	method, | in
0.290909	-13.1567	and | diverse
0.3125	-13.1568	number | random
0.75	-13.1568	broke | a
0.272727	-13.1568	the | keys
0.233333	-13.1568	are | enough
0.5	-13.1569	of | acm
0.454545	-13.157	calculating | between
0.222222	-13.157	sms | and
0.155779	-13.157	we | matching
0.164773	-13.1571	the | simplified
0.44	-13.1572	will | them.
0.341463	-13.1572	then | up
0.151659	-13.1572	the | baseline,
0.454545	-13.1573	achieved | overall
0.198347	-13.1573	the | pool
0.272727	-13.1574	toolkit | al.,
0.387097	-13.1575	the | gap.
0.833333	-13.1576	letters | symbols
0.120482	-13.1577	evaluation | our
0.217822	-13.1577	as | problem.
0.0718057	-13.1577	made | the
0.205128	-13.1577	shows | when
0.529412	-13.1578	tense | present
0.133094	-13.1578	for | generated
0.371429	-13.158	in | sense, | to
0.264706	-13.158	we demonstrate | a
0.0614912	-13.1581	we | to | our
0.128814	-13.1582	beginning | the
0.833333	-13.1582	fail | produce
0.833333	-13.1582	language | (lm),
0.20339	-13.1582	wordnet | as
0.0852018	-13.1582	the | translation.
0.134328	-13.1583	feature | word
0.75	-13.1583	two | harmonic
0.75	-13.1588	unsupervised | relations.
0.158163	-13.1589	x, | the
0.5	-13.1589	key | systems
0.319149	-13.159	have | them.
0.283333	-13.159	mappings | a
0.18705	-13.1592	which | better
0.75	-13.1593	prosodic | duration
0.153846	-13.1593	after | are
0.22	-13.1593	developed | al.,
0.5	-13.1593	objects | have
0.168605	-13.1594	the | letters
0.19084	-13.1594	conditions | of
0.211009	-13.1596	classes | are
0.264706	-13.1596	and | simplification
0.307692	-13.1597	ar | ar
0.7	-13.1597	also | extra
0.105505	-13.1597	for | model.
0.615385	-13.1597	missing | marker
0.615385	-13.1597	device | with
0.13834	-13.1597	and | decoding
0.4	-13.1598	5 | paper.
0.17284	-13.1598	the | ).
0.0645709	-13.1599	types | and
0.0469738	-13.16	in | two
0.151659	-13.16	we | results.
0.107914	-13.16	1 | 5
0.290909	-13.16	subsequent | in
0.162162	-13.1601	features | context
0.7	-13.1601	learner | produce
0.7	-13.1601	of | yet.
0.326087	-13.1602	finally, | also
0.246753	-13.1603	treebanks | the
0.146018	-13.1603	interest | and
0.454545	-13.1604	systems | multiple
0.137795	-13.1605	the | asked
0.423077	-13.1605	their | suggestions.
0.141079	-13.1605	image | the
0.35	-13.1608	on | subsets
0.238095	-13.1609	30 | 60 | 80
0.283333	-13.1609	data | comparable
0.3125	-13.161	or | name
0.454545	-13.161	these | could | be
0.387097	-13.1611	the | outcome.
0.133829	-13.1611	the | sufficient
0.0982318	-13.1611	accuracy | for
0.0781841	-13.1613	document | a
0.25	-13.1613	learnt | the
0.164773	-13.1614	we | count
0.215686	-13.1614	production | of
0.47619	-13.1614	only | surface
0.205357	-13.1615	but | does
0.833333	-13.1615	often | 1999).
0.211009	-13.1616	researchers | a
0.7	-13.1616	speakers | words
0.7	-13.1616	system | predefined
0.205357	-13.1616	we | filtering
0.141667	-13.1617	computational linguistics | department
0.211009	-13.1618	of | scores,
0.0365621	-13.1618	these | and
0.246753	-13.1618	of | recognizing
0.146018	-13.1618	to | properties
0.290909	-13.1618	an example | which
0.175	-13.162	course | the
0.529412	-13.1621	each | topic,
0.833333	-13.1621	(agarwal | al.,
0.132616	-13.1622	word | sentence
0.10084	-13.1622	models | are
0.207207	-13.1622	propbank | the
0.140496	-13.1623	with | similarity
0.117479	-13.1624	pattern | and
0.104213	-13.1627	a | of | types
0.333333	-13.1628	other | approaches
0.75	-13.1629	direct | complex
0.274194	-13.1629	consists of | set
0.15873	-13.1629	decoding | is
0.361111	-13.1631	performance | four
0.222222	-13.1631	in | messages
0.636364	-13.1631	small | sets.
0.636364	-13.1631	like | ones
0.0980392	-13.1631	the | takes
0.454545	-13.1632	expand | by
0.217822	-13.1632	can | directly
0.833333	-13.1632	distance | distances
0.833333	-13.1632	indicates | temporal
0.123028	-13.1632	responses | the
0.162162	-13.1633	above | for
0.225806	-13.1633	mention | entity
0.274194	-13.1633	an | dictionary
0.75	-13.1634	logical | logic.
0.146667	-13.1634	of | pairs,
0.571429	-13.1634	it | pronouns
0.189394	-13.1637	collected | a
0.833333	-13.1638	frequency | co-occurrence
0.25	-13.1639	learner | for
0.104784	-13.164	node | and
0.0916955	-13.164	one | or
0.230769	-13.164	the | items,
0.217822	-13.1641	project | for
0.44	-13.1641	utterances | single
0.12381	-13.1642	the | increasing
0.153465	-13.1643	model, | as
0.108911	-13.1643	and | up
0.454545	-13.1644	text-to-text | pages
0.7	-13.1644	0.65 | 0.90
0.7	-13.1644	shortness | breath
0.7	-13.1644	projected | treebank
0.636364	-13.1644	same | holds
0.636364	-13.1644	when | poor
0.181208	-13.1645	that | information.
0.235294	-13.1645	whether | sentence
0.198347	-13.1645	distance between | of
0.2	-13.1645	both | models
0.361111	-13.1646	is | granularity
0.102174	-13.1646	contain | of
0.47619	-13.1647	lu | al.
0.5	-13.1647	stream | streams
0.278689	-13.1648	our | baseline.
0.0936937	-13.1648	mentioned | the
0.4	-13.1649	point | but
0.0865533	-13.1649	the | via
0.833333	-13.1649	above | threshold,
0.833333	-13.1649	nouns, | adverbs.
0.306122	-13.165	to | probabilities.
0.571429	-13.165	mostly | but
0.137795	-13.1652	similarity | two
0.13834	-13.1652	it | very
0.18705	-13.1653	only | where
0.205357	-13.1653	of | derivations
0.25	-13.1654	in | sets,
0.0572391	-13.1654	the | proposed
0.235294	-13.1654	opinion | which
0.268657	-13.1654	in | acquisition
0.4	-13.1655	in | development,
0.326087	-13.1655	time | new
0.833333	-13.1655	rankers | set
0.833333	-13.1655	adopting | 2010).
0.351351	-13.1656	and | accordingly.
0.0715037	-13.1656	of | annotated
0.19084	-13.1656	explore | for
0.121951	-13.1656	the | method, | the
0.454545	-13.1657	each | vector.
0.0595065	-13.1661	the | selected
0.18	-13.1661	using | dependency
0.833333	-13.1661	classification | crfsgd
0.833333	-13.1661	d., | d.
0.134831	-13.1663	label | for
0.101952	-13.1664	for instance, | of
0.0991903	-13.1666	queries | the
0.833333	-13.1666	section 4 | discuss
0.243902	-13.1667	using | measure
0.155779	-13.1667	trained | on | and
0.130742	-13.1667	our | data.
0.319149	-13.1668	two | models.
0.130742	-13.1669	the | auxiliary
0.75	-13.167	teamingame | teamscoringall
0.636364	-13.167	rules | symbols
0.636364	-13.167	agree | gender
0.18705	-13.1671	give | for
0.25	-13.1671	year | and
0.0882825	-13.1671	presented | the | of
0.283333	-13.1671	tackle | and
0.833333	-13.1672	even | humans.
0.833333	-13.1672	seeds | seeds
0.126246	-13.1672	em | the
0.177632	-13.1673	published | the
0.0777917	-13.1674	with | number
0.238095	-13.1675	f, | the
0.529412	-13.1675	immediate | or
0.155	-13.1675	ratings | the
0.196721	-13.1676	first, | from
0.176471	-13.1676	going | the
0.47619	-13.1676	in | month
0.238095	-13.1677	established | and
0.127517	-13.1677	in | last
0.571429	-13.1677	since | avoid
0.278689	-13.1678	the | (buchholz
0.213592	-13.1678	of | production
0.274194	-13.1678	and | oov
0.5	-13.1678	in addition, | some
0.423077	-13.168	communication | an
0.253333	-13.1681	data-driven | and
0.351351	-13.1682	the | assignments.
0.0972763	-13.1682	described | a
0.268657	-13.1682	of | sufficiently
0.636364	-13.1683	moses | bleu
0.114441	-13.1685	complex | of
0.283333	-13.1685	when | multiple
0.407407	-13.1686	appear | both
0.3125	-13.1687	a | compiled
0.636364	-13.1687	the | observations:
0.636364	-13.1687	overall | what
0.135659	-13.1687	1.0 | 1.0
0.112272	-13.1687	there | an
0.296296	-13.1689	mert | a
0.107399	-13.1689	for | baseline
0.833333	-13.1689	listings | shoes
0.833333	-13.1689	spanish | innovation
0.75	-13.169	when | got
0.529412	-13.169	while | constraints
0.128472	-13.169	problem | as
0.3125	-13.1691	is | error.
0.361111	-13.1693	structural | are
0.19084	-13.1693	students | to
0.195122	-13.1693	thereby | the
0.2	-13.1693	the | predicates.
0.169697	-13.1693	also | performance
0.18705	-13.1694	description | is
0.217822	-13.1694	this | difference
0.122642	-13.1694	hierarchical | and
0.75	-13.1695	rare | corpus.
0.283333	-13.1696	the | modifies
0.192308	-13.1698	human | or
0.0955056	-13.1698	identify | and
0.5	-13.1698	(since | is
0.75	-13.17	dependencies | marneffe
0.283333	-13.17	a | researchers
0.529412	-13.1701	an | component.
0.166667	-13.1702	to | character
0.0763636	-13.1702	classification | of
0.228261	-13.1703	for | tuning
0.296296	-13.1703	have been | information
0.22	-13.1703	results | gold
0.44	-13.1704	theoretical | as
0.198347	-13.1705	association | word
0.47619	-13.1705	far | other
0.407407	-13.1705	further improve | of
0.833333	-13.1706	lexical | dictionaries.
0.278689	-13.1707	merged | a
0.361111	-13.1708	disambiguation | an
0.12381	-13.1708	a | relative
0.175325	-13.1708	therefore, | for
0.361111	-13.171	stacking | the
0.361111	-13.171	top | were
0.274194	-13.171	examine | on
0.111111	-13.171	one | it
0.0605602	-13.1711	training | for
0.0967118	-13.1711	computational linguistics | for
0.111688	-13.1712	argument | a
0.228261	-13.1713	c | v
0.361111	-13.1713	scores | significantly
0.0471612	-13.1713	this | not
0.209091	-13.1713	which | being
0.22	-13.1714	in | background
0.407407	-13.1714	by | lapata,
0.154229	-13.1715	path | a
0.253333	-13.1715	variable | that
0.1	-13.1716	method | that
0.0868878	-13.1716	count | the
0.833333	-13.1718	proteins | genes
0.454545	-13.1719	it | acceptable
0.371429	-13.1719	prediction | over
0.192308	-13.172	corpus | text
0.278689	-13.172	into | corresponding
0.0365752	-13.172	our | is | the
0.571429	-13.172	that | predictable
0.341463	-13.1721	utterances | more
0.454545	-13.1722	develop | models
0.129825	-13.1722	as | classification
0.301887	-13.1723	utility | a
0.145374	-13.1723	the | worse
0.0570087	-13.1724	however, | a
0.228261	-13.1724	the | l1
0.371429	-13.1724	model | trigram
0.361111	-13.1725	without | information
0.217822	-13.1725	approach | task
0.296296	-13.1725	pbsmt | the
0.106635	-13.1725	the | (for
0.225806	-13.1726	support | machine
0.100209	-13.1726	et | 2010;
0.221053	-13.1727	a | her
0.18797	-13.1729	to | usage
0.833333	-13.1729	analyze | characteristics
0.217822	-13.173	discover | of
0.238095	-13.173	distinct | for
0.205357	-13.173	set | extracted
0.375	-13.1731	a | interest.
0.5	-13.1732	methods | detection
0.15873	-13.1733	contribution | to
0.44	-13.1734	biomedical | june
0.44	-13.1734	these | designed
0.221053	-13.1735	with | i.e.
0.278689	-13.1735	all | translations
0.240964	-13.1735	the | candidate.
0.088141	-13.1736	of | entities
0.5	-13.1736	dataset, | which
0.361111	-13.1737	explains | in
0.529412	-13.1737	data | noise
0.149533	-13.1737	feature | it
0.290909	-13.1737	certainly | the
0.5	-13.1739	classifiers | features.
0.361111	-13.174	however, | higher
0.221053	-13.1741	of | compositionality
0.529412	-13.1741	svm | kernel
0.529412	-13.1741	resources | languages
0.833333	-13.1741	define | explicitly
0.319149	-13.1742	and | exhibit
0.0745413	-13.1742	assigned | the
0.201681	-13.1743	b, | the
0.44	-13.1743	bootstrapping | using
0.454545	-13.1744	baseline | unsupervised
0.454545	-13.1744	finally, | phrase
0.47619	-13.1744	a | heavy
0.230769	-13.1745	system | human
0.211538	-13.1745	in | creation
0.833333	-13.1746	t. | p.
0.75	-13.1747	especially | languages,
0.100629	-13.1747	construction | the
0.571429	-13.1747	europarl | parallel
0.529412	-13.1748	articles, | which
0.163842	-13.1751	the | toward
0.636364	-13.1753	norms, | and
0.4	-13.1753	data, | data,
0.269841	-13.1754	and | dictionaries
0.44	-13.1755	specify | which
0.25	-13.1755	we | arbitrary
0.423077	-13.1755	two | functions
0.454545	-13.1756	of | dictionaries,
0.285714	-13.1756	used | list
0.0672986	-13.1757	the | computed
0.3125	-13.1758	for | axioms
0.3125	-13.1758	unit | that
0.423077	-13.1758	important | quality
0.833333	-13.1758	described | soon
0.833333	-13.1758	interpretation | pronouns
0.257143	-13.1759	thus, | have
0.257143	-13.1759	as | typically
0.0742857	-13.1759	category | the
0.269841	-13.1761	query | ranking
0.19084	-13.1761	third | is
0.47619	-13.1761	similarity | weighted
0.387097	-13.1762	as | input.
0.529412	-13.1762	developed | only
0.264706	-13.1762	especially | with
0.185714	-13.1762	following | where
0.341463	-13.1763	(2010) | dependency
0.268657	-13.1763	a | w,
0.169697	-13.1763	of | edit
0.0527842	-13.1764	one | the | is
0.225806	-13.1766	by | prior
0.636364	-13.1766	(bender | al.,
0.19084	-13.1766	have | very
0.223404	-13.1767	is | relation.
0.0808219	-13.1767	the | languages.
0.26087	-13.1767	applied | other
0.12459	-13.1768	the | frequencies
0.351351	-13.1769	done | each
0.159574	-13.1769	4 | using
0.529412	-13.177	two | years
0.341463	-13.177	al | by
0.257143	-13.177	also | case
0.192308	-13.1772	get | for
0.134328	-13.1772	this | standard
0.189394	-13.1772	abstract | are
0.0502355	-13.1772	for | language
0.4	-13.1773	text | summarization
0.082153	-13.1775	propose | of
0.387097	-13.1775	load | in
0.341463	-13.1775	the | propagated
0.571429	-13.1775	grammar | analysis
0.264706	-13.1775	negation | a
0.108374	-13.1776	one | this
0.375	-13.1777	learn | models
0.0288528	-13.1777	the | were
0.4	-13.1778	is | name.
0.296296	-13.1778	(1 | is
0.125828	-13.1778	introduce | and
0.163842	-13.1779	module | a
0.192308	-13.1779	with | several
0.636364	-13.1779	3.3 | step
0.0679426	-13.1779	the | lower
0.325581	-13.178	because | available
0.3125	-13.1781	function | :
0.371429	-13.1781	(1), | a
0.155779	-13.1781	over | set
0.178808	-13.1782	experiments | use
0.75	-13.1783	two | common.
0.371429	-13.1783	of | cj
0.571429	-13.1783	(meyers | al.,
0.0963391	-13.1784	following | for
0.47619	-13.1784	and | effects.
0.47619	-13.1784	obtained | applying
0.325581	-13.1784	lemmatized | and
0.235294	-13.1786	to | projection
0.833333	-13.1786	ace | evaluation
0.198347	-13.1787	performance | over
0.0563287	-13.1788	they | in
0.529412	-13.1788	2009) | over
0.153465	-13.1789	in | process.
0.296296	-13.1789	to | keyphrases
0.407407	-13.1789	this | stream
0.833333	-13.1792	clusters | freebase
0.423077	-13.1793	achieves | f1
0.5	-13.1794	threshold | only
0.161111	-13.1795	takes | in
0.361111	-13.1795	list | extracted
0.333333	-13.1795	modifies | the
0.571429	-13.1795	2009), | 2008),
0.375	-13.1795	the | grows.
0.193548	-13.1796	we | approximate
0.134615	-13.1796	were | at
0.319149	-13.1797	only | content
0.75	-13.1798	component | build
0.75	-13.1798	lagrangian | problem
0.4	-13.1798	system | includes
0.833333	-13.1798	use | svmlight
0.215686	-13.1798	game | and
0.0621469	-13.18	new | and
0.351351	-13.18	triple | is
0.0775	-13.18	a | type
0.100418	-13.18	the | until
0.636364	-13.1801	using | ace
0.176471	-13.1801	(2008) | a | of
0.333333	-13.1801	hypothesis | using
0.0692076	-13.1803	patterns | the
0.75	-13.1804	(target | target
0.0822695	-13.1806	given | that
0.529412	-13.1806	into | computing
0.148837	-13.1806	we | nouns
0.217822	-13.1809	in | structural
0.127148	-13.1813	= | ,
0.306122	-13.1814	semantic | classes
0.75	-13.1814	expressed | expressed
0.75	-13.1814	use | ner.
0.636364	-13.1814	level | representation
0.4	-13.1814	where | known
0.571429	-13.1814	streams | single
0.229885	-13.1815	process | or
0.833333	-13.1815	exploiting | linguistic
0.5	-13.1818	are | perceived
0.221053	-13.182	we | variant
0.351351	-13.182	is | (from
0.186567	-13.1821	in proceedings of | conference
0.225806	-13.1821	candidates | are
0.269841	-13.1821	rules | but
0.833333	-13.1821	during | testing,
0.257143	-13.1821	clause | a
0.201754	-13.1824	the | adequacy
0.238095	-13.1825	corpus | contains
0.126246	-13.1826	of | whereas
0.571429	-13.183	a | white,
0.145374	-13.183	can | these
0.253521	-13.1832	issue, | the
0.151961	-13.1832	we found | in
0.833333	-13.1833	lack | space,
0.217822	-13.1834	• | have
0.232558	-13.1834	they | syntactic
0.117143	-13.1835	is | way
0.225806	-13.1836	1-best | the
0.529412	-13.1836	compared | approaches
0.4	-13.1836	corpus | 50
0.333333	-13.1836	that | messages
0.0672986	-13.1836	space | the
0.195122	-13.1836	parser | an
0.387097	-13.1837	three | single
0.407407	-13.1837	in | trie
0.217822	-13.1838	set | relations
0.169697	-13.1838	generated | by
0.0854962	-13.1839	a | class
0.75	-13.184	metrics | content
0.232558	-13.184	on | show that
0.47619	-13.184	in | msd
0.341463	-13.1841	but | several
0.19084	-13.1841	of | wi
0.15873	-13.1841	in | highly
0.5	-13.1843	in | scene,
0.101512	-13.1844	further | in
0.10514	-13.1845	the | tagged
0.4	-13.1845	topic | multinomial
0.375	-13.1845	3.2.1 | of
0.285714	-13.1847	of | sciences,
0.209524	-13.1847	of | earlier
0.375	-13.1847	indeed, | are
0.144737	-13.1847	terms | are
0.130909	-13.1848	dataset | a
0.319149	-13.1848	information | institute
0.0573997	-13.1848	the | shared
0.407407	-13.1848	might | better
0.215686	-13.1849	a | simplified
0.75	-13.185	tasks | entailment
0.319149	-13.185	semantic | those
0.5	-13.185	machine | performed
0.3125	-13.1851	this | mainly
0.196721	-13.1852	vs. | for
0.235294	-13.1852	= | denotes
0.264706	-13.1853	a | tokens,
0.636364	-13.1854	if | someone
0.076087	-13.1855	in section | we
0.319149	-13.1855	and | substantially
0.0845921	-13.1855	with | but
0.351351	-13.1856	thread | a
0.143478	-13.1856	of | especially
0.75	-13.1856	converted | representation
0.15544	-13.1856	attribute | of
0.144737	-13.1857	and | e.g.
0.15873	-13.1857	variation | of
0.235294	-13.1858	this, | that
0.177632	-13.1858	statements | the
0.319149	-13.1859	by | 2004).
0.454545	-13.186	xinhua | of
0.296296	-13.186	although | work
0.0754491	-13.186	classification | and
0.375	-13.186	clustering | different
0.533333	-13.1861	ontology | domain
0.533333	-13.1861	representations | 2008;
0.205357	-13.1862	probability that | is
0.178082	-13.1863	spelling | and
0.26087	-13.1864	model | knowledge
0.198347	-13.1864	come | and
0.122977	-13.1865	3.1 | in
0.3125	-13.1866	version | was
0.75	-13.1866	multiword | (mwes)
0.75	-13.1866	relation | direct
0.2	-13.1866	features | parse
0.75	-13.1871	formulated | problem
0.75	-13.1871	despite | lack
0.192	-13.1873	and | t.
0.21875	-13.1874	and | punctuation
0.207207	-13.1874	with | prior
0.342105	-13.1874	is | user.
0.75	-13.1876	method | uniform
0.75	-13.1876	our | though,
0.636364	-13.1876	question, | an
0.636364	-13.1876	queries | similar
0.0825036	-13.1876	the | of | sequence
0.47619	-13.1876	the | sites.
0.47619	-13.1876	on | mentions.
0.192	-13.1877	(we | the
0.24359	-13.1877	in | chain
0.225806	-13.1877	(2009) | by
0.167665	-13.1877	community | and
0.160221	-13.1877	of | errors.
0.407407	-13.1877	linear | algorithm
0.306122	-13.1878	the | learned.
0.423077	-13.1878	with | c.
0.361111	-13.188	some | appear
0.533333	-13.188	send | to
0.274194	-13.188	a | pattern.
0.223404	-13.1881	has | several
0.264706	-13.1881	30 | 30
0.533333	-13.1883	agree | disagree
0.636364	-13.1885	proposed | method,
0.0745921	-13.1885	side | the
0.205357	-13.1886	consists | which
0.24359	-13.1887	also | systems
0.423077	-13.1887	weighting | document
0.423077	-13.1887	an | interest
0.407407	-13.1888	use | limited
0.211538	-13.1888	again | and
0.121495	-13.1889	and | agreement
0.636364	-13.1889	optimal | search
0.296296	-13.1889	prior | has
0.215686	-13.1889	simple | can
0.128472	-13.189	2.1 | and
0.110256	-13.1891	in | discourse
0.5	-13.1891	some | express
0.213592	-13.1892	p | p r
0.161111	-13.1893	a | z
0.175325	-13.1893	interested in | of
0.529412	-13.1894	and | 2003.
0.215686	-13.1895	expansion | of
0.833333	-13.1896	the::d | what::w
0.833333	-13.1896	chinese | segmentation | machine
0.20354	-13.1897	that | share
0.3125	-13.1901	acquisition | is
0.130435	-13.1901	4.2 | in
0.104895	-13.1901	a | combination
0.253521	-13.1901	can | learn
0.325581	-13.1901	non-terminal | in
0.164706	-13.1902	number | total
0.423077	-13.1902	translated | language.
0.175325	-13.1902	selection | that
0.833333	-13.1902	pθ(f,a|e)
0.128028	-13.1903	= | (2)
0.361111	-13.1905	the | lms.
0.434783	-13.1905	wrong | wrong
0.5	-13.1905	hypothesis | tend
0.5	-13.1905	considers | sentence
0.278689	-13.1907	query | all
0.223404	-13.1908	1 | 0.5
0.180556	-13.1908	acknowledgments this | supported
0.113514	-13.1908	a | web
0.407407	-13.1908	on | right.
0.154639	-13.191	3.3 | in
0.0955166	-13.1911	in | output
0.342105	-13.1912	verify | in
0.375	-13.1913	0 | 180
0.119632	-13.1913	interesting | of
0.0729517	-13.1914	of | extracted
0.833333	-13.1914	analysis | (deerwester
0.833333	-13.1914	wikipedia | dbpedia
0.246753	-13.1915	incremental | in
0.167665	-13.1915	we | coverage
0.146119	-13.1915	table 6 | the
0.174194	-13.1917	issues | and
0.193548	-13.1917	task | one
0.571429	-13.1918	how | efficiently
0.0617482	-13.1919	additional | the
0.136187	-13.1919	in | users
0.152709	-13.192	tree, | the
0.833333	-13.192	in: | text
0.636364	-13.1921	test | hypothesis that
0.533333	-13.1921	β | beam
0.333333	-13.1922	human | given
0.571429	-13.1922	quality | learned
0.342105	-13.1922	these | occur
0.407407	-13.1922	properties | their
0.454545	-13.1923	al., | smith,
0.361111	-13.1923	and | ontologies
0.170886	-13.1924	moreover, | is
0.533333	-13.1924	feature | component
0.333333	-13.1924	if | change
0.333333	-13.1924	system | overall
0.181818	-13.1924	approach, | is
0.375	-13.1926	were | before
0.375	-13.1926	regression | which
0.0987654	-13.1927	constructed | the
0.175325	-13.1927	this | including
0.290909	-13.1927	have | comparable
0.174194	-13.1928	is | it.
0.75	-13.1929	obtaining | relevant
0.25	-13.1929	against | by
0.454545	-13.193	present | those
0.0604651	-13.193	an | from
0.280702	-13.1931	scores | both
0.246753	-13.1931	and | composition
0.157895	-13.1931	is | as follows.
0.361111	-13.1933	approaches | document
0.0979592	-13.1933	3 | with
0.306122	-13.1934	learn | data
0.75	-13.1934	same | reducing
0.0676983	-13.1934	for | semantic
0.454545	-13.1936	al., | denero
0.3	-13.1936	effect | features
0.232558	-13.1936	clearly | that
0.833333	-13.1937	simple, | effective
0.125424	-13.1938	to | evaluate
0.0840841	-13.1938	reported | the
0.407407	-13.1939	blast | and
0.110256	-13.194	metric | and
0.185185	-13.1943	to | scope
0.423077	-13.1943	and | symbols.
0.533333	-13.1943	used | feature.
0.342105	-13.1943	with | methods,
0.3125	-13.1944	our | robust
0.15544	-13.1945	with | length
0.213592	-13.1947	a | 0.
0.253521	-13.1947	that | expert
0.113889	-13.1947	support | and
0.5	-13.1947	though | many
0.149038	-13.1948	reliability | the
0.265625	-13.1949	a | defining
0.75	-13.195	table | confirms
0.434783	-13.195	although | often
0.333333	-13.195	of | times,
0.173077	-13.195	not | but | as
0.2	-13.195	no | at
0.0578279	-13.1951	of the | pages
0.195122	-13.1951	that | optimal
0.0840841	-13.1952	each | set
0.132576	-13.1953	is | during
0.174194	-13.1954	is | towards
0.0646018	-13.1954	which | from
0.318182	-13.1954	a | prominent
0.229885	-13.1955	base | on
0.15847	-13.1955	the | output,
0.423077	-13.1955	errors | those
0.833333	-13.1955	sentiment, | sentiment
0.24359	-13.1959	features | candidate
0.101996	-13.1959	an | has
0.3125	-13.196	related | similar
0.178082	-13.196	work | under
0.3	-13.1961	each | one or
0.387097	-13.1961	of | lemmas,
0.636364	-13.1961	thus | phrases
0.290909	-13.1961	scheme, | and
0.174194	-13.1962	larger | for
0.136	-13.1962	this | specific
0.318182	-13.1962	is | satisfy
0.290909	-13.1963	0.6 | 0.9
0.454545	-13.1965	claims | or
0.213592	-13.1965	the | low.
0.571429	-13.1966	head | identi
0.833333	-13.1966	because | shorter
0.833333	-13.1966	a. | f.
0.0776081	-13.1968	are | other
0.227273	-13.1968	from | users
0.290909	-13.1968	first | selected
0.533333	-13.1969	that | simultaneously.
0.636364	-13.197	(foster | al.,
0.265625	-13.1971	simple | only
0.182482	-13.1971	the | implies
0.101099	-13.1972	up | and
0.285714	-13.1972	sample | each
0.3	-13.1972	patterns | words
0.387097	-13.1972	in | unification
0.232558	-13.1972	created | from
0.833333	-13.1972	draw | conclusions.
0.104167	-13.1973	the | classified
0.423077	-13.1973	feature: | and
0.351351	-13.1974	information | without
0.102222	-13.1974	full | and
0.186567	-13.1977	results, | in
0.209524	-13.1977	the | ones,
0.0989474	-13.1978	describe | in
0.278689	-13.1979	by | likelihood
0.223404	-13.1981	assume that | that
0.75	-13.1981	annotators | annotators.
0.434783	-13.1981	rule | x1
0.434783	-13.1981	algorithm | line
0.108586	-13.1981	of | component
0.375	-13.1981	specifically, | word
0.232558	-13.1982	our | reported
0.265625	-13.1983	methods | both
0.0793011	-13.1983	for | text
0.407407	-13.1985	a | spectrum
0.211538	-13.1986	unlike | is
0.351351	-13.1987	developed | system
0.375	-13.1989	only | classes
0.211538	-13.1989	to | constraints.
0.118902	-13.199	frequent | in
0.265625	-13.199	exploited | and
0.833333	-13.199	2010a; | 2010).
0.833333	-13.199	employed | relation
0.423077	-13.1991	evaluation | done
0.342105	-13.1991	task | single
0.325581	-13.1993	5th | on
0.454545	-13.1994	the | segmenters
0.318182	-13.1994	accuracy | without
0.3	-13.1995	higher | scores
0.114846	-13.1996	of | constraints
0.246753	-13.1997	the | ranking | ranking
0.280702	-13.1998	support | by
0.290909	-13.1998	results | correct
0.434783	-13.1999	model | lattice
0.473684	-13.1999	improve | models.
0.153061	-13.2	utterances | in
0.387097	-13.2002	applied | sentences
0.833333	-13.2002	capture | phenomena.
0.833333	-13.2002	stimuli | linguistic
0.5	-13.2004	context | entities
0.265625	-13.2005	model | variables
0.473684	-13.2006	formality | that
0.047619	-13.2006	the | input
0.121795	-13.2007	of | usa
0.0352879	-13.2007	at | of
0.269841	-13.2008	with | f-score
0.240506	-13.2009	of | to reduce
0.26087	-13.2009	token | word
0.162011	-13.201	for | relative
0.571429	-13.201	was | prosodic
0.351351	-13.2011	guess | in
0.5	-13.2011	however, | easily
0.216495	-13.2012	complex | with
0.240506	-13.2012	the | composing
0.0343703	-13.2013	this | as
0.269841	-13.2013	extremely | in
0.434783	-13.2014	quality | annotations
0.178082	-13.2014	a | allowing
0.257143	-13.2014	pair | one
0.423077	-13.2015	correct | 3
0.265625	-13.2016	iteration | for
0.136546	-13.2016	given | where
0.257143	-13.2017	for | variants
0.571429	-13.2018	resource | languages
0.144144	-13.2018	alignment | word
0.454545	-13.2019	improve | improve
0.342105	-13.202	many | so
0.361111	-13.2022	i | about
0.363636	-13.2023	by | techniques.
0.175676	-13.2023	comparative | of
0.0822511	-13.2025	error | and
0.5	-13.2025	smt | translate
0.235294	-13.2026	a | hierarchy
0.75	-13.2029	small | current
0.195122	-13.2029	is | process.
0.351351	-13.2031	table 2 shows | on
0.232558	-13.2031	it | hard
0.833333	-13.2031	sparsifying | covariance
0.833333	-13.2031	explicit | reference
0.333333	-13.2032	research | models
0.149758	-13.2032	of | (and
0.318182	-13.2032	selected | were
0.325581	-13.2032	with | reduced
0.10804	-13.2033	words, | in
0.101996	-13.2033	phrases | in
0.75	-13.2034	for | system).
0.173077	-13.2036	to | errors.
0.166667	-13.2037	by | terms
0.533333	-13.2037	ck | size
0.333333	-13.2037	nevertheless, | that
0.290909	-13.2037	our | approaches.
0.2	-13.2038	know | is
0.423077	-13.2039	assignment | word
0.387097	-13.204	the | software.
0.285714	-13.2041	affective | of
0.434783	-13.2042	assuming | all
0.103926	-13.2043	to | good
0.25	-13.2043	4.3 | is
0.833333	-13.2043	enhancing | model
0.833333	-13.2043	statistics | n-grams
0.833333	-13.2043	journal | informatics
0.144796	-13.2043	system | can
0.407407	-13.2043	opinions | their
0.205607	-13.2044	if | rule
0.129496	-13.2045	compute | for
0.0260794	-13.2046	the | is | as
0.0596787	-13.2047	annotated | the
0.15544	-13.2047	path | to
0.15122	-13.2047	introduce | in
0.25	-13.2048	this | benefit
0.246753	-13.2049	previously | as
0.274194	-13.2049	8 | 12
0.833333	-13.2049	minimize | sum
0.306122	-13.205	of | characters,
0.75	-13.205	n-grams | tokens
0.363636	-13.2052	that | 80%
0.280702	-13.2053	of | balanced
0.0805556	-13.2053	that | given
0.5	-13.2053	use | simplification
0.363636	-13.2054	we | which,
0.571429	-13.2054	software | available
0.454545	-13.2055	and | kernels.
0.166667	-13.2055	to | summaries
0.833333	-13.2055	functions | roles.
0.112948	-13.2055	learning | with
0.227273	-13.2057	with | parameters
0.571429	-13.2058	segment | other
0.3	-13.2059	obtained | best
0.833333	-13.206	return | result
0.5	-13.206	base | model.
0.351351	-13.2061	space | only
0.75	-13.2061	output | word,
0.198276	-13.2061	defined | which
0.11326	-13.2062	to | identify
0.342105	-13.2062	services | and
0.636364	-13.2064	using | desired
0.134387	-13.2064	the | system’s
0.224719	-13.2065	5.1 | is
0.174497	-13.2066	classified | of
0.571429	-13.2067	to | company.
0.205607	-13.2068	fragments | a
0.26087	-13.2068	0.9 | of
0.092081	-13.207	and | output
0.227273	-13.207	where | is | length
0.318182	-13.207	exploits | and
0.75	-13.2071	variables | constraints.
0.75	-13.2071	systems | little
0.0559676	-13.2072	then | a
0.363636	-13.2072	random | probability
0.363636	-13.2072	we | counts.
0.25	-13.2073	is | area
0.407407	-13.2074	system | systems,
0.151515	-13.2075	provided | for
0.274194	-13.2075	removing | a
0.0877483	-13.2076	conclusion | the
0.0159987	-13.2076	is | in
0.285714	-13.2077	a | program
0.0840979	-13.2077	the | known
0.318182	-13.2077	actually | by
0.454545	-13.208	system | select
0.3	-13.208	algorithms | which
0.3	-13.208	the | pos-tag
0.325581	-13.208	when | languages
0.351351	-13.2081	in | phase.
0.5	-13.2081	integrated | models
0.24359	-13.2082	a | similarity.
0.2	-13.2082	analysis | analysis | the
0.0839695	-13.2083	our | use
0.0432692	-13.2083	a | where
0.153061	-13.2083	labeled | as
0.112948	-13.2083	from | where
0.0818966	-13.2084	the | sets.
0.285714	-13.2085	have | thus
0.144144	-13.2085	start | and
0.269841	-13.2086	modifications | and
0.17931	-13.2086	of | medical
0.75	-13.2087	collection | through
0.178082	-13.2087	of | internal
0.113889	-13.2087	which | other
0.0514124	-13.2088	• | of
0.164706	-13.2088	given | source
0.280702	-13.2089	we now | that
0.0231545	-13.2089	the | model | the
0.143498	-13.2089	this | general
0.454545	-13.209	hypothesis | probability
0.423077	-13.209	finite-state | in
0.473684	-13.209	< | $.
0.216495	-13.2091	to | decoding.
0.269841	-13.2092	different | methods
0.181159	-13.2093	the | another.
0.318182	-13.2093	make | task
0.1875	-13.2094	9 | for
0.205607	-13.2094	the | descriptions.
0.21875	-13.2094	a | (which
0.0704376	-13.2094	the | added | the
0.423077	-13.2096	5.4 | with
0.12766	-13.2097	weight | is
0.196581	-13.2098	of | decoder
0.131086	-13.2098	web | a
0.375	-13.21	problems | our
0.170886	-13.2101	strength | the
0.533333	-13.2101	ppt | figure
0.306122	-13.2102	perhaps | is
0.136	-13.2102	• | or
0.833333	-13.2102	different | densities.
0.833333	-13.2102	feature-based | features
0.193548	-13.2102	sometimes | to
0.318182	-13.2102	of | informed
0.105263	-13.2103	the | constructed
0.0944123	-13.2104	analysis | that
0.234568	-13.2104	used | parse
0.135458	-13.2105	the | settings
0.269841	-13.2105	number | correctly
0.342105	-13.2105	national | no.
0.3	-13.2106	and | h.
0.05	-13.2106	that | have
0.407407	-13.2106	on | topic,
0.0967742	-13.2108	2. | a
0.1875	-13.2109	furthermore, | as
0.294118	-13.2111	consequence, | the
0.3	-13.2112	authors | necessarily reflect
0.16185	-13.2112	is | to | current
0.125	-13.2112	that | significant
0.318182	-13.2113	further | these
0.0888889	-13.2114	alignment | in
0.325581	-13.2115	the | “as
0.0238707	-13.2116	is | as
0.5	-13.2117	lexical | grammatical
0.176871	-13.2118	the | reflects
0.211538	-13.2118	for | turn
0.454545	-13.2119	assume | null
0.434783	-13.2119	with | seven
0.149038	-13.2122	of | tagging
0.1875	-13.2122	are | text.
0.285714	-13.2124	system | knowledge
0.186567	-13.2124	context | words
0.131579	-13.2124	investigate | and
0.205607	-13.2125	for | comments
0.153061	-13.2125	abstract | from
0.234568	-13.2126	ri | the
0.636364	-13.2127	most | online
0.182482	-13.2127	observed | as
0.169811	-13.2128	required | a
0.240506	-13.2128	the | leave
0.75	-13.213	method, | method.
0.0522607	-13.213	for | set
0.473684	-13.213	possible | resulting
0.290909	-13.213	consistent with | that
0.125424	-13.2131	return | the
0.392857	-13.2131	for | percent
0.571429	-13.2132	paraphrases | sentences
0.833333	-13.2132	improve | setting.
0.833333	-13.2132	avoids | problem
0.21875	-13.2132	were | about
0.274194	-13.2133	to | published
0.0669253	-13.2135	the | provide
0.434783	-13.2135	merge | are
0.21875	-13.2135	size | set
0.407407	-13.2135	of | sampling.
0.5	-13.2135	normalized | which
0.105263	-13.2136	is | being
0.583333	-13.2136	comparable | slightly
0.583333	-13.2136	types | 1).
0.583333	-13.2136	48th | the
0.209524	-13.2136	system | three
0.666667	-13.2137	there | ample
0.666667	-13.2137	especially | linguistic
0.666667	-13.2137	67 | 67
0.666667	-13.2137	mc | mc
0.232558	-13.2137	the | environment.
0.833333	-13.2138	main | b)
0.325581	-13.2138	forms | word
0.240506	-13.214	average | between
0.190476	-13.214	distortion | the
0.75	-13.2141	will | others.
0.636364	-13.2141	negation | other
0.0763454	-13.2141	that | over
0.234568	-13.2141	as | root
0.318182	-13.2141	for | items.
0.157609	-13.2141	experiments | section
0.666667	-13.2142	arabic | dialectal
0.173333	-13.2142	from | small
0.196581	-13.2144	steps | a
0.473684	-13.2144	unary | are
0.166667	-13.2145	therefore | as
0.229885	-13.2145	various | such
0.285714	-13.2146	with | people
0.666667	-13.2146	got | score
0.387097	-13.2148	artificial | in
0.194915	-13.2148	every | by
0.151515	-13.2149	the | penalty
0.583333	-13.2149	graph-based | scoring
0.209524	-13.2151	for | twitter
0.75	-13.2152	from | mt08
0.75	-13.2152	instance | techniques
0.75	-13.2152	components | systems,
0.571429	-13.2152	decipherment | using
0.583333	-13.2153	english | swedish
0.115607	-13.2154	all | our
0.209524	-13.2154	requires | that
0.285714	-13.2155	we have | how
0.124161	-13.2155	for | we have
0.138075	-13.2156	active | and
0.666667	-13.2156	nodes | coalesced.
0.333333	-13.2156	event | arguments.
0.473684	-13.2157	also | beyond
0.265625	-13.2158	for | creation
0.178082	-13.2158	a | direction
0.636364	-13.2159	grammar | derivations
0.123746	-13.2159	using | translation
0.342105	-13.2159	only | argument
0.130597	-13.2159	our experiments | of
0.407407	-13.2159	different | annotations
0.325581	-13.2159	nlp | language
0.164706	-13.216	it | out
0.434783	-13.216	section | conclusions.
0.571429	-13.216	more | type.
0.109948	-13.2161	a | entities
0.280702	-13.2161	contributed | the
0.0824742	-13.2161	2006; | et
0.473684	-13.2161	train | statistical
0.2	-13.2161	trigram | and
0.75	-13.2162	local | patterns
0.142222	-13.2163	scoring | of
0.149758	-13.2164	testing | and
0.0377069	-13.2165	the | phrase
0.100877	-13.2166	to | task.
0.666667	-13.2166	robust | noise
0.666667	-13.2166	role-specific | role
0.16185	-13.2166	2002), | and
0.285714	-13.2169	main | use
0.246575	-13.2169	approach | extraction
0.224719	-13.2169	creating | in
0.325581	-13.2169	function | ∈
0.154639	-13.2171	we | them.
0.0658514	-13.2172	to | task
0.0882852	-13.2173	development | and
0.833333	-13.2173	similar | are,
0.181159	-13.2174	discriminative | a
0.666667	-13.2175	simple | trigger
0.666667	-13.2175	mckeown, | callison-burch,
0.351351	-13.2176	were | five
0.205607	-13.2176	a | graphical
0.571429	-13.2177	terms | accuracy.
0.571429	-13.2177	method | measuring
0.0837139	-13.2178	as | feature
0.636364	-13.2178	they | cover
0.533333	-13.2178	added | lexicon
0.128571	-13.2179	expect | to
0.833333	-13.2179	simultaneously | different
0.407407	-13.2179	improve | better
0.222222	-13.218	a | ccg
0.0645455	-13.2181	with | word
0.114943	-13.2181	4 | and | in
0.5	-13.2181	entailment | lexical
0.154639	-13.2182	the | and/or
0.186047	-13.2183	of | chunk
0.224719	-13.2183	method | but
0.181159	-13.2184	for | queries
0.342105	-13.2184	internal | are
0.16875	-13.2184	the | w.
0.306122	-13.2185	of | ambiguities
0.181159	-13.2185	a | efficient
0.185185	-13.2185	approach | information
0.434783	-13.2185	for | dependents
0.833333	-13.2185	some | possibility
0.833333	-13.2185	induce | grammar
0.375	-13.2186	simple | results
0.636364	-13.2187	a. | t.
0.201754	-13.2187	the | entity,
0.311111	-13.2187	that | shorter
0.392857	-13.2187	use | contextual
0.454545	-13.2188	then | 0.
0.114613	-13.2189	where | an
0.375	-13.2189	only | four
0.5	-13.2189	of | trivial.
0.179856	-13.2191	with | lower
0.257143	-13.2191	in | speed
0.246575	-13.2193	from | observed
0.212121	-13.2193	approaches, | and
0.205607	-13.2195	in | light
0.166667	-13.2196	from | so
0.142222	-13.2196	in | e.g.,
0.121795	-13.2196	for | 10
0.15847	-13.2197	also | information
0.434783	-13.2197	in | discarded.
0.434783	-13.2197	using | information,
0.159341	-13.2197	constituent | of
0.175676	-13.2197	modeled | the
0.571429	-13.2197	their | identifying
0.571429	-13.2197	application, | for
0.833333	-13.2197	estimate | real
0.407407	-13.2197	approaches | context
0.135458	-13.2198	provided | a
0.666667	-13.2199	close | space
0.181159	-13.22	a | his
0.75	-13.22	likely | increase
0.454545	-13.2201	held-out | set
0.306122	-13.2201	usefulness | and
0.533333	-13.2201	every | length
0.533333	-13.2201	from | commercial
0.571429	-13.2201	by | individuals
0.203704	-13.2202	effort | to
0.583333	-13.2203	4.5 | from
0.666667	-13.2204	referent | referent
0.392857	-13.2204	segmentation | morphological
0.0305668	-13.2204	as | for
0.123333	-13.2206	in | take
0.125874	-13.2206	events | a
0.16875	-13.2206	employ | to
0.306122	-13.2207	labels, | a
0.125436	-13.2208	versions | the
0.185185	-13.2208	work | grant
0.224719	-13.2208	gives | on
0.0997831	-13.2209	in | gold
0.833333	-13.2209	reasons | this.
0.833333	-13.2209	tweet | tweets.
0.311111	-13.221	the | computation.
0.139831	-13.2211	with | similar
0.407407	-13.2211	set | types.
0.294118	-13.2212	has | (see
0.666667	-13.2213	documents | media
0.257576	-13.2213	labels, | of
0.280702	-13.2214	set | articles
0.392857	-13.2215	leverage | for
0.109948	-13.2216	of | performed
0.166667	-13.2216	year | the
0.2	-13.2216	of | parsed
0.214286	-13.2216	the | scheme,
0.0705755	-13.2217	above | the
0.227273	-13.2217	split | training
0.666667	-13.2218	patern | annotation
0.25	-13.2218	although | such
0.325581	-13.2218	the | said.
0.636364	-13.2219	match | original
0.0791269	-13.222	in | single
0.0798319	-13.222	dictionary | the
0.833333	-13.2221	did | ?
0.25	-13.2223	simply | from
0.198276	-13.2223	extraction, | and
0.143498	-13.2223	is | matrix
0.129151	-13.2224	3.2 | a
0.133858	-13.2224	that | labeled
0.306122	-13.2225	category | word
0.179856	-13.2225	are | in | 5.
0.0698413	-13.2225	present | and
0.201835	-13.2226	in | typical
0.571429	-13.2226	authors | texts
0.407407	-13.2226	0.04 | 0.08
0.454545	-13.2227	bicycles | are
0.285714	-13.2227	figure | presents
0.0824742	-13.2227	2. | and
0.183824	-13.2227	they | information
0.120635	-13.2227	different | or
0.833333	-13.2227	without | generality,
0.666667	-13.2228	less | ones.
0.229885	-13.2228	constant | of
0.342105	-13.2228	however, | few
0.285714	-13.2229	transcripts | of
0.294118	-13.2229	to | unlikely
0.214286	-13.2229	with | increase
0.306122	-13.223	to | words)
0.0958084	-13.223	label | a
0.333333	-13.223	the | trigger.
0.342105	-13.223	the | dependency trees.
0.285714	-13.2231	all | models.
0.243243	-13.2231	corpus | size
0.096	-13.2231	among | and
0.203704	-13.2232	evaluating | in
0.583333	-13.2232	resources | available.
0.392857	-13.2232	using | 2011).
0.16185	-13.2233	tool | to
0.15	-13.2234	turn | to
0.325581	-13.2234	clusters | their
0.311111	-13.2235	features | results.
0.666667	-13.2237	dots | totally
0.181159	-13.2237	use | method
0.583333	-13.2237	is | limited.
0.318182	-13.2237	the | 8).
0.246575	-13.2238	of | eight
0.571429	-13.2238	of | straight
0.407407	-13.2238	parameter | development
0.227273	-13.2241	number | extracted
0.392857	-13.2241	sentence | assigned
0.533333	-13.2243	kernels | kernel
0.167702	-13.2244	is | clustering
0.124161	-13.2244	that | contain
0.152284	-13.2244	using | similarity
0.392857	-13.2244	system | pairwise
0.333333	-13.2244	explore | features
0.16185	-13.2245	thus, | as
0.257576	-13.2245	from | gigaword
0.198276	-13.2246	two | similar
0.131086	-13.2246	for | simple
0.2	-13.2246	is | parent
0.126761	-13.2247	improvements | and
0.285714	-13.2247	a | point.
0.351351	-13.2247	finally, | into
0.375	-13.2248	denotes | on
0.583333	-13.2249	framework | specific
0.16875	-13.2249	a | building
0.0704225	-13.2249	are | same
0.583333	-13.2253	building | building
0.75	-13.2254	text | tokens,
0.533333	-13.2254	algorithms | compute
0.234568	-13.2255	precision | recall | recall
0.306122	-13.2256	do | make
0.0363531	-13.2256	the | models
0.21	-13.2256	systems | have
0.275862	-13.2257	set | trees
0.0667976	-13.2257	sequence | the
0.833333	-13.2257	13 | 33
0.271186	-13.2257	then | rules
0.269841	-13.2258	both | syntactic
0.342105	-13.2258	an | null,
0.214286	-13.2258	speaker | in
0.198276	-13.2259	initial | by
0.198276	-13.2259	90% | the
0.231707	-13.226	capture | between
0.636364	-13.226	become | become
0.342105	-13.226	n | nodes
0.131579	-13.226	of | span
0.306122	-13.2263	school | computer
0.222222	-13.2263	examples | each
0.163743	-13.2264	language, | and
0.122112	-13.2264	recognition | of
0.318182	-13.2264	responses | each
0.407407	-13.2264	the | 1989).
0.75	-13.2265	transition-based | zhang
0.533333	-13.2266	that | compositionality.
0.0747549	-13.2266	similar | to
0.231707	-13.2267	are | concepts
0.16092	-13.2268	we | unigram
0.169811	-13.2268	wikipedia | for
0.111717	-13.2268	collected | the
0.240506	-13.227	occurring | a
0.141593	-13.2271	we | before
0.318182	-13.2271	the | fillers.
0.143498	-13.2271	1 | .
0.257576	-13.2271	the | modifying
0.179856	-13.2272	cost | a
0.5	-13.2272	corpora | linguistic
0.229885	-13.2273	full | model
0.0385113	-13.2274	each | is
0.583333	-13.2275	not | available,
0.240506	-13.2275	system | uses | and
0.833333	-13.2275	idss | sense
0.25	-13.2276	fragment | in
0.363636	-13.2276	of | frames.
0.571429	-13.2276	i, | j
0.571429	-13.2276	measure | relative
0.231707	-13.2277	6 conclusions | the
0.311111	-13.2277	we | record
0.0849057	-13.2277	the | elements
0.375	-13.2278	an | necessarily
0.115727	-13.2279	the | trees.
0.306122	-13.228	our | entire
0.75	-13.2281	find | granularity
0.833333	-13.2282	therefore | choice
0.0700431	-13.2282	how | a
0.0737218	-13.2284	then | for
0.257576	-13.2284	next | our
0.285714	-13.2286	joint | over
0.119497	-13.2287	such | have
0.75	-13.2287	while | strictly
0.159091	-13.2288	3 | used
0.0706522	-13.2288	relevant | the
0.162791	-13.2288	recently, | and
0.833333	-13.2288	decoder | current
0.533333	-13.2289	all | largest
0.533333	-13.2289	improvement | additional
0.288462	-13.229	way | using
0.2375	-13.2291	wrong | in
0.155914	-13.2294	expression | a
0.288462	-13.2295	9 | 7
0.454545	-13.2296	as | incorporating
0.182482	-13.2296	we | observe
0.253521	-13.2296	for | minimal
0.2375	-13.2297	the | evidences
0.363636	-13.2297	correct, | in
0.407407	-13.2297	differing | of
0.106173	-13.2298	this | since
0.138075	-13.2299	number | sentences
0.454545	-13.2299	both | contain
0.131274	-13.2299	describes | of
0.25	-13.2299	to | detection.
0.392857	-13.23	for | side,
0.060778	-13.23	the | thus
0.231707	-13.2301	for | six
0.434783	-13.2301	approach | results.
0.261538	-13.2302	instance | it
0.246575	-13.2302	in | selection.
0.0341662	-13.2302	is | which
0.280702	-13.2303	from | 2010
0.375	-13.2303	approach | directly
0.193277	-13.2303	each | form
0.352941	-13.2304	the | j-th
0.253521	-13.2304	coverage | by
0.375	-13.2305	flexible | for
0.294118	-13.2306	standard | use
0.636364	-13.2306	these | giza++
0.833333	-13.2306	word | partof-speech
0.157609	-13.2306	in | separate
0.275862	-13.2309	the | synthesized
0.75	-13.2309	information about | distribution
0.75	-13.2309	crucially, | that
0.571429	-13.2309	chooses | most
0.214286	-13.2309	where | are | from
0.533333	-13.2312	parameters | learning
0.280702	-13.2313	a | covers
0.265625	-13.2313	first, | word
0.363636	-13.2313	collocations | in
0.15508	-13.2313	other | used
0.140969	-13.2313	summary | in
0.75	-13.2314	generalize | unseen
0.085622	-13.2314	multiple | of
0.21978	-13.2314	1 | 1.
0.174497	-13.2314	to | incorrect
0.0885417	-13.2315	a | being
0.363636	-13.2315	for | alone
0.392857	-13.2315	establishing | and
0.102564	-13.2316	when | not
0.122112	-13.2316	is | in | context
0.375	-13.2316	of | use,
0.271186	-13.2316	together | their
0.0518257	-13.2317	when | is | the
0.583333	-13.2317	within | including
0.14554	-13.2317	3 | 7
0.325581	-13.2317	the | numbers.
0.163636	-13.2318	depth | of
0.833333	-13.2318	fairly | given
0.275862	-13.2319	paraphrases | that
0.143519	-13.232	features | these
0.533333	-13.232	position | sentence.
0.16875	-13.232	explicitly | and
0.583333	-13.2321	does not | score
0.16092	-13.2324	s | .
0.666667	-13.2324	œ | k.
0.833333	-13.2324	done | steps.
0.318182	-13.2324	and | variance
0.318182	-13.2324	operators | for
0.275862	-13.2325	it | quite
0.636364	-13.2325	none, | and
0.571429	-13.2325	by | costs
0.571429	-13.2325	82 | number
0.257576	-13.2325	amounts | a
0.196581	-13.2326	binary | that
0.363636	-13.2326	function | scores
0.0868114	-13.2326	automatically | and
0.117117	-13.2327	given | or
0.153439	-13.2329	underlying | of
0.666667	-13.2329	our | agnostic
0.148515	-13.2329	highest | in
0.12963	-13.233	the | pairwise
0.833333	-13.233	allophonic | lexical
0.833333	-13.233	p(t|m)
0.342105	-13.2332	which | answer
0.114286	-13.2333	all | data
0.0883882	-13.2333	a | d
0.583333	-13.2334	when | off
0.179856	-13.2334	factors | in
0.363636	-13.2334	on average, | in
0.833333	-13.2336	46 | 40
0.351351	-13.2337	parameters: | of
0.311111	-13.2338	output | two
0.571429	-13.2338	surprisal | surprisal
0.571429	-13.2338	a2 | a2
0.571429	-13.2338	utah | city,
0.033691	-13.2338	each | in
0.257576	-13.2338	in | possible.
0.325581	-13.2338	summed | the
0.666667	-13.2339	methods | making
0.294118	-13.234	usually | than
0.3	-13.2341	part | nsf
0.141593	-13.2341	pi | pi
0.134387	-13.2341	interface | the
0.5	-13.2341	does not | significantly
0.172185	-13.2342	six | and
0.243243	-13.2342	evident | the
0.833333	-13.2342	set | 0.01,
0.833333	-13.2342	si | sj
0.169935	-13.2342	1000 | of
0.636364	-13.2343	author | papers
0.288462	-13.2343	the | tier
0.234568	-13.2344	the | clustering.
0.0935673	-13.2345	2.1 | the
0.142857	-13.2345	problem | are
0.0645756	-13.2345	the | of | systems
0.126761	-13.2346	which | information
0.473684	-13.2346	is | speaker.
0.222222	-13.2347	furthermore, | on
0.533333	-13.2347	bigrams | or
0.533333	-13.2347	use | describing
0.407407	-13.2347	study | corpus
0.5	-13.2348	gains | performance
0.253731	-13.2351	augmented | a
0.0753769	-13.2352	we have | in
0.3	-13.2352	results | baseline.
0.087931	-13.2352	the | modeling
0.375	-13.2352	only | maximum
0.124138	-13.2353	identification | of
0.636364	-13.2353	such | system’s
0.191667	-13.2354	so, | to
0.0857605	-13.2356	is | best
0.473684	-13.2356	word | morphemes
0.186047	-13.2357	scoring | for
0.636364	-13.2357	targets | only
0.0910747	-13.2358	experiment | of
0.191667	-13.2358	four | on
0.363636	-13.2358	we use | version
0.193277	-13.2358	this | cannot
0.666667	-13.2359	numerical expressions | can
0.583333	-13.2359	estimate | confidence
0.21	-13.236	in | correction
0.311111	-13.2361	to | baselines.
0.636364	-13.2362	used | huang
0.224719	-13.2362	recall, | and
0.271186	-13.2362	2006; | al., 2009).
0.234568	-13.2363	2002), | et
0.288462	-13.2364	results | automatically
0.583333	-13.2364	j., | m.,
0.125874	-13.2365	system | all
0.194915	-13.2365	to | unseen
0.352941	-13.2366	domain | labeled
0.3	-13.2366	bilingual | extraction
0.203704	-13.2366	furthermore, | are
0.142857	-13.2366	applied to | to
0.833333	-13.2366	maximizing | objective
0.0675403	-13.2366	there | that
0.636364	-13.2367	aims | identify
0.363636	-13.2368	bilingual | model.
0.0894454	-13.2368	2 | 4
0.352941	-13.2369	another | its
0.454545	-13.2369	total | over
0.280702	-13.2369	of | redundancy
0.171053	-13.237	optimized | the
0.0660836	-13.237	the | the | models.
0.342105	-13.237	ag | ag
0.159091	-13.2371	conclusion | that
0.392857	-13.2372	wmt | evaluation
0.392857	-13.2372	the | bought
0.234568	-13.2373	tests | in
0.407407	-13.2374	took | as
0.1875	-13.2376	decreases | the
0.636364	-13.2376	constraint | requires
0.636364	-13.2376	it | fair
0.117284	-13.2377	and | parameters
0.205607	-13.2377	the | ranking.
0.0405797	-13.2377	is | that | a
0.666667	-13.2378	curves | at
0.126316	-13.2378	the | raw
0.0604839	-13.2378	all | that
0.0805755	-13.2379	these | from
0.265625	-13.2381	for | control
0.158192	-13.2381	to | accuracy.
0.416667	-13.2382	model | predicting
0.311111	-13.2382	also | unsupervised
0.533333	-13.2382	novel | feature
0.375	-13.2382	with | preference
0.126316	-13.2383	may | more
0.275862	-13.2384	as | response
0.275862	-13.2384	evaluate | method
0.198198	-13.2384	we | numbers
0.311111	-13.2384	the | desr
0.149254	-13.2384	transformation | the
0.14486	-13.2385	discussed | and
0.342105	-13.2385	presented | algorithm
0.112676	-13.2385	our | sentences
0.116418	-13.2387	identified | of
0.318182	-13.2387	words | present
0.5	-13.2388	annotation | across
0.0868114	-13.239	three | for
0.222222	-13.2391	recognize | of
0.352941	-13.2392	metrics | system
0.454545	-13.2392	any | system.
0.5	-13.2392	or | outside
0.261538	-13.2393	we | removed
0.151832	-13.2393	perform | for
0.16092	-13.2396	also, | of
0.833333	-13.2397	conference | weblogs
0.342105	-13.2397	optimized | on
0.666667	-13.2398	errors | find
0.0477627	-13.2398	only | in
0.106329	-13.2398	provided | and
0.434783	-13.2399	anaphora | that
0.352941	-13.24	weighted | which
0.0848	-13.2401	the | here
0.15	-13.2402	we | prediction
0.833333	-13.2403	classified | categories
0.12963	-13.2406	data | different
0.288462	-13.2406	h | w
0.37931	-13.2406	additionally, | which
0.0579495	-13.2407	the | content
0.363636	-13.2408	widespread | of
0.280702	-13.2412	if | take
0.0723454	-13.2412	a | possible
0.434783	-13.2412	et | 1995).
0.434783	-13.2412	an | potentially
0.257576	-13.2412	absolute | for
0.144186	-13.2413	to | whereas
0.188976	-13.2413	only | few
0.0429272	-13.2413	1 | 2 | 2
0.304348	-13.2414	different | (e.g.
0.37931	-13.2415	german | translation
0.105392	-13.2417	of | selection
0.222222	-13.2417	number | size
0.0859504	-13.2419	candidates | the
0.0818452	-13.2419	the | context.
0.666667	-13.2422	resources | entailment
0.261538	-13.2422	used | specific
0.407407	-13.2422	observed | one
0.37931	-13.2423	filter | an
0.266667	-13.2424	can | increase
0.0534351	-13.2424	the | defined
0.392857	-13.2424	are | valuable
0.171053	-13.2425	features | pos
0.140351	-13.2425	with | label
0.0919325	-13.2426	partial | the
0.533333	-13.2426	that | ?
0.666667	-13.2427	random | compare
0.234568	-13.2427	value | 1
0.434783	-13.2428	result | baseline
0.193277	-13.2428	cue | the
0.275862	-13.243	(or | are
0.16	-13.243	the | material
0.363636	-13.243	meant | a
0.473684	-13.2432	option, | the
0.24	-13.2432	reveals | of
0.0430152	-13.2433	used | a
0.392857	-13.2433	various | were
0.168831	-13.2435	is | against
0.1875	-13.2435	an | representation
0.75	-13.2435	slu | first
0.24	-13.2435	including | or
0.2	-13.2435	to | parent
0.583333	-13.2436	can | alleviate
0.294118	-13.2437	in contrast, | are
0.533333	-13.2437	resources | these
0.186047	-13.2439	results | significantly
0.37931	-13.244	pronouns | are
0.363636	-13.244	is | betwen
0.352941	-13.2441	indicator | on
0.533333	-13.2441	system | generating
0.434783	-13.2441	used | automated
0.21978	-13.2441	(3) | are
0.109333	-13.2442	semantic | between
0.188525	-13.2442	that | easily
0.375	-13.2443	group | other
0.275862	-13.2444	terms | they
0.266667	-13.2444	it | unsupervised
0.148515	-13.2444	system | task
0.25	-13.2444	that | core
0.25	-13.2444	and | laboratory
0.222222	-13.2445	align | and
0.533333	-13.2445	frame | overall
0.342105	-13.2445	we | distributions.
0.253731	-13.2446	instructor | the
0.75	-13.2446	newswire | mt
0.333333	-13.2446	concept | which
0.151832	-13.2447	many | which
0.0955285	-13.2449	text | as
0.407407	-13.2449	follows: | ,
0.176056	-13.245	using | lexical
0.149254	-13.245	give | a
0.75	-13.2451	select | until
0.636364	-13.2451	study | significantly
0.333333	-13.2451	2000), | al.,
0.352941	-13.2452	if | share
0.311111	-13.2452	scenario | for
0.123288	-13.2452	wikipedia | to
0.473684	-13.2453	adverb | verb
0.198198	-13.2455	example | our
0.165644	-13.2455	english | corpus
0.128676	-13.2456	in | initial
0.666667	-13.2456	modular | type
0.666667	-13.2456	result | greater
0.392857	-13.2456	have | parent
0.75	-13.2457	history | previous
0.25	-13.2457	is | practice
0.533333	-13.2457	transition | weight
0.257576	-13.2457	various | or
0.13913	-13.2457	, | 1
0.169935	-13.2459	this, | a
0.222222	-13.246	continuous | of
0.636364	-13.246	approach | (2010),
0.179856	-13.246	of | 200
0.16092	-13.2461	of | affect
0.0912477	-13.2461	using | which
0.5	-13.2462	mt | alignment
0.311111	-13.2464	relations | text
0.311111	-13.2464	obtained | over
0.407407	-13.2464	only | required
0.533333	-13.2465	seven | are
0.666667	-13.2466	2 5 | .
0.246575	-13.2466	a | producing
0.25	-13.2466	salience | the
0.119741	-13.2468	do | have
0.304348	-13.247	model, | only
0.666667	-13.2471	all | punctuation.
0.21978	-13.2472	for | interaction
0.152632	-13.2472	pairs | or
0.135802	-13.2473	current | is
0.75	-13.2473	limit | size
0.0548128	-13.2473	difference | the
0.407407	-13.2473	the | (artiles
0.261538	-13.2474	the | decreased
0.636364	-13.2474	task | pubmed
0.0538611	-13.2474	the | of | new
0.275862	-13.2475	sentence | phrases
0.162651	-13.2476	labels | are
0.137339	-13.2476	expressions | a
0.434783	-13.2476	if | edit
0.434783	-13.2476	label | vertices
0.533333	-13.2477	rules | patterns.
0.833333	-13.2477	highlight | importance
0.833333	-13.2477	grid | optimal
0.21	-13.2477	are | against
0.275862	-13.2479	estimated | with
0.144186	-13.248	post | the
0.217391	-13.248	system | input
0.666667	-13.2481	present | preliminary
0.294118	-13.2481	simple | where
0.311111	-13.2482	(2002) | for
0.280702	-13.2484	all | pairs.
0.636364	-13.2484	microsoft | asia
0.186047	-13.2485	have | them
0.392857	-13.2485	nsummationdisplay | where
0.473684	-13.2485	experimental | datasets
0.342105	-13.2485	pairs | comparable
0.352941	-13.2486	if | enough
0.253731	-13.2486	use | (e.g.,
0.253731	-13.2486	is | (section
0.0452468	-13.2489	information | and
0.325	-13.249	instead, | are
0.0412566	-13.249	words | in
0.087931	-13.2491	monolingual | the
0.311111	-13.2491	earlier | that
0.205882	-13.2492	it | thus
0.190083	-13.2492	the | sequences.
0.434783	-13.2495	how | represent
0.75	-13.2496	0.57 | 0.49
0.75	-13.2496	(agichtein | 2000;
0.75	-13.2496	listings | men’s
0.75	-13.2496	cornell | ithaca,
0.75	-13.2496	isr | isr
0.75	-13.2496	0.23 | 0.18
0.75	-13.2496	backward | backward
0.533333	-13.2496	equivalence | by
0.144186	-13.2497	as | being
0.583333	-13.2497	@ | ¸@
0.271186	-13.2497	corpus | development
0.069869	-13.2498	because | in
0.416667	-13.2499	all | none
0.416667	-13.2499	many | patterns
0.243243	-13.2499	use | corpora
0.0786207	-13.2499	and | systems
0.21978	-13.2499	the | observing
0.304348	-13.25	me | in
0.137339	-13.25	we | performance.
0.167742	-13.25	while | data
0.181818	-13.25	a | ...
0.266667	-13.2501	that | bad
0.833333	-13.2501	overall | product.
0.294118	-13.2502	ith | is
0.45	-13.2502	corpus | modern
0.181818	-13.2502	learned | for
0.325	-13.2504	(2010), | that
0.21	-13.2504	increasing | in
0.0321387	-13.2505	the | over
0.392857	-13.2505	of | positions.
0.109333	-13.2506	occurrences | the
0.666667	-13.2506	allows | domain
0.158192	-13.2506	sentence | translation
0.140351	-13.2507	while | this
0.0715935	-13.2508	now | the
0.0870307	-13.2508	selection | of
0.201835	-13.2508	modified | to
0.0452882	-13.2508	the | annotation
0.318182	-13.2509	mt | reference
0.143519	-13.251	to | global
0.5	-13.251	approaches | automatically
0.294118	-13.2511	belief | for
0.0859504	-13.2511	for | pairs
0.217391	-13.2512	splits | the
0.318182	-13.2514	the | lower.
0.45	-13.2515	this | table 3.
0.164634	-13.2515	a | he
0.275862	-13.2516	statistics | data
0.342105	-13.2516	other | models.
0.236842	-13.2517	the | translates
0.636364	-13.2517	hereafter, | and
0.25	-13.2518	as | language,
0.833333	-13.252	task, | current
0.666667	-13.2521	constraints | ilp
0.636364	-13.2521	created | reference
0.37931	-13.2521	for | operators
0.0528967	-13.2521	the | is | one
0.318182	-13.2521	corpus | frequency
0.318182	-13.2521	over | input
0.205882	-13.2522	value | each
0.473684	-13.2523	and | katakana
0.215054	-13.2524	perplexity | and
0.152632	-13.2524	for | 2007).
0.45	-13.2525	are | handled
0.0858086	-13.2525	of | common
0.666667	-13.2526	are | following.
0.583333	-13.2527	joint | independent
0.363636	-13.2527	of | total.
0.407407	-13.2527	situation | by
0.333333	-13.2528	the | annotators’
0.45	-13.2529	21 | 25
0.24	-13.2529	from | (the
0.325	-13.2529	in | belongs to
0.173611	-13.253	also | models
0.198198	-13.2531	extended | for
0.434783	-13.2531	relation | entities.
0.112717	-13.2532	instance | a
0.37931	-13.2532	due to | some
0.145631	-13.2534	update | of
0.0221952	-13.2534	is | that
0.116923	-13.2535	the | identical
0.283019	-13.2535	parameter | training
0.176056	-13.2536	w/o | w/o
0.533333	-13.2536	assumption | often
0.0993377	-13.2537	they | by
0.257576	-13.2537	however, | methods
0.0580256	-13.2538	for | use
0.37931	-13.2538	screenshot | the
0.325	-13.2538	methods | new
0.416667	-13.2539	evolution | a
0.111732	-13.254	a | linear
0.0727056	-13.254	events | the
0.37931	-13.2541	of | posts.
0.158192	-13.2541	such | they
0.294118	-13.2542	an | left
0.318182	-13.2542	overall | over
0.117284	-13.2543	inference | to
0.236842	-13.2543	the | suggestions
0.176056	-13.2544	paraphrases | a
0.25	-13.2544	merely | the
0.416667	-13.2545	” | ”
0.636364	-13.2545	of | plausible.
0.152632	-13.2545	using | described
0.0926641	-13.2546	the | complete
0.37931	-13.2546	filled | a
0.45	-13.2546	experiments, | best
0.103118	-13.2547	the | constituent
0.416667	-13.2548	our | posts
0.666667	-13.255	performs | tasks
0.666667	-13.255	tuned | optimize
0.636364	-13.255	patterns | include
0.271186	-13.255	agreement | annotation
0.155556	-13.2551	as | within
0.196429	-13.2551	threshold | for
0.130268	-13.2551	mining | and
0.0926641	-13.2551	to | class
0.533333	-13.2552	features | unseen
0.103118	-13.2553	resulting | of
0.288462	-13.2553	extensive | in
0.318182	-13.2554	examples | sentences
0.207921	-13.2557	a | symbols
0.174825	-13.2558	the | valuable
0.198198	-13.2558	pair | not
0.325	-13.2558	measure | lexical
0.123288	-13.2558	to | nlp
0.636364	-13.2559	baselines | compare
0.363636	-13.2559	investigating | a
0.666667	-13.256	show | improvement.
0.325	-13.256	object | an
0.212121	-13.2562	used | those
0.234568	-13.2562	λ | a
0.21	-13.2563	we | 200
0.253731	-13.2565	of | collections
0.24	-13.2565	the | classifies
0.2	-13.2566	author | to
0.271186	-13.2566	our | classifiers
0.0960334	-13.2567	setting | the
0.363636	-13.2567	li | for
0.392857	-13.2567	we | prove
0.193277	-13.2567	in | rest
0.113043	-13.2568	we | train
0.45	-13.2569	processing | sign
0.13913	-13.2569	contains | for
0.416667	-13.257	(i.e., | its
0.113703	-13.257	and | tasks
0.271186	-13.257	follow | with
0.21978	-13.2571	the | gathered
0.203883	-13.2573	of | targets
0.434783	-13.2573	language | grammatical
0.181818	-13.2573	this | perform
0.0860034	-13.2574	identify | of
0.45	-13.2576	these | variants
0.283019	-13.2577	system | provided
0.0818182	-13.2579	the | factor
0.24	-13.2579	university | research
0.108466	-13.2579	while | not
0.118971	-13.258	include | as
0.142202	-13.258	in | semantics
0.231707	-13.2581	fixed | for
0.363636	-13.2581	they | under
0.124113	-13.2582	which | semantic
0.174825	-13.2582	the | convergence
0.45	-13.2582	of | year.
0.833333	-13.2582	with | relationships.
0.833333	-13.2582	& | 2010.
0.257576	-13.2582	the | exposure
0.0501126	-13.2583	(1) | the
0.275862	-13.2584	corpus | web
0.583333	-13.2584	corpus | scientific
0.257576	-13.2584	it | future
0.666667	-13.2585	rule | yield
0.363636	-13.2586	j, | is
0.103704	-13.2587	had | of
0.318182	-13.2587	which | it.
0.833333	-13.2588	repetitions | same
0.283019	-13.2588	are | otherwise
0.416667	-13.2589	prosodic | speech
0.434783	-13.2589	ny | usa
0.205882	-13.259	wm | the
0.157303	-13.259	typically | in
0.142202	-13.259	the | somewhat
0.246575	-13.2591	for | hindi
0.173611	-13.2591	each | node
0.176056	-13.2592	is | entire
0.304348	-13.2592	ambiguities | and
0.45	-13.2593	correlation | judgment
0.121212	-13.2593	importance | the
0.363636	-13.2594	together, | in
0.333333	-13.2595	section | gives
0.833333	-13.2595	sentiwordnet 3.0: | for
0.0926641	-13.2596	translation | from
0.294118	-13.2596	a | segmentation.
0.45	-13.2596	user’s | with
0.636364	-13.2597	compressions | deletion
0.636364	-13.2597	skip | skip
0.636364	-13.2597	[task | subtask
0.636364	-13.2597	filled | slot3,
0.636364	-13.2597	5k | 5k
0.257576	-13.2597	experiments | proposed
0.150259	-13.2598	of | english.
0.0947581	-13.2598	the | transition
0.169935	-13.2599	for | performance.
0.136752	-13.2601	original | is
0.833333	-13.2601	paths | structure
0.434783	-13.2602	constructions | by
0.161677	-13.2603	of | descriptions
0.128302	-13.2603	2000; | and
0.0883392	-13.2603	based | on
0.152632	-13.2604	the | content.
0.342857	-13.2604	that | incorrectly
0.666667	-13.2605	data | mixed
0.186992	-13.2605	which | them
0.288462	-13.2605	direct | or
0.228916	-13.2605	to | classifier,
0.0750636	-13.2605	the | tasks
0.304348	-13.2606	of | genes
0.45	-13.2606	identify | types
0.163636	-13.2606	approach | use
0.342857	-13.2606	approaches | data.
0.5	-13.2607	agreement | judges
0.0808824	-13.2608	done | the
0.416667	-13.2608	is | critical
0.342105	-13.2608	in general, | with
0.113703	-13.2609	labels | in
0.37931	-13.2609	since | specific
0.45	-13.2609	syntactic | complex
0.45	-13.2609	by | tagger.
0.0562278	-13.2609	and | results
0.666667	-13.261	measure | useful
0.311111	-13.261	to | output,
0.101896	-13.2612	the | minimum
0.533333	-13.2612	systems | outperform
0.833333	-13.2613	modeling | syntax
0.352941	-13.2614	of | cavity
0.666667	-13.2615	values | below
0.136752	-13.2615	we have | from
0.246575	-13.2615	this | fully
0.079096	-13.2616	the | generation
0.352941	-13.2617	comparison | previous
0.304348	-13.2617	this | providing
0.0746835	-13.2617	of | sentence.
0.37931	-13.2617	since | words,
0.181818	-13.2617	he | in
0.266667	-13.2618	this | greater
0.138528	-13.2618	there | a | between
0.311111	-13.2619	12 | by
0.434783	-13.2619	parsing | linear
0.666667	-13.262	its | neighbors
0.833333	-13.262	target) | we
0.150259	-13.2622	with | function
0.257576	-13.2622	that | mostly
0.666667	-13.2625	represent | features,
0.666667	-13.2625	users | type
0.190083	-13.2625	partially | and
0.158824	-13.2626	of | beyond
0.040293	-13.2626	target | the
0.833333	-13.2626	information | specificity
0.21	-13.2626	dependency | syntactic
0.119741	-13.2629	of | (section
0.342857	-13.2629	tests | on
0.0798258	-13.263	is | no
0.212766	-13.2631	of | predicates.
0.13964	-13.2632	learning | training
0.25	-13.2632	do | contain
0.833333	-13.2632	interpretation | predicates
0.416667	-13.2633	m. | m.
0.0825545	-13.2633	after | to
0.21	-13.2634	proposed | word
0.11746	-13.2635	good | in
0.297872	-13.2636	often | same
0.154696	-13.2636	evidence | a
0.125899	-13.2637	in | 4,
0.283019	-13.2637	exploits | of
0.283019	-13.2637	expression | an
0.161677	-13.2638	record | the
0.0909091	-13.2639	define | and
0.45	-13.264	it | pass
0.434783	-13.2641	as | essential
0.277778	-13.2644	all | final
0.666667	-13.2645	following | a.
0.294118	-13.2645	selected, | the
0.833333	-13.2645	repeated | until
0.215054	-13.2646	aligned | target
0.45	-13.2647	free | order
0.149485	-13.2648	4 | 8
0.180451	-13.2648	classes, | the
0.666667	-13.265	model | explanatory
0.145631	-13.2651	sources | and
0.833333	-13.2651	now | just
0.257576	-13.2651	result | our
0.190083	-13.2652	to | structures.
0.311111	-13.2652	up | data
0.288462	-13.2653	our | predictions
0.583333	-13.2654	et | 2011a).
0.24	-13.2655	vp | pp
0.833333	-13.2657	english: | from
0.325	-13.2657	our | generating
0.416667	-13.2658	lexicons | words
0.207921	-13.2658	to | normal
0.164557	-13.2658	4 | 10
0.121212	-13.2658	used | words
0.127341	-13.2659	output | for
0.271186	-13.2659	used | bilingual
0.145631	-13.266	the | unseen
0.217391	-13.2662	whether | has
0.123675	-13.2662	data | corpus
0.0576194	-13.2663	a | text
0.257576	-13.2663	the | subsequently
0.5	-13.2663	experiments | strong
0.352941	-13.2665	for | years.
0.266667	-13.2666	obtained | corpus
0.0971922	-13.2666	in | content
0.311111	-13.2666	for | humans
0.125899	-13.2667	we | select
0.363636	-13.2668	complexity, | a
0.236842	-13.2669	in | tweet
0.108401	-13.2669	are | set of
0.666667	-13.267	as | bed
0.304348	-13.267	learning | knowledge
0.297872	-13.267	the | arguably
0.177778	-13.267	of | modifier
0.22619	-13.2671	section, | describe
0.191667	-13.2671	a | labels.
0.304348	-13.2675	within | between
0.333333	-13.2675	resource | semantic
0.24	-13.2675	not | overall
0.342857	-13.2676	of | eisner
0.0793651	-13.2677	after | and
0.196429	-13.2678	explain | of
0.185484	-13.2678	the | lies
0.277778	-13.2678	approach | without
0.243243	-13.2679	as | towards
0.416667	-13.268	for | predicate,
0.416667	-13.268	and | 2006b;
0.533333	-13.268	complete | its
0.233766	-13.268	a | configuration
0.333333	-13.2682	likewise | the
0.0648855	-13.2682	over | in
0.434783	-13.2684	the | kept.
0.434783	-13.2684	human | identify
0.192982	-13.2686	two | where | is
0.257576	-13.2686	that | act
0.210526	-13.2686	an | metric
0.5	-13.2686	a | degree,
0.275862	-13.2687	fill | the
0.13253	-13.2687	oregon, | june
0.0774457	-13.2688	represented | the
0.304348	-13.2688	for | adjacent
0.37931	-13.2689	of | scientists
0.5	-13.2689	many | ways
0.5	-13.2689	target | always
0.5	-13.2689	but | restricted
0.236842	-13.2691	two | gold
0.304348	-13.2691	syllables | the
0.0907372	-13.2691	all | or
0.203883	-13.2691	is | reasonable
0.317073	-13.2691	and | another.
0.0824261	-13.2692	2007; | al.,
0.183206	-13.2692	to | around
0.103448	-13.2693	it | different
0.0718492	-13.2693	while | for
0.583333	-13.2693	models | handle
0.123675	-13.2693	is | larger
0.212766	-13.2695	which | out
0.833333	-13.2695	strategies | techniques
0.392857	-13.2696	sa | user:
0.170068	-13.2697	follow | in
0.0493622	-13.2697	the | of | also
0.228916	-13.2697	provides | which
0.0822981	-13.2697	as | only
0.45	-13.2697	is | plain
0.325	-13.2697	the | transferred
0.325	-13.2697	defines | that
0.124113	-13.2698	to | features,
0.262295	-13.27	the | lattice.
0.106771	-13.2701	in | development
0.0948454	-13.2701	the | grammatical
0.325	-13.2702	trained | domain
0.666667	-13.2705	dependency | probabilities
0.666667	-13.2705	introduce | variable
0.22619	-13.2706	the | really
0.583333	-13.2706	thus, | annotation
0.583333	-13.2706	list | items
0.352941	-13.2707	that | optimize
0.246377	-13.2707	document | distribution
0.283019	-13.2707	with | prosodic
0.533333	-13.2708	contrast, | its
0.392857	-13.2708	assigned | value
0.0532468	-13.2708	evaluation | and
0.129278	-13.2709	unsupervised | for
0.666667	-13.271	10, | 10
0.106218	-13.271	test | from
0.253731	-13.271	pair, | to
0.120401	-13.271	produced | of
0.24	-13.2711	analysed | the
0.228916	-13.2712	using | statistical
0.533333	-13.2712	although | increase
0.21978	-13.2713	2 related work | al.,
0.0895795	-13.2713	then | with
0.205882	-13.2714	for | too
0.22619	-13.2715	classified | as
0.342857	-13.2715	nnp | nn
0.277778	-13.2716	using | latent
0.0775623	-13.2717	d | the
0.271186	-13.2717	about | target
0.666667	-13.2721	find | exact
0.416667	-13.2721	parameters | model,
0.228916	-13.2721	(c) | a
0.154696	-13.2722	because | an
0.0331408	-13.2722	are | are
0.192982	-13.2724	and | feedback
0.533333	-13.2724	represent | represents
0.533333	-13.2724	requires | domain
0.533333	-13.2724	model | augmented
0.352941	-13.2726	non-terminal | is
0.352941	-13.2726	composition | with
0.666667	-13.2726	co-occurrence | lexical
0.666667	-13.2726	2005), | 2005),
0.666667	-13.2726	observing | from
0.37931	-13.2726	used | filter
0.833333	-13.2726	multi-class | problem
0.277778	-13.2726	and | solutions
0.0903955	-13.2727	representation | and
0.416667	-13.2727	metadata | for
0.262295	-13.2727	may | better
0.434783	-13.2727	analysis | performed
0.0815385	-13.2728	case | and
0.13617	-13.2728	like | are
0.145631	-13.2729	choose | and
0.133065	-13.2729	of | tree.
0.2	-13.273	come | a
0.366667	-13.2731	corpus | 2005)
0.325	-13.2731	0 | (1)
0.304348	-13.2732	this | around
0.45	-13.2732	of | correlation.
0.833333	-13.2733	books | books,
0.833333	-13.2733	curve | obtained
0.833333	-13.2733	settings | settings
0.352941	-13.2734	preference | with
0.22619	-13.2734	experiments | both | and
0.141553	-13.2735	to | alignments
0.165605	-13.2735	real | and
0.416667	-13.2736	proof | in
0.246377	-13.2736	part | our
0.533333	-13.2736	cartesian | of
0.262295	-13.2737	use | supervised
0.363636	-13.2737	blog | on
0.366667	-13.2737	i=1 | )
0.366667	-13.2737	that | viable
0.0547196	-13.2738	user | the
0.0971922	-13.2739	has | or
0.116719	-13.274	information, | the
0.473684	-13.274	values | average
0.236842	-13.2741	with | million
0.666667	-13.2741	also, | features
0.666667	-13.2741	definite | demonstrative
0.180451	-13.2741	former | of
0.583333	-13.2742	appears | once
0.114458	-13.2742	for | (e.g.,
0.152174	-13.2743	these | may
0.271186	-13.2743	value | was
0.392857	-13.2744	in | instructor
0.352941	-13.2745	that | transformations
0.179104	-13.2745	the | right-hand
0.833333	-13.2745	open | close
0.833333	-13.2745	classifier | performance.
0.160714	-13.2746	the | dependency tree
0.283019	-13.2746	an | point
0.288462	-13.2747	annotated | their
0.4	-13.2748	polarity | using
0.416667	-13.2749	only | ambiguous
0.19469	-13.2749	a | image
0.253731	-13.275	we | type.
0.147208	-13.275	in | appears
0.158824	-13.2751	a | portion
0.583333	-13.2751	relative | across
0.311111	-13.2751	2004; | 2009).
0.253731	-13.2752	the | identified.
0.129921	-13.2752	of | authors
0.266667	-13.2753	global | from
0.311111	-13.2753	an | application
0.366667	-13.2754	order | create
0.148718	-13.2754	srl | the
0.317073	-13.2754	set | independent
0.12782	-13.2754	only | features
0.100939	-13.2754	the | previously
0.5	-13.2754	are | nature.
0.271186	-13.2754	an | utterance
0.416667	-13.2755	results | 7.
0.149485	-13.2756	we | procedure
0.666667	-13.2756	an | factor,
0.161677	-13.2758	sequences | to
0.333333	-13.2758	for | p.
0.24	-13.2758	cr | the
0.833333	-13.2758	cost | tuned
0.5	-13.2758	information | therefore
0.583333	-13.2759	in | 2008b;
0.583333	-13.2759	aac | their
0.583333	-13.2759	researchers | algorithms
0.132	-13.276	available | for
0.174825	-13.276	of | grammar.
0.311111	-13.276	to | descriptions.
0.666667	-13.2761	consists | applying
0.333333	-13.2761	noticeable | the
0.392857	-13.2762	thus, | would
0.095339	-13.2762	the | people
0.185484	-13.2763	our | provide
0.184	-13.2763	become | in
0.833333	-13.2764	received | 5
0.833333	-13.2764	n | γ
0.533333	-13.2765	are | ignored.
0.434783	-13.2766	a | sentiment.
0.0868056	-13.2767	present | for
0.416667	-13.2768	joint | syntactic
0.128906	-13.2768	collection | a
0.366667	-13.2768	ii | to
0.304348	-13.2769	of | academy
0.210526	-13.2769	that: | the
0.22619	-13.277	the | independence
0.243243	-13.277	evaluator | evaluator
0.434783	-13.277	emnlp | processing,
0.833333	-13.2771	utility | information
0.352941	-13.2772	and | usages
0.583333	-13.2773	elementary | functions
0.228916	-13.2773	a) | and
0.473684	-13.2773	although | good
0.5	-13.2773	count | any
0.25	-13.2775	the | string-to-tree
0.363636	-13.2775	also | existing
0.352941	-13.2777	extracted | bilingual
0.24	-13.2777	used | models.
0.833333	-13.2777	model | f|e
0.104592	-13.2778	term | to
0.288462	-13.2779	are usually | to
0.291667	-13.2779	data | 2010
0.135021	-13.278	5 | results
0.342857	-13.278	k | l
0.666667	-13.2781	koo | (2010)
0.203883	-13.2781	a | classes.
0.124555	-13.2781	to | instance
0.583333	-13.2782	improve | current
0.228916	-13.2782	when | given
0.416667	-13.2784	that | connotation
0.283019	-13.2784	each | against
0.101695	-13.2785	that | much
0.0680851	-13.2785	for | feature
0.223529	-13.2785	shown | results
0.161677	-13.2786	techniques | as
0.666667	-13.2786	extraction | pantel,
0.666667	-13.2786	2003), | 2010),
0.215054	-13.2786	for | predicates
0.18254	-13.2786	the | knowledge base
0.113095	-13.2787	a | during
0.266667	-13.2788	less | when
0.114114	-13.2789	second, | of
0.144928	-13.2791	equation | is
0.116352	-13.2791	for | parser
0.13278	-13.2793	structures | to
0.236842	-13.2793	a | triggers
0.533333	-13.2793	determining | sentiment
0.124113	-13.2796	also | used
0.144928	-13.2796	can | better
0.233766	-13.2796	the | requirements
0.131474	-13.2797	this | without
0.271186	-13.2797	(1998) | and
0.0834671	-13.2798	and | type
0.0948454	-13.2798	the | representations
0.288462	-13.2799	our | traditional
0.288462	-13.2799	this results in | of
0.583333	-13.2799	grouped | into
0.208333	-13.2799	probably | of
0.283019	-13.2799	we | feature selection
0.253731	-13.28	implementation | with
0.242857	-13.28	first, | an
0.5	-13.28	attitude | attitude
0.132	-13.2801	is | random
0.100939	-13.2801	of | existing
0.666667	-13.2802	t, | r
0.25	-13.2802	unlike | are
0.363636	-13.2803	der | den
0.363636	-13.2803	acc | acc
0.114804	-13.2803	the | affect
0.352941	-13.2804	are | free
0.128405	-13.2804	the | positions
0.22619	-13.2804	of | schemes
0.135593	-13.2804	study, | the
0.392857	-13.2804	distribution | corpus.
0.123239	-13.2804	selection | for
0.473684	-13.2805	on | aac
0.4	-13.2806	that | stated
0.14218	-13.2808	we | dictionary
0.583333	-13.2808	and | shot
0.325	-13.2808	grows | of
0.366667	-13.2809	relevant | given
0.0816641	-13.2809	furthermore, | the
0.113433	-13.2809	to | constraints
0.201923	-13.281	through | or
0.114907	-13.281	for | example
0.4	-13.2812	involve | which
0.176471	-13.2812	benefit from | the
0.123675	-13.2813	if | only | the
0.473684	-13.2813	method | evaluated
0.217391	-13.2815	the | closed
0.262295	-13.2815	all | nouns
0.228916	-13.2817	(2009), | of
0.277778	-13.2817	compositionality | in
0.0738854	-13.2817	type | and
0.266667	-13.2818	details | section
0.0811639	-13.2818	of | patterns
0.366667	-13.2818	the | three.
0.325	-13.2818	(%) | for
0.201923	-13.2819	producing | of
0.0594463	-13.2819	syntactic | and
0.191304	-13.2819	a | itself
0.0409583	-13.2819	we | the | using
0.5	-13.2819	that | image.
0.5	-13.2819	convolution | and
0.0458273	-13.282	the | terms
0.186992	-13.282	had | on
0.666667	-13.2822	rule | β
0.157895	-13.2822	much | as
0.473684	-13.2823	and | baselines,
0.342857	-13.2824	phase, | in
0.0931174	-13.2826	of | trees
0.366667	-13.2826	3.2 | algorithm
0.333333	-13.2827	algorithm | model.
0.0330058	-13.2828	they | the
0.236842	-13.2828	on | 23
0.416667	-13.2828	such | seen
0.236842	-13.2829	select | each
0.0574713	-13.2829	the | argument
0.152174	-13.2829	entries | and
0.185484	-13.283	expanded | the
0.325	-13.283	an | analysis.
0.666667	-13.2832	reordering | translations.
0.304348	-13.2832	in | wordnet.
0.102941	-13.2833	with | target
0.304348	-13.2834	the | assignment.
0.0836013	-13.2834	comparing | the
0.112094	-13.2834	it | two
0.228916	-13.2836	of | partially
0.283019	-13.2836	of | distance.
0.666667	-13.2837	when | hit
0.291667	-13.2837	to | expectation
0.210526	-13.2837	facts | the
0.416667	-13.2838	statistically | from
0.121107	-13.2838	the | descriptions
0.137778	-13.2838	tweet | the
0.272727	-13.2839	a | verbs,
0.333333	-13.2839	lda | from
0.122807	-13.2839	concepts | and
0.05995	-13.2839	we | its
0.10219	-13.284	method | from
0.262295	-13.2841	consequently, | to
0.45	-13.2841	haitian | sms
0.115987	-13.2842	the | direction
0.473684	-13.2842	task | providing
0.215054	-13.2843	how | information
0.325	-13.2843	required, | the
0.366667	-13.2846	the | too.
0.0945378	-13.2847	the | independent
0.666667	-13.2848	measures | results.
0.24	-13.2848	jump | the
0.133891	-13.2848	a | named
0.271186	-13.2849	it | information.
0.102941	-13.285	not | other
0.208333	-13.285	the | µ
0.416667	-13.2851	reductions | for
0.253731	-13.2851	third | with
0.179104	-13.2851	an | search
0.666667	-13.2853	dictionaries | languages.
0.297872	-13.2853	two | models,
0.161491	-13.2853	widely | in
0.473684	-13.2853	best | arabic
0.297872	-13.2855	marks | in
0.147208	-13.2856	from | single
0.283019	-13.2856	of | countries
0.196429	-13.2857	each | 10
0.5	-13.2857	lateen | objective
0.666667	-13.2858	p(s|t)
0.277778	-13.2858	denotes | number
0.277778	-13.2858	states | or
0.45	-13.2859	metric | evaluate
0.203883	-13.2861	back | to
0.127341	-13.2862	we | existing
0.583333	-13.2862	show | achieving
0.0374715	-13.2863	a | with | and
0.366667	-13.2863	this | generalized
0.131687	-13.2863	of | prediction
0.126866	-13.2863	thus | for
0.188525	-13.2863	words | syntactic
0.416667	-13.2864	second, | only
0.236842	-13.2866	the | treelets
0.833333	-13.2866	tweet | get
0.304348	-13.2867	use | framework
0.165605	-13.2867	consider | from
0.212766	-13.2868	the | name,
0.666667	-13.2868	rates | rates
0.079063	-13.2868	human | to
0.233766	-13.2869	including | which
0.325	-13.287	each | multinomial
0.258065	-13.2871	following | these
0.666667	-13.2873	or | sequences.
0.272727	-13.2874	with | sequences
0.25	-13.2875	engine | a
0.583333	-13.2875	says | then
0.317073	-13.2875	simple | text
0.137778	-13.2876	sentence | words
0.198113	-13.2877	for | used.
0.309524	-13.2877	messages | from
0.666667	-13.2878	baseline, | model.
0.155172	-13.2878	considered | for
0.304348	-13.2879	to | an optimal
0.179104	-13.2879	data, | are
0.533333	-13.2879	step | step.
0.833333	-13.2879	its | translations,
0.833333	-13.2879	once | selected
0.121951	-13.2879	the | phrases.
0.104592	-13.288	and | argument
0.156069	-13.2881	that | made
0.266667	-13.2882	in | modules
0.266667	-13.2882	present | corpus
0.22619	-13.2882	and | fine-grained
0.198113	-13.2882	from | manually
0.325	-13.2883	clusters | more
0.143541	-13.2883	word | given
0.19469	-13.2883	because | was
0.666667	-13.2884	state | star
0.666667	-13.2884	improved | combination
0.45	-13.2886	assume | so
0.223529	-13.2886	and | operations
0.533333	-13.2887	arguments | frame
0.177778	-13.2887	common | by
0.342857	-13.2887	slot | from
0.37931	-13.2888	example | annotated
0.5	-13.2888	our | possibly
0.352941	-13.2891	about | documents
0.0945378	-13.2891	model | data
0.0440449	-13.2891	in | one
0.156069	-13.2892	the | tags,
0.5	-13.2892	quality | translated
0.181102	-13.2893	the | efficiency
0.583333	-13.2893	proposed | extended
0.473684	-13.2893	subject | direct
0.666667	-13.2894	my | .
0.37931	-13.2894	say | or
0.121107	-13.2896	trees | in
0.191304	-13.2896	are | aligned
0.1625	-13.2897	5 | each
0.0842975	-13.2897	to | linguistic
0.45	-13.2897	because | include
0.583333	-13.2898	references | text
0.833333	-13.2898	many | exist,
0.352941	-13.2899	ctb | and
0.266667	-13.29	a | shallow
0.218391	-13.29	to | length.
0.173913	-13.29	pp | np
0.176471	-13.29	we | representations
0.366667	-13.29	the | believes
0.392857	-13.29	0.03 | 0.03
0.473684	-13.29	kernel | relation
0.266667	-13.2902	clusters | each
0.156977	-13.2902	path | in
0.230769	-13.2902	expert | for
0.139013	-13.2903	3.1 | we
0.0657895	-13.2904	reference | the
0.126394	-13.2905	proposed | with
0.833333	-13.2905	paraphrases | sentences,
0.0957447	-13.2907	that | large
0.147208	-13.2907	edge | is
0.212766	-13.2908	understanding | is
0.352941	-13.291	performance | individual
0.833333	-13.2911	drew | from
0.155172	-13.2912	to | text,
0.297872	-13.2913	an initial | is
0.201923	-13.2915	2004), | al.,
0.262295	-13.2916	to | ratings
0.533333	-13.2916	or | nodes.
0.333333	-13.2916	each | nonterminal
0.0641762	-13.2916	not | or
0.37931	-13.2918	has | parameters
0.192982	-13.2919	model | shown
0.333333	-13.2919	1994; | et
0.253968	-13.2919	the | disambiguated
0.666667	-13.292	children | tree.
0.37931	-13.292	provide | they
0.153005	-13.2921	a few | and
0.272727	-13.2923	could | any
0.127413	-13.2923	the | problem,
0.5	-13.2923	eecs | of
0.342857	-13.2924	information | further
0.666667	-13.2925	candidates | pattern
0.666667	-13.2925	first | load
0.242857	-13.2925	polarity | negative
0.0695553	-13.2927	problem | and
0.236842	-13.293	f-scores | the
0.666667	-13.293	condition | {
0.5	-13.293	alignments | generated
0.272727	-13.2931	reports | a
0.147208	-13.2931	source | target | is
0.198113	-13.2931	a | s,
0.177778	-13.2931	(2009), | the
0.139535	-13.2931	the | sms
0.22619	-13.2933	(2007), | and
0.277778	-13.2933	the | open source
0.262295	-13.2934	the | error,
0.145	-13.2935	some | words
0.317073	-13.2936	it is | natural
0.2	-13.2936	examined | and
0.473684	-13.2937	of | thread.
0.210526	-13.2937	this | require
0.0643275	-13.2938	this | is | as
0.416667	-13.2938	dan | and
0.0865724	-13.2939	nouns | the
0.309524	-13.294	training | objective
0.186992	-13.2941	presents a | of
0.4	-13.2941	(2010) | english
0.101942	-13.2942	used | word
0.242857	-13.2942	in | category.
0.583333	-13.2943	problem | address
0.154286	-13.2943	is | metric
0.236842	-13.2944	compute | using
0.533333	-13.2945	systems | question
0.333333	-13.2945	implemented | an
0.0726601	-13.2946	score | a
0.127413	-13.2946	allowed | the
0.0947368	-13.2947	and | through
0.416667	-13.2947	precision | both
0.185484	-13.2948	weighted | for
0.583333	-13.2948	when | gets
0.366667	-13.2948	dataset | pairs
0.119863	-13.2951	with | no
0.139535	-13.2952	of | strategies
0.37931	-13.2953	such | already
0.19469	-13.2954	which | multiple
0.5	-13.2954	german | was
0.352941	-13.2956	liang | al.
0.666667	-13.2956	whether | holds
0.246377	-13.2956	discussions | and
0.139535	-13.2956	similarly | the
0.538462	-13.2957	methods | 1998;
0.5	-13.2957	shared | organizers
0.45	-13.2959	an | y.
0.473684	-13.2959	multinomial | with
0.215909	-13.296	our | corpus,
0.19469	-13.296	needs | and
0.666667	-13.2961	pages | those
0.198113	-13.2961	method | use
0.538462	-13.2961	only | terms.
0.473684	-13.2962	existing | document
0.45	-13.2963	these | object
0.333333	-13.2963	the | mt output.
0.266667	-13.2964	was | 5
0.5	-13.2965	but | minimal
0.4	-13.2966	is | for.
0.583333	-13.2966	some | whole
0.132231	-13.2966	objective | is
0.115625	-13.2966	and | work.
0.533333	-13.2966	analyzed | an
0.473684	-13.2966	each | has | own
0.473684	-13.2966	of | bitext.
0.317073	-13.2966	we | transform
0.0583067	-13.2966	that | different
0.115987	-13.2969	a | (the
0.246377	-13.2969	of | predicate.
0.233766	-13.2969	finally | to
0.538462	-13.2969	provide | basis
0.538462	-13.2969	approaches | e.g.
0.266667	-13.297	results | several
0.206186	-13.297	vector, | the
0.0875	-13.297	concept | the
0.0990783	-13.297	4 | 1
0.106667	-13.297	important | that
0.666667	-13.2971	a | assistant
0.304348	-13.2971	then | output
0.177778	-13.2971	clear | in
0.258065	-13.2971	can | instances
0.114551	-13.2972	main | is | of
0.258065	-13.2973	event | protein
0.5	-13.2973	3.3 | algorithm
0.167785	-13.2974	it | much
0.242857	-13.2974	were | could
0.152174	-13.2974	this | way
0.342857	-13.2975	ongoing | in
0.666667	-13.2976	2010; | 2011),
0.223529	-13.2976	a | sometimes
0.317073	-13.2976	(2010) | more
0.283019	-13.2976	nine | and
0.333333	-13.2978	< | at
0.583333	-13.2979	2003; | 2007)
0.116129	-13.298	f | and
0.45	-13.298	once | its
0.283019	-13.298	secondly, | in
0.153005	-13.2981	for | range
0.333333	-13.2981	text | news
0.473684	-13.2981	metrics | mt
0.291667	-13.2981	a | ie
0.5	-13.2981	asbestos | and
0.160494	-13.2982	obvious | the
0.366667	-13.2982	by | experts
0.833333	-13.2982	relations, | general
0.0856164	-13.2984	all | which
0.583333	-13.2984	represents | words.
0.304348	-13.2985	internet | and
0.262295	-13.2985	only | specific
0.5	-13.2985	noticeably | the
0.5	-13.2985	identifying | pairs
0.538462	-13.2986	meaning | meaning.
0.666667	-13.2987	inference | tractable
0.666667	-13.2987	texts | get
0.0829268	-13.2987	given | as
0.583333	-13.2988	computationally | large
0.325	-13.2988	of | reaction
0.714286	-13.2989	phrases | keywords
0.714286	-13.2989	ratinov | (2009)
0.714286	-13.2989	c(a) | c(b)
0.714286	-13.2989	curve | 2,
0.714286	-13.2989	|t|
0.333333	-13.2989	synset | wordnet
0.210526	-13.299	if | ∈
0.277778	-13.2991	alignment | other
0.666667	-13.2992	via | 2008).
0.285714	-13.2993	simple | at
0.179688	-13.2993	that | close
0.304348	-13.2994	pos | tag
0.4	-13.2994	more | i.e.
0.45	-13.2994	file | one
0.105541	-13.2995	features | each
0.714286	-13.2995	units, | units
0.833333	-13.2995	by | researchers,
0.833333	-13.2995	uas | las
0.0893536	-13.2996	of | relevant
0.666667	-13.2997	lower | overall
0.272727	-13.2997	repetition | the
0.171429	-13.2997	possible | each
0.583333	-13.2997	del | de
0.196262	-13.2997	parser | which
0.239437	-13.2998	is | closest
0.538462	-13.2999	differences | systems.
0.714286	-13.3	(2007) | technique
0.37931	-13.3	in | matching.
0.253968	-13.3	divided | two
0.304348	-13.3001	science | engineering
0.210526	-13.3001	these | human
0.666667	-13.3002	aim | compare
0.666667	-13.3002	k | 10.
0.833333	-13.3002	availability | tools
0.473684	-13.3003	little | this
0.5	-13.3004	by | tagger,
0.297872	-13.3005	is | successful
0.154286	-13.3005	for | categories
0.109551	-13.3006	name | and
0.714286	-13.3006	known | brands
0.714286	-13.3006	sent | john
0.714286	-13.3006	neutral | direction
0.714286	-13.3006	based on | observations,
0.267857	-13.3006	for | roughly
0.533333	-13.3007	a | extractor.
0.107735	-13.3008	measures | to
0.262295	-13.3008	and | value.
0.185484	-13.3008	a | (figure
0.285714	-13.301	section 2 | related
0.297872	-13.301	however, | languages
0.583333	-13.3011	match | terms
0.583333	-13.3011	a. | m.
0.416667	-13.3012	each | npk
0.416667	-13.3012	and | identical.
0.4	-13.3012	revealed | for
0.352941	-13.3014	of | tools,
0.0489589	-13.3014	problem | the
0.108033	-13.3014	who | to
0.833333	-13.3015	beliefs | rule
0.156977	-13.3016	each | null
0.262295	-13.3016	a | said
0.583333	-13.3016	keyphrases | topic
0.5	-13.3016	because | (e.g.
0.5	-13.3016	formal | but
0.0707547	-13.3017	if | we
0.416667	-13.3018	test | significance
0.473684	-13.3018	distributions | data
0.230769	-13.3018	the | varied
0.22093	-13.302	model, | each
0.533333	-13.302	following | features:
0.333333	-13.302	50 | all
0.198113	-13.302	the | charniak
0.538462	-13.302	displays | by
0.833333	-13.3021	i | it.
0.146465	-13.3022	and | disambiguation
0.4	-13.3022	constituent | an
0.172662	-13.3023	typical | and
0.242857	-13.3024	to | decrease
0.297872	-13.3024	precision | f-measure
0.0882353	-13.3024	grammar | a
0.583333	-13.3025	triggering | from
0.317073	-13.3025	use | joint
0.170068	-13.3026	information | semantic
0.0382363	-13.3027	is | than
0.304348	-13.3027	requirement | a
0.666667	-13.3028	n-gram | estimated
0.714286	-13.3028	languages | non-translated
0.533333	-13.3028	(b3) | (ceaf)
0.533333	-13.3028	paris | paris,
0.0851789	-13.3029	the | relatively
0.111111	-13.3029	the | tools
0.2	-13.3029	accuracies | of
0.283019	-13.3029	sentence | relation
0.428571	-13.303	annotated | data,
0.227848	-13.3031	pos tagging | and
0.142857	-13.3031	of | intended
0.0410522	-13.3031	different | and
0.208333	-13.3031	0 | accuracy
0.131148	-13.3031	is | linear
0.5	-13.3031	statistics | corpora
0.128906	-13.3032	global | of
0.146465	-13.3032	parts | and
0.538462	-13.3032	binary | vs.
0.366667	-13.3033	classes | using
0.184	-13.3034	we | access
0.833333	-13.3034	experiments | 5).
0.833333	-13.3034	which | graphic
0.416667	-13.3035	use | statistics
0.218391	-13.3035	in | em
0.253968	-13.3035	evaluation | reference
0.45	-13.3036	instance | both
0.309524	-13.3036	framework | semantic
0.309524	-13.3036	words | (e.g.
0.138889	-13.3037	task | or
0.538462	-13.3037	sentiment | thesaurus
0.538462	-13.3037	because | fixed
0.416667	-13.3038	from | ∼
0.0977273	-13.3038	the | alternative
0.172662	-13.3038	difference in | of
0.12069	-13.3039	each | other
0.134199	-13.304	is | low
0.5	-13.3043	as | units.
0.666667	-13.3044	cox | on
0.666667	-13.3044	em | local
0.714286	-13.3045	0.83 | -0.00
0.45	-13.3047	foundation | program
0.833333	-13.3047	its | hypernyms
0.277778	-13.3049	of | (where
0.253968	-13.3051	promising | for
0.0615522	-13.3052	the | discourse
0.583333	-13.3052	dataset | constructed
0.583333	-13.3052	derivations | same
0.317073	-13.3052	specifying | of
0.666667	-13.3054	work | involve
0.267857	-13.3054	null | nullnullnullnull
0.0461153	-13.3054	is | where | the
0.352941	-13.3055	next | how
0.285714	-13.3055	some | found
0.201923	-13.3055	is | semantically
0.5	-13.3055	which | desired
0.583333	-13.3057	information | central
0.583333	-13.3057	depth | wordnet
0.122302	-13.3059	introduced | and
0.297872	-13.3059	on | reduced
0.143564	-13.306	largest | the
0.175182	-13.306	beginning | of
0.147368	-13.3061	because | can
0.714286	-13.3062	results | promising.
0.184	-13.3062	preposition | a
0.0687711	-13.3064	show | in
0.107735	-13.3064	for | experiments
0.428571	-13.3064	several | related
0.09375	-13.3064	difficult | the
0.10951	-13.3064	p | r
0.666667	-13.3065	uses | distortion
0.4	-13.3066	or | having
0.416667	-13.3067	use | vocabulary
0.140845	-13.3067	computing | and
0.833333	-13.3067	tokenizer | into
0.354839	-13.3068	report | standard
0.352941	-13.3069	systems | system.
0.127413	-13.307	variables | and
0.114198	-13.307	in | translated
0.236111	-13.3071	closure | the
0.45	-13.3072	outperform | data
0.201923	-13.3073	the | xml
0.666667	-13.3075	x. | j.
0.666667	-13.3075	are | metaphors
0.246377	-13.3075	” | the
0.583333	-13.3075	contexts | event
0.583333	-13.3075	analysis, | only
0.45	-13.3075	• | last
0.0972851	-13.3076	includes | of
0.233766	-13.3076	results | combination
0.352941	-13.3077	algorithm | decoding
0.104439	-13.3077	large | for
0.297872	-13.3077	overall, | that
0.168919	-13.3077	to | (and
0.5	-13.3078	output | matches
0.714286	-13.3079	α(s0,|w|)β(s0,|w|)
0.181102	-13.3079	datasets | in
0.352941	-13.308	to | number,
0.285714	-13.308	learn | language
0.125954	-13.308	syntactic | are
0.0543933	-13.3082	the | selected | the
0.25	-13.3083	for | configuration
0.262295	-13.3084	candidate | sentence
0.123636	-13.3084	a | effect
0.45	-13.3086	few | many
0.45	-13.3086	gives | better
0.833333	-13.3086	“gold | in
0.833333	-13.3086	andrés | and
0.5	-13.3086	disjunctive | in
0.137615	-13.3087	the | removing
0.428571	-13.3087	but | easily
0.100719	-13.3087	then | from
0.538462	-13.3087	newswire | news
0.236111	-13.3088	as | above.
0.4	-13.3088	similar | found
0.131687	-13.3089	in | collection
0.666667	-13.3091	success | methods
0.118644	-13.3091	and | tag
0.0619872	-13.3092	words | as
0.0764951	-13.3092	the | concept
0.272727	-13.3093	some | those
0.366667	-13.3094	raw | raw
0.666667	-13.3096	representing | documents
0.538462	-13.3096	n | lists
0.184874	-13.3097	wish to | the
0.0759669	-13.3098	uses | of
0.37931	-13.3098	for | digital
0.366667	-13.31	do | add
0.538462	-13.31	that | untrained
0.714286	-13.3101	words | substituted
0.236111	-13.3103	we | test data
0.213483	-13.3103	4 | three
0.10951	-13.3103	the | element
0.22093	-13.3104	to | opinions
0.111111	-13.3105	is | within
0.470588	-13.3106	annotation | utterance
0.5	-13.3106	et | 2009a).
0.1625	-13.3107	10 | with
0.583333	-13.3107	tree, | only
0.262295	-13.3108	use | terms
0.178295	-13.3108	i | this
0.218391	-13.3109	concepts | by
0.470588	-13.3109	our | keyphrases
0.107735	-13.3109	identifying | of
0.285714	-13.3111	in | recognition.
0.285714	-13.3111	contains | information
0.242857	-13.3111	concerns | of
0.714286	-13.3112	into | slot
0.262295	-13.3112	the | (since
0.354839	-13.3112	{ | =
0.146597	-13.3113	observed | is
0.470588	-13.3113	targets | extracted
0.833333	-13.3113	objects | visual
0.0373677	-13.3113	language | and
0.0927835	-13.3114	for | parsing
0.333333	-13.3114	for | ).
0.0929752	-13.3114	the | states
0.5	-13.3114	indicators | by
0.354839	-13.3115	my | that
0.317073	-13.3115	ccg | for
0.168919	-13.3115	are | directly
0.246377	-13.3116	chunk | to
0.152542	-13.3117	for | inference
0.258065	-13.3117	preliminary | on
0.253968	-13.3117	to | boundaries.
0.0631579	-13.3118	has | on
0.188034	-13.3118	then | then
0.135965	-13.3119	for | total
0.225	-13.312	every | or
0.168919	-13.312	to | annotation.
0.110465	-13.3121	after | is
0.583333	-13.3121	that | optimal.
0.538462	-13.3121	task | attributes
0.666667	-13.3122	it | preferable
0.666667	-13.3122	given | l,
0.302326	-13.3123	for | length.
0.416667	-13.3123	order | ensure
0.416667	-13.3123	weighted | features
0.354839	-13.3123	sent | 0.60
0.0464286	-13.3124	used to | the
0.342857	-13.3125	the | further.
0.291667	-13.3125	found | performance
0.416667	-13.3126	and | machines.
0.120996	-13.3126	in | i.e.
0.267857	-13.3126	like | other
0.325	-13.3126	translation | aligned
0.5	-13.3126	5-gram | language
0.272727	-13.3127	the | appears.
0.714286	-13.3129	changes | made
0.117264	-13.3129	from | then
0.112121	-13.313	this | problem
0.366667	-13.3131	below | an
0.666667	-13.3133	study | design
0.258065	-13.3133	given | class
0.0251506	-13.3134	a | is | a
0.242857	-13.3134	in | framenet
0.384615	-13.3134	measures | both
0.538462	-13.3134	impression | computed
0.45	-13.3135	n-gram | probabilities
0.167785	-13.3136	structured | in
0.0782866	-13.3136	the | structures
0.181102	-13.3136	increased | of
0.309524	-13.3136	primarily | on
0.176923	-13.3138	for | spanish
0.140777	-13.3139	words | were
0.150838	-13.3139	systems | or
0.384615	-13.314	some | approaches
0.191304	-13.3141	the | annotator.
0.131915	-13.3142	construct | and
0.135965	-13.3143	twitter | and
0.131915	-13.3145	to | people
0.236111	-13.3146	segmented | a
0.0812102	-13.3147	of | after
0.583333	-13.3148	propose | scheme
0.416667	-13.3149	constructional | in
0.208333	-13.3149	subjects | a
0.225	-13.315	as | dictionary
0.083612	-13.315	the | column
0.184874	-13.315	theoretical | of
0.342857	-13.315	current | new
0.227848	-13.3151	the | are marked
0.242857	-13.3151	are | rules.
0.366667	-13.3152	of | one’s
0.45	-13.3153	extraction | performed
0.666667	-13.3154	eliminate | pairs
0.470588	-13.3154	tag | label
0.191304	-13.3154	two | –
0.538462	-13.3155	summation | all
0.333333	-13.3156	functions | or
0.714286	-13.3157	lack | consensus
0.239437	-13.3158	connective | a
0.0637066	-13.3158	but | for
0.470588	-13.3158	generate | sequence
0.140777	-13.3158	the | are | words.
0.20202	-13.3158	al | in
0.0954447	-13.3158	with | information
0.666667	-13.3159	figure 2: | model.
0.164474	-13.316	performance | two
0.206186	-13.3161	in | expert
0.0689266	-13.3161	to | large
0.100737	-13.3161	the | attribute
0.146597	-13.3162	for | joint
0.198113	-13.3162	web | are
0.416667	-13.3163	that | ignore
0.714286	-13.3163	natural | (nl)
0.714286	-13.3163	however, | imply
0.714286	-13.3163	simplified | strategy
0.0195295	-13.3163	are | to
0.18254	-13.3163	of | relatedness
0.130081	-13.3163	designed | and
0.267857	-13.3168	improvements | using
0.45	-13.3168	task | among
0.714286	-13.3169	|s
0.242857	-13.3169	are | semantically
0.297872	-13.3169	two | containing
0.666667	-13.317	gain | hypotheses
0.317073	-13.317	acoustic | with
0.45	-13.3171	report | when
0.106849	-13.3171	is | current
0.236111	-13.3172	data | provided
0.583333	-13.3172	text | identified
0.833333	-13.3172	coreference resolution | mention
0.416667	-13.3173	ratio | 1
0.714286	-13.3174	technology | dallas
0.0760719	-13.3174	a | c
0.100244	-13.3174	for | class
0.151685	-13.3175	then | a | that
0.297872	-13.3176	side | rule
0.4	-13.3176	incorporated | as
0.0792227	-13.3176	second | is
0.149733	-13.3176	each | also
0.285714	-13.3177	(2005), | and
0.137615	-13.3178	guidelines | the
0.236111	-13.3179	every | it
0.416667	-13.3179	haitian | for
0.4	-13.3179	with | tutor
0.833333	-13.3179	aside | use
0.714286	-13.318	complexity | spoken
0.145078	-13.3181	extend | and
0.583333	-13.3181	contain | often
0.538462	-13.3181	lexical, | features
0.143564	-13.3182	such | semantic
0.053682	-13.3182	2 | a
0.354839	-13.3182	since | directly
0.0934183	-13.3183	to | languages
0.277778	-13.3183	seemed | the
0.179688	-13.3184	proposed | based
0.416667	-13.3186	is | derivation.
0.416667	-13.3186	classification | 2009).
0.108262	-13.3186	verbs | in
0.242857	-13.3188	since | syntactic
0.428571	-13.3189	in | domingos,
0.538462	-13.3189	our | extensions
0.242857	-13.319	easier | and
0.120141	-13.319	than | other
0.324324	-13.319	as | example.
0.324324	-13.319	methods | very
0.6	-13.319	colour | distractor
0.088015	-13.319	and | 2010;
0.22093	-13.3191	the | target-side
0.060262	-13.3192	nodes | the
0.139535	-13.3193	and | improved
0.239437	-13.3194	google | for
0.136986	-13.3194	on | topic
0.263158	-13.3194	convergence | in
0.263158	-13.3194	a | diverse
0.6	-13.3195	other | aac
0.6	-13.3195	bicycles | road
0.133621	-13.3196	figure 1 | and
0.666667	-13.3196	α | 1,
0.416667	-13.3196	context | content
0.428571	-13.3196	to | (2003).
0.213483	-13.3197	because | language
0.135747	-13.3198	two | of | two
0.833333	-13.3198	project | developing
0.833333	-13.3198	verbs | direct
0.218391	-13.3199	however, | both
0.272727	-13.3199	despite | with
0.45	-13.32	output | found
0.666667	-13.3202	increase | decoding
0.317073	-13.3202	have | 2005).
0.37931	-13.3203	particular | context
0.297872	-13.3204	trees | syntactic
0.366667	-13.3204	system | stanford
0.384615	-13.3204	while | annotations
0.129555	-13.3205	for example, | for | and
0.25	-13.3205	… | of
0.172662	-13.3205	an | annotation
0.666667	-13.3207	usually | problem
0.213483	-13.3209	and | versus
0.194444	-13.3209	composition | and
0.6	-13.3209	instance | (e.g.
0.6	-13.3209	improved | models.
0.45	-13.321	set | modifiers
0.45	-13.321	order | calculate
0.0418936	-13.321	second | the
0.538462	-13.3211	overall, | better
0.138889	-13.3212	structured | and
0.333333	-13.3212	a | phrasebased
0.148148	-13.3212	we | 20
0.106557	-13.3213	a | question
0.186441	-13.3213	0 | 30 | 50
0.5	-13.3213	these | annotations.
0.714286	-13.3214	log-linear | various
0.6	-13.3214	normal: | a
0.0236655	-13.3214	and | on
0.538462	-13.3215	association | associations
0.213483	-13.3216	below, | to
0.206186	-13.3217	table | gives
0.333333	-13.3217	dictionaries | by
0.175573	-13.3218	and | center
0.0571652	-13.3218	of | context
0.165563	-13.3219	trained | data.
0.047619	-13.3219	of | models
0.6	-13.3219	access | user
0.538462	-13.3219	make | just
0.538462	-13.3219	for | conflated
0.714286	-13.322	arabic | msa
0.302326	-13.3221	provided | shared
0.25	-13.3223	contained | or
0.181102	-13.3224	100% | the
0.0833333	-13.3224	a | analysis
0.258065	-13.3225	but | general
0.272727	-13.3227	of | entity,
0.208333	-13.3227	it | observed
0.666667	-13.3228	in | cause.
0.354839	-13.3228	we | predictors
0.470588	-13.3229	indeed | which
0.232877	-13.3229	with | e.g.,
0.146597	-13.323	our | parser
0.225	-13.3231	extract | each
0.714286	-13.3231	run | models.
0.384615	-13.3231	are | strictly
0.309524	-13.3231	of | influences
0.176923	-13.3231	the | expressing
0.20202	-13.3232	may | when
0.45	-13.3232	these | link
0.45	-13.3232	training | classifier,
0.833333	-13.3232	regularization | objective
0.302326	-13.3233	when | pair
0.227848	-13.3233	where | value
0.102041	-13.3233	we will | a
0.666667	-13.3234	particular | (or
0.666667	-13.3234	operator | tree
0.0961969	-13.3235	and | phrases
0.0806962	-13.3235	and | extraction
0.4	-13.3236	links | wikipedia
0.354839	-13.3236	this | knowledge.
0.291667	-13.3236	lm | stream
0.253968	-13.3236	obvious | in
0.714286	-13.3237	remark | a
0.714286	-13.3237	single-beam | multiple-beam
0.0992908	-13.3237	first, | a
0.277778	-13.3238	rules | english
0.291667	-13.3238	0.1 | 0.2
0.428571	-13.324	model | adjacent
0.0886792	-13.324	the | disambiguation
0.181102	-13.3241	features | improve
0.302326	-13.3243	bleu scores | for
0.428571	-13.3243	(1) | possible
0.45	-13.3243	explore | performance
0.189655	-13.3243	it | common
0.333333	-13.3244	and | k,
0.309524	-13.3244	(2011) | for
0.833333	-13.3245	employ | strategy
0.354839	-13.3247	with | constant
0.384615	-13.3247	trie | a
0.6	-13.3247	cue | patterns
0.6	-13.3247	are | heads.
0.666667	-13.3249	work | collaboration
0.0801233	-13.325	the | amount of
0.297872	-13.325	model | languages.
0.215909	-13.325	topics | each
0.246154	-13.325	randomly | training
0.324324	-13.325	supported | projects
0.104558	-13.325	the | allowing
0.154286	-13.325	it | lexical
0.470588	-13.3251	annotated | labels
0.37931	-13.3251	 | 
0.366667	-13.3251	hypotheses | an
0.309524	-13.3251	this | typical
0.232877	-13.3251	counts | from
0.324324	-13.3252	returns | for
0.6	-13.3252	tokens | appear
0.333333	-13.3254	(2006) | al.
0.0551876	-13.3254	et | 2010).
0.12749	-13.3254	a | i.e.,
0.538462	-13.3254	optimization | weights
0.302326	-13.3255	to | systematic
0.666667	-13.3255	(2008b) | a
0.175573	-13.3255	by | linguistic
0.291667	-13.3255	investigating | to
0.0583882	-13.3256	for | where
0.28	-13.3257	string | from
0.6	-13.3257	were | heuristics
0.5	-13.3257	the | backchannels.
0.15	-13.3258	the | chain
0.285714	-13.3259	most | words.
0.285714	-13.3259	using | 2005).
0.666667	-13.326	domain | effective
0.196262	-13.326	method | information
0.0868762	-13.3261	we use | to | and
0.317073	-13.3261	performs | using
0.5	-13.3261	learning | behavior
0.302326	-13.3262	therefore, | number
0.6	-13.3262	since | regular
0.239437	-13.3263	of | discovered
0.20202	-13.3263	used | evaluate
0.2	-13.3263	we | update
0.666667	-13.3265	entities | arguments
0.0680804	-13.3265	is | that | not
0.714286	-13.3265	match | reference.
0.263158	-13.3265	value | 0
0.833333	-13.3265	trbm | isbn
0.833333	-13.3265	al-haj | wintner
0.833333	-13.3265	hits | (kleinberg,
0.833333	-13.3265	cluster-ranking | rahman
0.833333	-13.3265	mid | .2
0.833333	-13.3265	linear | (lp)
0.833333	-13.3265	94.7 | 96.4
0.833333	-13.3265	action-value | q(s,a)
0.833333	-13.3265	downspeak | upspeak.
0.833333	-13.3265	void | {
0.833333	-13.3265	folds | fold.
0.833333	-13.3265	earth | wind
0.833333	-13.3265	cu-bojar | pc-trans
0.833333	-13.3265	clm | bow
0.833333	-13.3265	duluth-1 | duluth-3
0.833333	-13.3265	(kate | mooney,
0.833333	-13.3265	e., | vouk,
0.833333	-13.3265	lp-mert | 1d-mert
0.833333	-13.3265	ponzetto | strube,
0.833333	-13.3265	p(e1) | p(e2)
0.833333	-13.3265	operator | re
0.272727	-13.3266	but | correct
0.0763791	-13.3266	an | word
0.134529	-13.3267	account | in
0.151685	-13.3267	of | interpretation
0.0847751	-13.3268	step | to
0.2	-13.3268	rules | source
0.120996	-13.3269	similar | with
0.470588	-13.327	plan | further
0.204082	-13.327	confirms | the
0.384615	-13.3271	quality | reference
0.309524	-13.3271	in | boundaries.
0.232877	-13.3271	don’t | to
0.20202	-13.3273	15 | in
0.583333	-13.3274	methods, | over
0.428571	-13.3274	and | (eq.
0.428571	-13.3274	systems | 2003)
0.139423	-13.3274	to | recognition
0.538462	-13.3275	rest | paper,
0.11745	-13.3275	for | top
0.366667	-13.3277	corpus | specific
0.258065	-13.3277	a | mass
0.0826446	-13.3279	words | or
0.0270175	-13.328	the | translation
0.183333	-13.328	using | bleu
0.4	-13.3281	on | 25
0.181818	-13.3281	not | significant
0.184874	-13.3282	non-terminal | the
0.583333	-13.3283	avoids | by
0.302326	-13.3284	a | head.
0.239437	-13.3284	central | and
0.333333	-13.3286	a | combination.
0.309524	-13.3286	our | design
0.6	-13.3286	there | mismatch
0.6	-13.3286	negation | polarity
0.157576	-13.3286	the | newly
0.666667	-13.3287	action | take
0.666667	-13.3287	aim | produce
0.219512	-13.3287	and | determining
0.272727	-13.3288	which | existing
0.149171	-13.3289	when | words
0.354839	-13.329	mentions | their
0.101781	-13.329	are | while
0.156627	-13.3291	the | alone
0.666667	-13.3292	first | redundant
0.272727	-13.3292	is | here.
0.538462	-13.3292	α | β,
0.342857	-13.3292	begins | and
0.384615	-13.3293	word | substitution
0.204082	-13.3293	entire | is
0.5	-13.3293	mae | for
0.225	-13.3294	is | easier
0.714286	-13.3294	mt08 | baseline
0.4	-13.3294	histograms | for
0.0768137	-13.3295	whether | a
0.428571	-13.3295	3.1 | documents
0.196262	-13.3295	each | 5
0.196262	-13.3295	of | describing
0.6	-13.3295	without | correct
0.166667	-13.3296	that | correctly
0.317073	-13.3296	most | least
0.25	-13.3297	is | structure,
0.5	-13.3297	and | understanding.
0.470588	-13.33	joint | other
0.470588	-13.33	investigated | features
0.213483	-13.33	adds | and
0.196262	-13.33	representation | by
0.583333	-13.3302	case-insensitive | (papineni
0.188034	-13.3302	the | laboratory
0.184874	-13.3303	the | biological
0.470588	-13.3304	new | general
0.242857	-13.3304	(2008), | and
0.354839	-13.3304	however, | useful
0.45	-13.3304	optimization | objective
0.135135	-13.3305	v | =
0.384615	-13.3305	simply | their
0.5	-13.3305	baseline | learned
0.219512	-13.3306	(2008) | on
0.4	-13.3307	comparing | or
0.190909	-13.3308	experiment, | a
0.666667	-13.3308	agreement | transcribers
0.45	-13.3308	generative | model.
0.45	-13.3308	bridge | between
0.0292349	-13.3309	a | which | the
0.428571	-13.3309	are | summaries.
0.263158	-13.3309	the | valence
0.204082	-13.3309	significantly | from
0.0692488	-13.3309	effect | the
0.20202	-13.331	analyses | a
0.6	-13.331	candidate | list.
0.186441	-13.331	generalized | of
0.155689	-13.3312	overview | of
0.178295	-13.3312	demonstrate | in
0.28	-13.3313	each | latent
0.107042	-13.3313	in | frequency
0.666667	-13.3314	for | bits.
0.263158	-13.3314	proposed | system
0.151685	-13.3314	adaptation | to
0.366667	-13.3315	adaptation | translation
0.6	-13.3315	mappings | sequences
0.714286	-13.3317	messages | chat
0.714286	-13.3317	because | know | are
0.0421687	-13.3317	corpus | and
0.285714	-13.3318	mckeown, | al.,
0.45	-13.3318	our | ten
0.109467	-13.3318	the | scale
0.291667	-13.3318	different | features.
0.470588	-13.3319	by | permutation
0.1125	-13.3319	of | technology
0.107955	-13.3319	its | are
0.16129	-13.3319	of | parsing.
0.267857	-13.332	using | metrics
0.366667	-13.3321	since | common
0.714286	-13.3323	users, | users
0.666667	-13.3324	k-nearest | graph
0.0884615	-13.3324	this | first
0.6	-13.3324	competitive | which
0.239437	-13.3326	for | book
0.222222	-13.3326	the | attributes,
0.113565	-13.3326	n-gram | in
0.139423	-13.3326	to | relatively
0.538462	-13.3327	if | discarded.
0.0529891	-13.3328	described | the
0.0969977	-13.3329	definition | of
0.130802	-13.3329	finding | a
0.4	-13.3329	patterns | instances
0.428571	-13.3329	way | possible
0.0821018	-13.3332	moreover, | the
0.317073	-13.3332	perceptron | with
0.0521452	-13.3332	this paper | the
0.0815661	-13.3332	set | which
0.470588	-13.3334	performance, | than
0.152542	-13.3334	in | textual
0.253968	-13.3334	graph | using
0.583333	-13.3335	help | how
0.0922432	-13.3335	of | tasks
0.324324	-13.3336	but | show
0.384615	-13.3336	and | summaries,
0.538462	-13.3336	usually, | are
0.181818	-13.3336	ranking | as
0.190909	-13.3337	supervised | with
0.222222	-13.3337	distinguishing | and
0.0510397	-13.3338	text | and
0.583333	-13.3339	extractive | summarization.
0.384615	-13.3339	a | attach
0.342857	-13.3339	is | primarily
0.666667	-13.334	occur | once
0.128	-13.334	statistical | for
0.0563272	-13.3343	the | phrases
0.342857	-13.3344	cluster | its
0.239437	-13.3345	the | reference,
0.714286	-13.3346	go | event
0.285714	-13.3348	on | english.
0.0774854	-13.3348	with | one
0.140777	-13.3348	that | achieve
0.219512	-13.3348	of | classifier,
0.583333	-13.3349	lmbot | finite
0.166667	-13.335	this | relevant
0.20202	-13.335	methods | word
0.428571	-13.335	low | score
0.666667	-13.3351	phrases | contiguous
0.141463	-13.3353	the | mwes
0.183333	-13.3353	for | definition
0.117057	-13.3354	point | and
0.153846	-13.3354	proper | of
0.0791476	-13.3354	this | where
0.258621	-13.3355	when | parser
0.0694118	-13.3356	probabilities | the
0.156627	-13.3357	the | of | feature.
0.583333	-13.3358	this | lexemes
0.150838	-13.336	the | received
0.384615	-13.3361	which | fit
0.5	-13.3361	roman | and
0.302326	-13.3362	limited | training
0.272727	-13.3362	word | seen
0.109467	-13.3362	we | frequency
0.714286	-13.3363	product | web
0.714286	-13.3363	translations, | human
0.714286	-13.3363	formulation | problem.
0.583333	-13.3363	relations | patterns.
0.0267222	-13.3363	we | from
0.156627	-13.3364	feedback | and
0.0784314	-13.3365	in | very
0.5	-13.3365	their | relations,
0.227848	-13.3366	method | learning
0.302326	-13.3367	of | objects,
0.666667	-13.3367	that | seldom
0.0808241	-13.3368	of | (see
0.285714	-13.3369	annotation | events
0.34375	-13.3371	most | way
0.211111	-13.3372	iterations, | the
0.470588	-13.3372	chinese | 2005).
0.291667	-13.3372	a | prefer
0.666667	-13.3373	ontologies | domain
0.45	-13.3373	since | nouns
0.142857	-13.3373	where | at
0.6	-13.3373	is | tectogrammatical
0.6	-13.3373	16 | 23
0.181818	-13.3374	paradigm | the
0.354839	-13.3376	we | binomial
0.22973	-13.3377	a | decrease
0.6	-13.3377	(kipper | al.,
0.151685	-13.3377	3. | a
0.25	-13.3378	to | fairly
0.216867	-13.3379	while | models
0.324324	-13.338	axioms | as
0.354839	-13.3381	et | 2009b).
0.428571	-13.3381	noisy | data
0.295455	-13.3381	very | also
0.5	-13.3381	small | best
0.239437	-13.3382	in | 1999;
0.150838	-13.3382	only | their
0.666667	-13.3383	corrects | a
0.131915	-13.3383	require | and
0.147541	-13.3384	= | v
0.160256	-13.3385	has | some
0.28	-13.3386	modifier | is
0.232877	-13.3388	using | instead
0.181818	-13.3388	examples | training
0.366667	-13.3389	allows | user
0.366667	-13.3389	the | perspective.
0.128514	-13.339	of | german
0.470588	-13.3391	2008) | 2009)
0.102041	-13.3391	remaining | the
0.167832	-13.3391	model | any
0.0311394	-13.3391	these | of
0.147541	-13.3391	the | american
0.227848	-13.3392	from | who
0.109467	-13.3392	with | new
0.106742	-13.3392	four | in
0.157576	-13.3392	scores | all
0.167832	-13.3393	different | were
0.105121	-13.3393	method | et
0.45	-13.3395	we | as implemented
0.333333	-13.3397	models | simple
0.14433	-13.3398	a | finite
0.239437	-13.3398	color | the
0.0829103	-13.3398	a | model, | the
0.45	-13.3398	linguistics | uppsala
0.384615	-13.3398	argument | classification
0.5	-13.3398	3 | questions
0.181818	-13.3398	deletion | the
0.208791	-13.3399	found | all
0.165517	-13.3401	method | used
0.309524	-13.3401	zhao | al.,
0.470588	-13.3403	by | authors,
0.324324	-13.3404	large | size
0.666667	-13.3405	under | experimental
0.666667	-13.3405	its | capturing
0.163399	-13.3406	we compute | and
0.428571	-13.3406	challenge, | and
0.428571	-13.3406	scenes | target
0.309524	-13.3406	and | interpretations
0.103175	-13.3406	recognition | and
0.0868762	-13.3407	the | improved
0.222222	-13.3407	using | available
0.232877	-13.3407	from | real
0.159236	-13.3409	and | computing
0.538462	-13.341	one | sometimes
0.538462	-13.341	summaries | specific
0.227848	-13.3411	local | features
0.100251	-13.3411	existing | and
0.216867	-13.3412	these | need
0.213483	-13.3412	vary | and
0.291667	-13.3412	2002; | 2005;
0.272727	-13.3413	the | shifts
0.366667	-13.3413	training | classifiers
0.146597	-13.3414	which | lexical
0.714286	-13.3415	users | topic.
0.169014	-13.3415	according to | with
0.174242	-13.3415	store | the
0.666667	-13.3416	features | chiang
0.151685	-13.3416	are | particular
0.0903614	-13.3418	the | characters
0.2	-13.3418	online | for
0.157576	-13.3418	transitions | the
0.101781	-13.3419	the | 1).
0.538462	-13.3419	automatically | articles
0.056294	-13.342	in | over
0.0614525	-13.342	the | set,
0.384615	-13.342	reference | translations
0.384615	-13.342	to | revision
0.302326	-13.3421	mean | score
0.666667	-13.3421	christodoulopoulos | al.
0.0907258	-13.3421	1, | and
0.6	-13.3421	asked | summaries
0.236111	-13.3422	the | became
0.285714	-13.3422	for | plus
0.333333	-13.3424	baldwin, | and
0.181818	-13.3424	case | or
0.324324	-13.3425	collective | a
0.242424	-13.3426	of | ir
0.470588	-13.3426	approach | generating
0.052	-13.3426	because | of
0.666667	-13.3427	several | ie
0.140777	-13.3427	of | derivation
0.538462	-13.3427	always | majority
0.0237832	-13.3428	a | on
0.165517	-13.3428	and | surface
0.45	-13.3428	for | inference,
0.0581395	-13.3428	the | set, | the
0.183333	-13.3428	used, | and
0.384615	-13.3429	gold-standard | are
0.0990338	-13.343	is | following
0.324324	-13.343	of | communication.
0.102302	-13.343	best | on
0.0809061	-13.3431	graph | and
0.115132	-13.3431	of | evaluated
0.125984	-13.3431	selected | for
0.0164216	-13.3431	a | for
0.14433	-13.3432	a | usually
0.138756	-13.3432	the | reduces
0.34375	-13.3432	inter-annotator agreement | for
0.470588	-13.3433	also | little
0.583333	-13.3434	sets | (see
0.0820771	-13.3434	edge | the
0.246154	-13.3434	it | help
0.0833333	-13.3434	list | and
0.22973	-13.3436	indeed, | to
0.120879	-13.3436	already | and
0.291667	-13.3436	in | sentence, | are
0.470588	-13.3437	several | during
0.27451	-13.3439	work | generation
0.45	-13.3439	making | about
0.2	-13.3439	idea | by
0.4	-13.344	abstract | annotation
0.324324	-13.3441	then | selected
0.0859232	-13.3442	models | with
0.5	-13.3442	attachment | an
0.5	-13.3442	for | failure
0.138756	-13.3443	discussion | a
0.138756	-13.3443	to | whole
0.16129	-13.3444	finally, | by
0.285714	-13.3445	in | another.
0.258621	-13.3445	ensure | are
0.178862	-13.3445	we | experiments on
0.6	-13.3446	could | direct
0.5	-13.3446	zhu, | and
0.219512	-13.3446	the | action.
0.0969977	-13.3447	form | in
0.666667	-13.3448	mailing | to
0.272727	-13.3449	0.001 | 0.001
0.538462	-13.3449	ranker | all
0.538462	-13.3449	highly | high
0.125	-13.3449	true | of
0.285714	-13.345	corpus | consists of
0.133333	-13.3452	an | task
0.409091	-13.3452	ranged | and
0.216867	-13.3453	the | sentences).
0.216867	-13.3453	can | performance.
0.20202	-13.3454	us | our
0.302326	-13.3455	1 introduction | data
0.167832	-13.3455	accuracy | from
0.6	-13.3455	alignment | segmentations.
0.216867	-13.3456	the | digital
0.714286	-13.3456	trained | papers
0.4	-13.3456	find | relation
0.258621	-13.3456	tasks | which
0.27451	-13.3457	these | combined
0.45	-13.3457	m. | university
0.366667	-13.3458	add | between
0.109467	-13.3458	entropy | the
0.666667	-13.3459	so | look
0.0938865	-13.3459	the | tweets
0.6	-13.346	i’m | my
0.102362	-13.346	the | underlying
0.366667	-13.3461	may | make
0.714286	-13.3462	lack | fit
0.384615	-13.3464	has | considerable
0.192661	-13.3464	the | counting
0.409091	-13.3465	flexible | with
0.6	-13.3465	hence, | any
0.213483	-13.3466	used | parsing
0.246154	-13.3466	to | translations,
0.538462	-13.3467	texts | authors
0.538462	-13.3467	answer, | to
0.295455	-13.3467	all | utterances
0.189189	-13.3469	regression | for
0.208791	-13.3471	times, | the
0.101781	-13.3471	created | of
0.5	-13.3471	used | reverse
0.236111	-13.3472	using | classification
0.185841	-13.3472	inclusion | the
0.428571	-13.3472	approach | structured
0.159236	-13.3473	morphological | for
0.714286	-13.3473	therefore, | surface
0.714286	-13.3473	twitter | tweets.
0.714286	-13.3473	textrunner-r | and
0.315789	-13.3473	that | closest
0.114379	-13.3473	another | that
0.155689	-13.3474	for | position
0.10585	-13.3474	of | semantics
0.4	-13.3475	in | abduction
0.6	-13.3475	help | find
0.6	-13.3475	having | now
0.5	-13.3475	we now | two
0.183333	-13.3475	are | certain
0.141414	-13.3476	states | and
0.302326	-13.3477	orthographic | and
0.211111	-13.3477	w | c
0.222222	-13.3477	and | list.
0.315789	-13.3478	among | other
0.409091	-13.3478	human | (callison-burch
0.714286	-13.3479	during | dialog
0.4	-13.3479	each | d,
0.6	-13.348	under | framework.
0.6	-13.348	prediction | often
0.16129	-13.348	one | translation
0.666667	-13.3481	using | (em)
0.428571	-13.3482	interpretation | it
0.246154	-13.3482	in | constant
0.45	-13.3483	adapted | other
0.384615	-13.3483	lexical | dictionary
0.6	-13.3485	tend | often
0.106443	-13.3485	the | synset
0.172932	-13.3486	each | them
0.0668151	-13.3486	full | the
0.232877	-13.3486	and | nearly
0.285714	-13.3487	third, | and
0.324324	-13.3488	try | with
0.538462	-13.3489	provide | measure
0.0456176	-13.349	that | a | that
0.189189	-13.349	objects | and
0.22973	-13.3491	their | over
0.0603604	-13.3491	its | in
0.107042	-13.3491	models | the | model
0.132159	-13.3492	we | monolingual
0.151163	-13.3492	5 | 4
0.172932	-13.3493	the | lexicon,
0.0524881	-13.3493	also | that
0.120438	-13.3493	at | of | at
0.6	-13.3495	using | as:
0.152047	-13.3496	for | low
0.27451	-13.3496	2) | are
0.34375	-13.3497	systems | combination
0.104396	-13.3498	in | study
0.2	-13.3498	formally, | is
0.0307334	-13.3499	for each | the | the
0.0664488	-13.3499	it | or
0.28	-13.35	all | subset
0.181818	-13.35	that | mention
0.253968	-13.3501	by | group
0.28	-13.3502	results | under
0.714286	-13.3502	entire | corpus,
0.226667	-13.3502	in | others.
0.0703704	-13.3502	very | to
0.194175	-13.3502	while | do
0.263158	-13.3502	their | similar
0.20202	-13.3503	correspondence | and
0.253968	-13.3503	of | remain
0.309524	-13.3504	implementation | an
0.138756	-13.3504	then | this
0.37037	-13.3505	were | ones
0.194175	-13.3505	entity, | the
0.409091	-13.3505	this | exponential
0.470588	-13.3506	referring | target
0.0635721	-13.3507	distance | the
0.0617871	-13.3508	for | between | and
0.666667	-13.3508	only | m2
0.714286	-13.3508	terms | entailed
0.714286	-13.3508	a | 447
0.714286	-13.3508	full | text.
0.333333	-13.3508	ideal | for
0.34375	-13.3508	words | topic.
0.37037	-13.3509	tasks | their
0.165517	-13.3509	4 | system
0.6	-13.3509	solving | optimization
0.232877	-13.3509	should | be | as
0.180328	-13.351	our | dataset
0.428571	-13.351	training, | sets
0.185841	-13.3511	linked | of
0.155689	-13.3511	documents | from
0.128631	-13.3511	grammars | the
0.538462	-13.3511	being | specific
0.409091	-13.3512	n+1 | n
0.666667	-13.3514	dependency | (smith
0.37037	-13.3515	in | decision.
0.0569366	-13.3515	method | a
0.409091	-13.3515	all | beginning
0.09	-13.3515	here | and
0.45	-13.3516	moreover, | information
0.121771	-13.3516	the | covered
0.113636	-13.3517	whose | is | the
0.45	-13.3519	document | selected
0.333333	-13.3519	2 | overview
0.6	-13.3519	building | classifiers
0.714286	-13.352	university | university.
0.538462	-13.352	if | your
0.538462	-13.352	parser, | parse
0.19802	-13.3521	the | segmentations
0.470588	-13.3522	where | j.
0.333333	-13.3522	state | then
0.131004	-13.3522	classifier | as
0.104396	-13.3523	names | the
0.109145	-13.3523	approaches | al.,
0.15625	-13.3523	analysis | this
0.152941	-13.3524	can | about
0.28	-13.3525	the | expected.
0.0462029	-13.3525	is | one
0.714286	-13.3526	characters | characters,
0.714286	-13.3526	overlap | linguistic
0.216867	-13.3527	performance | three
0.181818	-13.3527	we | detect
0.104972	-13.3528	the | (a
0.167832	-13.3528	literature | of
0.354839	-13.3529	is | summarization.
0.6	-13.3529	approach, | terms
0.075	-13.353	represents | the
0.103175	-13.353	of | presented
0.19802	-13.3531	to | occurring
0.136792	-13.3531	to | queries
0.122677	-13.3531	a | means
0.27451	-13.3532	obtain | data
0.0926724	-13.3532	is | human
0.538462	-13.3533	heuristic | search
0.0940919	-13.3534	within | and
0.4	-13.3534	an | feature.
0.263158	-13.3534	optimization | with
0.0973872	-13.3536	of | initial
0.666667	-13.3536	interesting | here
0.0754986	-13.3536	section | section
0.232877	-13.3536	baseline, | a
0.5	-13.3536	neutral | positive
0.302326	-13.3537	we | def=
0.27451	-13.3538	unigram | features
0.20202	-13.3538	attempts to | of
0.116041	-13.3538	to | range
0.333333	-13.3538	however, | what
0.154762	-13.3539	simplified | the
0.409091	-13.3539	have | rules.
0.163265	-13.3541	classifiers | in
0.354839	-13.3541	whether | present
0.123552	-13.3541	via | and
0.0750708	-13.3541	of | tokens
0.45	-13.3542	θ | i=1
0.128099	-13.3543	then | which
0.22973	-13.3544	trend | of
0.714286	-13.3544	therefore, | techniques
0.354839	-13.3544	1990; | and
0.6	-13.3544	verbs | dependency
0.0784615	-13.3544	systems | a
0.216867	-13.3545	identifier | the
0.0854545	-13.3545	of | full
0.171642	-13.3545	would | such
0.104683	-13.3546	a | name
0.409091	-13.3546	was | resulting
0.4	-13.3548	our | (table
0.104972	-13.3549	that | related
0.154762	-13.3549	sentences | one
0.714286	-13.3549	however, | classifier.
0.113269	-13.3549	they | which
0.157233	-13.3549	system | best
0.159236	-13.355	an | rule
0.0969267	-13.355	and | development
0.302326	-13.3552	above | all
0.19802	-13.3552	sentiment | and | sentiment
0.5	-13.3553	content | automatic
0.37037	-13.3554	= | ...
0.4	-13.3554	stage | our
0.102094	-13.3555	of | empirical
0.166667	-13.3556	to | active
0.133929	-13.3556	the | outperform
0.226667	-13.3557	annotation | were
0.0833333	-13.3557	senses | the
0.666667	-13.3558	cross-language | language
0.0181759	-13.3558	are | and
0.163265	-13.3559	word | also
0.285714	-13.356	with | amounts
0.120879	-13.356	the | differ
0.263158	-13.3562	previously | that
0.0785824	-13.3562	measure | and
0.154762	-13.3564	ability | a
0.354839	-13.3564	fill | in
0.0362035	-13.3564	of | different
0.295455	-13.3564	3.1 | two
0.0381021	-13.3565	the | task | the
0.384615	-13.3565	documents | score
0.0872865	-13.3567	it is | with
0.204301	-13.3567	how | they
0.714286	-13.3567	following | idea
0.714286	-13.3567	presents | challenges
0.714286	-13.3567	two | d2
0.226667	-13.3567	the | jth
0.4	-13.3567	the | countries.
0.324324	-13.3568	true | which
0.37037	-13.3569	the | mistakes.
0.105263	-13.3569	had | to
0.108187	-13.357	all | features
0.288889	-13.357	genetic | and
0.242424	-13.3571	confusion | as
0.154762	-13.3572	systems | which
0.288889	-13.3572	the | (first
0.232877	-13.3572	dialogues | a
0.164384	-13.3573	earlier | of
0.5	-13.3573	partition | training
0.6	-13.3574	characters | present
0.181818	-13.3575	recent | has
0.22973	-13.3576	factors | are
0.267857	-13.3576	just | not
0.34375	-13.3576	a | meaning,
0.190909	-13.3577	the | bits
0.0815308	-13.3577	number | words
0.15625	-13.3577	a | separate
0.309524	-13.3578	granularity | in
0.714286	-13.3579	θ | value
0.324324	-13.3579	in addition, | model
0.145946	-13.3579	and | distributional
0.6	-13.3579	edge | null.
0.6	-13.3579	function | constant
0.6	-13.3579	ir | creative
0.22973	-13.358	any | only
0.121771	-13.358	pair, | the
0.0756063	-13.358	we | rules
0.27451	-13.3581	determiner | in
0.14359	-13.3581	estimate | in
0.366667	-13.3581	almost | or
0.366667	-13.3581	by | nouns,
0.315789	-13.3582	of | usa.
0.315789	-13.3582	as | encoding
0.25	-13.3582	we | stream
0.232877	-13.3582	systems | only
0.192308	-13.3583	utilized | the
0.194175	-13.3583	manually | from
0.166667	-13.3584	we | extracting
0.288889	-13.3587	of | coverage.
0.118705	-13.3587	a | wikipedia
0.131579	-13.3587	effective | and
0.226667	-13.3589	it | require
0.15625	-13.3589	texts | from
0.0926724	-13.359	sequences | the
0.37037	-13.3591	built | two
0.22973	-13.3593	this | procedure
0.0785824	-13.3593	et | 2009;
0.6	-13.3593	make | attempt
0.162162	-13.3594	german | in
0.0997506	-13.3594	of | who
0.0741758	-13.3594	linguistic | and
0.133929	-13.3594	the | chunk
0.666667	-13.3596	features | source-side
0.354839	-13.3596	single | while
0.208791	-13.3596	email | and
0.714286	-13.3597	towards | end
0.180328	-13.3597	basic | that
0.0417936	-13.3597	accuracy | the
0.27451	-13.3599	the | nature.
0.216867	-13.3599	short | for
0.263158	-13.3599	and | aid
0.37037	-13.36	use | map
0.181818	-13.36	we | actual
0.0916844	-13.3601	the | is,
0.145161	-13.3601	most | our
0.189189	-13.3602	co-occurrence | a
0.5	-13.3602	how | components
0.124514	-13.3603	algorithm | from
0.190909	-13.3604	a | for | sentence.
0.0903491	-13.3604	table | table
0.538462	-13.3604	task | monolingual
0.34375	-13.3604	modifiers | are
0.28	-13.3606	ideas | of
0.254237	-13.3606	although | their
0.0640496	-13.3606	our | this
0.0897959	-13.3606	the | to improve
0.131579	-13.3607	unlike | of
0.172932	-13.3608	(or | is
0.0150148	-13.3609	the | is | to
0.134259	-13.3609	the | development | test
0.258621	-13.3609	of | entries.
0.122137	-13.3609	for | general
0.142857	-13.3609	scores, | the
0.288889	-13.361	generates | with
0.27451	-13.3611	using | 2004).
0.101299	-13.3611	of | against
0.263158	-13.3612	several | at
0.116838	-13.3613	the | answers
0.089613	-13.3614	the | sample
0.714286	-13.3614	correct | thus
0.4	-13.3614	able | produce
0.4	-13.3614	morphemes | pos
0.140704	-13.3614	this | provide
0.37037	-13.3615	speech | spoken
0.288889	-13.3615	useful | information
0.152047	-13.3616	the | stop
0.315789	-13.3616	several | text
0.538462	-13.3617	only | constant
0.190909	-13.3618	the | anchor
0.114754	-13.3618	the | recent
0.6	-13.3618	abstract | up
0.22973	-13.3619	of | tutoring
0.5	-13.3619	comparison | approaches
0.27451	-13.362	via | from
0.714286	-13.362	adaptation | background
0.714286	-13.362	7th | grant
0.138095	-13.362	2.1 | a
0.409091	-13.362	as | mentioned,
0.315789	-13.3621	task | discourse
0.333333	-13.3623	syntactic | improve
0.236111	-13.3625	leverage | of
0.122137	-13.3625	extraction | as
0.384615	-13.3625	and | boundaries,
0.0964706	-13.3625	2) | the
0.0939597	-13.3626	of | pairs.
0.111821	-13.3627	the | emotion
0.5	-13.3627	it | normal
0.354839	-13.3629	other | techniques
0.0836237	-13.3629	important | to
0.139303	-13.363	over | each
0.470588	-13.3631	projection | target
0.2	-13.3631	set | k
0.538462	-13.3631	that, | obtained
0.714286	-13.3632	where | realized
0.226667	-13.3632	(at | the
0.135514	-13.3633	to | simply
0.409091	-13.3633	over | types.
0.124031	-13.3633	is | another
0.6	-13.3633	set | bounded
0.211111	-13.3634	t. | and
0.288889	-13.3634	using | system,
0.666667	-13.3635	yu, | and
0.5	-13.3635	web | data.
0.128099	-13.3636	finding | and
0.285714	-13.3636	for | interactions
0.122605	-13.3636	and | resources
0.0918803	-13.3637	for | possible
0.333333	-13.3637	validation | test
0.171642	-13.3639	is | having
0.142857	-13.3641	of | sources
0.232877	-13.3642	with | social
0.0995025	-13.3644	might | in
0.174242	-13.3644	sense | which
0.288889	-13.3644	previous | methods
0.0712468	-13.3645	add | the
0.470588	-13.3646	use | connective
0.470588	-13.3646	their | content.
0.117647	-13.3647	of | true
0.1875	-13.3647	an | meaning
0.295455	-13.3647	popped | the
0.208791	-13.3648	set | annotated
0.6	-13.3648	discourse | relations.
0.37037	-13.3649	to | edge.
0.714286	-13.365	added | previous
0.714286	-13.365	a | cells.
0.666667	-13.3651	17 | 13
0.384615	-13.3651	to | variable.
0.28	-13.3652	french-english | the
0.6	-13.3653	conduct | experiment
0.106628	-13.3654	found | for
0.288889	-13.3654	here | have
0.202128	-13.3655	and | multilingual
0.176	-13.3656	significant | over
0.714286	-13.3656	and | without.
0.295455	-13.3657	break | a
0.302326	-13.3659	scores | three
0.285714	-13.3659	build | each
0.269231	-13.366	the | objective.
0.214286	-13.366	model | able
0.214286	-13.366	the | null1
0.2	-13.3661	the | thousand
0.2	-13.3661	of | possible.
0.302326	-13.3662	max | max
0.12931	-13.3663	with | those
0.133333	-13.3664	models | word
0.37037	-13.3665	we | 10k
0.27451	-13.3665	also | up
0.107558	-13.3667	the | google
0.538462	-13.3667	of | suffixes.
0.0582192	-13.3667	that | other
0.666667	-13.3668	on | clases
0.151163	-13.3668	these | these
0.0561709	-13.3669	list | the
0.285714	-13.3669	when | less
0.0744382	-13.3671	with | other
0.0666667	-13.3672	that | into
0.295455	-13.3672	since | similar
0.0997506	-13.3673	information | or
0.428571	-13.3673	and | tales
1	-13.3673	0.41 | 0.34
1	-13.3673	cinderella | godfather
1	-13.3673	(dev) | cll
1	-13.3673	article/presentation | author/presenter
1	-13.3673	language | (talip),
1	-13.3673	{china, | .
1	-13.3673	gate | (cunningham
1	-13.3673	series, | predications.
1	-13.3673	np03 | np04
1	-13.3673	solve | problem:
1	-13.3673	ministerio | ciencia
1	-13.3673	mcccj=mcclosky-charniak-johnson | sd=stanford
1	-13.3673	structuring | own
1	-13.3673	-が | (lit.
1	-13.3673	translation: | _吗
1	-13.3673	company1 | person1
1	-13.3673	word | closest in
1	-13.3673	[a1,a2] | n[2]
1	-13.3673	system | matcher.
1	-13.3673	a* | admissible
1	-13.3673	nsf | darpa.
1	-13.3673	transliteration: | al*y
1	-13.3673	j | med
1	-13.3673	cotton* | closed*
1	-13.3673	besonderen | stimme
1	-13.3673	“sentiment | psychology”
1	-13.3673	“less | more”
1	-13.3673	〈tense〉pres | 〈mood〉dcl
1	-13.3673	eisner | satta
1	-13.3673	p(a|α)
1	-13.3673	y∗ | 1...n
1	-13.3673	stony | 11794,
1	-13.3673	insurance | insurance
1	-13.3673	y., | lester,
1	-13.3673	2000; | sang,
1	-13.3673	end | carnival-like
1	-13.3673	(prop-dep), | (noprop-dep),
1	-13.3673	0.00% | 6.00%
1	-13.3673	if: | node:case
1	-13.3673	u,v | s.t.
1	-13.3673	clark | pulman,
1	-13.3673	?r | geo:river
1	-13.3673	simpleew | revilp
1	-13.3673	terminals | terminals
1	-13.3673	chris | baltimore,
1	-13.3673	conditional | fields:
1	-13.3673	expectations | inside-outside
1	-13.3673	pf | 85.07
1	-13.3673	north | rada@cs.unt.edu
1	-13.3673	[j|β], | λ1
1	-13.3673	acquisition | allophonic
1	-13.3673	γ1 | γ2
1	-13.3673	here | sponsored
1	-13.3673	knight, | wei
1	-13.3673	“to” | “to”
1	-13.3673	2w2t | 4w4t
1	-13.3673	pennsylvania | 19104
1	-13.3673	essen | →dass
1	-13.3673	baldwin, | dan
1	-13.3673	juan | spain
1	-13.3673	review | helpfulness.
1	-13.3673	tone: | phrasal
1	-13.3673	workshop | 13,
1	-13.3673	.56 | .88
1	-13.3673	34450 | istanbul,
1	-13.3673	mt04 | mt05
1	-13.3673	v-measure | hirschberg,
1	-13.3673	.49 | .60
1	-13.3673	(+) | (−)
1	-13.3673	computing | keynes,
1	-13.3673	authorized | distribute
1	-13.3673	ep | gw
1	-13.3673	rubenstein | goodenough
1	-13.3673	(r) | 33.49
1	-13.3673	cooperative | (dst-jst)
1	-13.3673	none | 2966
1	-13.3673	dissociation | dissociation
1	-13.3673	1; | <=
1	-13.3673	ordered | score,
1	-13.3673	(on | left)
1	-13.3673	absoluteness | relativeness
1	-13.3673	“statistical | translation”.
1	-13.3673	noted | tend
1	-13.3673	(technologies | federal
1	-13.3673	φn | order
1	-13.3673	(golub | kahan,
1	-13.3673	chest | 05/14/08
1	-13.3673	grasp(grasper:washer, | reach(reacher:washer,
1	-13.3673	state: | song-title
1	-13.3673	scoretype-2 | scoretype-3
1	-13.3673	model | attr.↓
1	-13.3673	base | morphos
1	-13.3673	tense | disambiguation
1	-13.3673	beautiful | lovely
1	-13.3673	abduction | logic,
1	-13.3673	beneficial | english-side
1	-13.3673	windowdiff | hearst,
1	-13.3673	huddleston | pullum
1	-13.3673	reviews | f­measure
1	-13.3673	d1 | avg
1	-13.3673	(mcnaughton | papert,
1	-13.3673	h∈h | c∈c(h)
1	-13.3673	bioscope | (vincze
1	-13.3673	sent | me).'
1	-13.3673	ocr | character
1	-13.3673	c | ��'wt�luk��b.
1	-13.3673	 | +1
1	-13.3673	(majumdar | bhattacharyya,
1	-13.3673	subjectivity | (ssa)
1	-13.3673	 beam | f-­‐measure
1	-13.3673	and | leacock.
1	-13.3673	design∧ | 〈prop〉(p
1	-13.3673	btng | bpng
1	-13.3673	erotic | nonerotic
1	-13.3673	las | tedeval
1	-13.3673	(agirre | edmonds,
1	-13.3673	type | pre-clusterpair
1	-13.3673	alarmierende | der
1	-13.3673	t(z| | no
1	-13.3673	9.4 | colour.
1	-13.3673	slower, | em,
1	-13.3673	stephan | alex
1	-13.3673	19c | dictionary.
1	-13.3673	pivoting | callison-burch,
1	-13.3673	, | tjong
1	-13.3673	francisco, | november
1	-13.3673	livejournal | blogspot
1	-13.3673	marcus, | ann
1	-13.3673	yejin | brook
1	-13.3673	(yarowsky | ngai,
1	-13.3673	topi | kho
1	-13.3673	living | living
1	-13.3673	personality | personality
1	-13.3673	lookahead (depth | 0)
1	-13.3673	54 | 55
1	-13.3673	callison-burch, | fordyce,
1	-13.3673	kim | 2000b;
1	-13.3673	ng1,2 | sciences
1	-13.3673	btsg | btsg
1	-13.3673	†microsoft | asia,
1	-13.3673	plant | animal
1	-13.3673	community’s | 247762
1	-13.3673	desk) | específico
1	-13.3673	axiom | z),
1	-13.3673	abstractive | summaries,
1	-13.3673	ss | ss_n
1	-13.3673	“not | colour”
1	-13.3673	[[u1]]c | →utt1
1	-13.3673	vorbei:by | können
1	-13.3673	84.76 | 89.81
1	-13.3673	hal ri | pls
1	-13.3673	verbs | nominalizations.
1	-13.3673	h | �
1	-13.3673	yuji | 630-0192,
1	-13.3673	service | nj
1	-13.3673	chris | noah
1	-13.3673	queens | city
1	-13.3673	quadrants | octants
1	-13.3673	alfred | sloan
1	-13.3673	nicta | broadband,
1	-13.3673	cahill | riester
1	-13.3673	songs | mates
1	-13.3673	possessive, | (collection
1	-13.3673	(simplification | understand
1	-13.3673	explains | preferences
1	-13.3673	r1(x, | r2(y,
1	-13.3673	(witten | bell,
1	-13.3673	(jonker | volgenant,
1	-13.3673	stemmed | fini
1	-13.3673	u | expert.
1	-13.3673	npk | unseen.
1	-13.3673	’inhibitor’, | ’transformation’,
1	-13.3673	country, | profession.
1	-13.3673	passenger | 22%.
1	-13.3673	“about” | “about”
1	-13.3673	rb | wp
1	-13.3673	(bangalore | 1999;
1	-13.3673	〈arg〉(d | 〈artifact〉d
1	-13.3673	wsf_n | wsf_v
1	-13.3673	np | 1.n,
1	-13.3673	nflplayoffgame | nflteam
1	-13.3673	1998 | 93
1	-13.3673	pp left | avg.
1	-13.3673	< | >.
1	-13.3673	giza | ilp_seg
1	-13.3673	(sharifi | inouye,
1	-13.3673	〈creator〉 | relations—lexicalized
1	-13.3673	white | green,
1	-13.3673	’a’, | ’of’,
1	-13.3673	owl/dl | specifications
1	-13.3673	(andreevskaia | bergler,
1	-13.3673	i=0 t(fj|ei)
1	-13.3673	<prop> | (is|’s)
1	-13.3673	strengths | weaknesses.
1	-13.3673	processing | september,
1	-13.3673	%imp | 43.85
1	-13.3673	transverse | specific
1	-13.3673	nb | bmm
1	-13.3673	(using | features)
1	-13.3673	π23, | low
1	-13.3673	[condition | [conditioned
1	-13.3673	zahl | ihre
1	-13.3673	-wh | pein
1	-13.3673	amia | symposium,
1	-13.3673	set, | words).
1	-13.3673	relief | $5,000
1	-13.3673	ss_v | sf_v
1	-13.3673	(liang | 2006),
1	-13.3673	(r), | (p),
1	-13.3673	charles | mathematics
0.288889	-13.3673	a | prototype
0.6	-13.3673	here | 2010).
0.5	-13.3673	w=1.0 | sir
0.5	-13.3673	fairy | densities
0.102094	-13.3674	(2) | is
0.666667	-13.3674	sc | de=
0.666667	-13.3674	0.20 | 0.24
0.666667	-13.3674	<sy | ore
0.216867	-13.3674	segments | a
0.139303	-13.3675	documents | as
0.538462	-13.3675	for | xi,
0.37037	-13.3677	which | linked
0.149425	-13.3677	i.e., | a
0.444444	-13.3677	sign | valletta,
0.5	-13.3677	replacement | replacement
0.409091	-13.3678	h | ...
0.0590308	-13.3679	on | other
0.171642	-13.3679	was | under
0.211765	-13.368	to | outside
0.165517	-13.368	it | case
0.384615	-13.368	compositional | as
0.288889	-13.368	in | collections
0.0920771	-13.368	scores | for
0.538462	-13.368	(2010) | evaluate
0.8	-13.368	coreference | coreferences
0.8	-13.368	( | λ2
0.8	-13.368	⊃ | ...
0.8	-13.368	token | non-verb
0.8	-13.368	tsl | sl
0.8	-13.368	tfidf | lda1
0.8	-13.368	exploiting | corpora.
0.8	-13.368	compositional | zamparelli,
0.8	-13.368	1992; | kozareva
0.8	-13.368	third-order | reranking
0.324324	-13.3681	aid | to
0.157233	-13.3681	describes | for
0.5	-13.3681	0.68 | 0.78
0.16	-13.3682	(with | and
0.114865	-13.3682	summary | and
0.214286	-13.3682	and | constant
0.140704	-13.3683	tools | and
0.384615	-13.3683	(brants | franz,
0.324324	-13.3684	(see | 4).
0.324324	-13.3684	be | classes.
0.538462	-13.3684	supplementary | for
0.538462	-13.3684	of | ii.
0.0397772	-13.3685	the | set.
0.165517	-13.3685	latent | al.,
0.409091	-13.3685	require | annotation
0.714286	-13.3686	discussions | political
0.384615	-13.3686	label | assigned
0.8	-13.3686	- | entries.
0.8	-13.3686	diversity | perspectives
0.8	-13.3686	segmenting | labeling
0.8	-13.3686	networks | networks.
0.8	-13.3686	segment | segment,
0.34375	-13.3687	first | full
0.428571	-13.3688	the | unknown,
0.6	-13.3688	lda, | average
0.6	-13.3688	system | constraints.
0.190476	-13.3689	vary | of
0.242424	-13.369	the | revealed
0.34375	-13.369	is | search.
0.714286	-13.3691	using | durme
0.180328	-13.3692	characters | a
0.246154	-13.3692	classes | from
0.8	-13.3693	ieee | systems,
0.8	-13.3693	stored | trie
0.8	-13.3693	expensive | create.
0.8	-13.3693	informative | gap-fill
0.8	-13.3693	gender, | rationality
0.8	-13.3693	ambiguous | intermediate
0.8	-13.3693	kappa | landis
0.8	-13.3693	over | employed.
0.125	-13.3694	oracle | the
0.37037	-13.3695	when | aligned
0.204301	-13.3695	in | documents,
0.307692	-13.3697	occur | an
0.714286	-13.3697	features | cluster-level
0.190476	-13.3697	both | they
0.12931	-13.3697	processing | to
0.428571	-13.3698	use | recent
0.307692	-13.3699	in | reflecting
0.113269	-13.37	is | argument
0.8	-13.37	binary | subsampling
0.8	-13.37	highway | about
0.8	-13.37	cc | vbz
0.8	-13.37	total | spent
0.263158	-13.3702	of | covariance
0.538462	-13.3702	improvement | absolute
0.161074	-13.3703	supervised | on
0.295455	-13.3704	of | acl,
0.109091	-13.3706	to | various
0.180328	-13.3706	to | 2011).
0.163265	-13.3706	is | cost
0.157233	-13.3706	the | something
0.8	-13.3706	hit | ...
0.8	-13.3706	problem | tackled
0.8	-13.3706	dimensions | svd
0.8	-13.3706	sivaji | jadavpur
0.8	-13.3706	bars | show
0.8	-13.3706	answering | answers
0.13615	-13.3707	online | and
0.6	-13.3708	formula | used
0.714286	-13.3709	up | billion
0.714286	-13.3709	hence | context
0.167883	-13.371	to | minimum
0.246154	-13.371	project | al.,
0.5	-13.371	predict | scores
0.0788644	-13.3711	the | true
0.538462	-13.3711	gains | approach
0.117438	-13.3711	discuss | in
0.0685579	-13.3712	to | while
0.8	-13.3713	29 | 11
0.8	-13.3713	sampled | distribution.
0.8	-13.3713	greedy | guarantees
0.8	-13.3713	hhmm | n-gram
0.8	-13.3713	→ | 100.0%
0.8	-13.3713	approaches | proposed,
0.8	-13.3713	performs | best,
0.106017	-13.3714	to | opinion
0.714286	-13.3715	absolute | (or
0.714286	-13.3715	types | tokens.
0.182609	-13.3715	are | ways
0.151163	-13.3716	a | hidden
0.37037	-13.3717	in | resolving
0.17037	-13.3718	by | out
0.0882353	-13.3718	last | of
0.211765	-13.3719	in | 2003),
0.150289	-13.3719	are | multiple
0.151163	-13.3719	article | to
0.0787402	-13.3719	the | in | model.
0.288889	-13.3719	0.50 | of
0.5	-13.3719	rank | coefficient
0.5	-13.3719	able | sense
0.428571	-13.372	use | features:
0.538462	-13.372	considered | following
0.8	-13.372	against | held-out
0.8	-13.372	vs | ?
0.8	-13.372	percentage | games
0.8	-13.372	enhanced | mining.
0.8	-13.372	bootstrapping | resources.
0.8	-13.372	interaction | simulated
0.8	-13.372	dutch, | spanish
0.8	-13.372	outcome | study
0.384615	-13.3721	discovering | in
0.0726783	-13.3721	function | a
0.0604089	-13.3722	about | and
0.177419	-13.3723	and | normalized
0.5	-13.3723	importance | features
0.5	-13.3723	parsers | syntactic
0.196078	-13.3724	normal | and
0.27451	-13.3725	modifiers | in
0.0969267	-13.3725	all | using
0.444444	-13.3725	case | head
0.444444	-13.3725	research | 2006).
0.204301	-13.3726	(11) | the
0.315789	-13.3726	the | likelihood.
0.354839	-13.3726	topic | distributions
0.8	-13.3726	(galley | 2004),
0.171642	-13.3726	when | will
0.714286	-13.3727	c. | r.
0.174603	-13.3727	while | one
0.0916844	-13.3728	4 | with
0.384615	-13.3728	for | boundaries.
0.133028	-13.3729	to | sentences,
0.158228	-13.3729	overview | and
0.1875	-13.373	always | for
0.384615	-13.3731	ngram | for
0.354839	-13.3732	by | alignment.
0.714286	-13.3733	degree | polysemy
0.6	-13.3733	to | (1991).
0.8	-13.3733	degrades | performance.
0.8	-13.3733	violation | property
0.8	-13.3733	preference | connotative
0.17037	-13.3734	methods, | of
0.428571	-13.3734	the | package.
0.254237	-13.3737	annotated | different
0.238806	-13.3737	among | which
0.196078	-13.3737	if | node
0.0789889	-13.3737	to | user
0.25	-13.3738	there | cases
0.428571	-13.3738	pathway | event
0.384615	-13.3738	moreover, | may
0.109756	-13.3739	we | selection
0.140704	-13.3739	results | better
0.153374	-13.374	and | sequences
0.315789	-13.374	bleu | tuning
0.333333	-13.374	64 | a
0.8	-13.374	specified | file.
0.8	-13.374	197number | best
0.8	-13.374	natural | (emnlp).
0.8	-13.374	improvement | reranking
0.07	-13.3741	parser | and
0.254237	-13.3741	similarly, | as
0.089613	-13.3742	required | the
0.126016	-13.3742	and | manually
0.27451	-13.3743	model | include
0.1875	-13.3743	state-of-the-art | et
0.119134	-13.3743	mention | a
0.470588	-13.3744	by | annotator,
0.0708861	-13.3744	the | learned
0.354839	-13.3744	sentence | gold
0.444444	-13.3744	verb | roles
0.120755	-13.3745	those | with
0.154321	-13.3745	models | different
0.8	-13.3746	uj | .
0.8	-13.3746	structure | (rst)
0.216867	-13.3747	framework, | a
0.238806	-13.3747	s | all
0.444444	-13.3747	that | wizards
0.145946	-13.3748	is | frequent
0.6	-13.3748	thus, | larger
0.1875	-13.3749	any | those
0.25	-13.3749	the | synsets.
0.154321	-13.3749	learning | language
0.315789	-13.375	all | systems.
0.714286	-13.3751	with | coindexed
0.714286	-13.3751	kapa | for
0.25	-13.3751	and | i.
0.154321	-13.3751	n-grams | in
0.177419	-13.3751	(with | of
0.083779	-13.3752	c©2011 | linguistics | university
0.27451	-13.3753	w2 | w1
0.8	-13.3753	learner | predicates
0.8	-13.3753	cluster | together.
0.1875	-13.3755	often | al.,
0.16	-13.3755	to | necessary
0.130435	-13.3755	from | first
0.22973	-13.3756	µ | the
0.12987	-13.3756	context | a | a
0.37037	-13.3757	one | them.
0.161074	-13.3757	especially | a
0.238806	-13.3757	third, | of
0.258621	-13.3758	th | of
0.108761	-13.376	we | compute
0.384615	-13.376	from | remove
0.8	-13.376	maximizes | sum
0.8	-13.376	full | parsing,
0.8	-13.376	user | save
0.0877193	-13.3761	note that | to
0.12931	-13.3761	between | different
0.5	-13.3761	instance | ranking
0.246154	-13.3762	at | those
0.1875	-13.3763	years | and
0.384615	-13.3763	allow | rules
0.37037	-13.3764	first, | will
0.119565	-13.3764	topics | in
0.223684	-13.3764	on | scores.
0.092511	-13.3764	the original | and
0.163265	-13.3765	others | and
0.130435	-13.3766	these | use
0.190476	-13.3766	difference | not
0.444444	-13.3766	proposed | state-of-the-art
0.177419	-13.3766	20 | 30
0.409091	-13.3767	label, | in
0.8	-13.3767	label | artists
0.302326	-13.3768	syntactic | phrase
0.220779	-13.3768	this way, | a
0.5	-13.3769	constraints | structure
0.102981	-13.377	empirical | in
0.538462	-13.377	a | per-word
0.538462	-13.377	preposition | prepositions
0.0642782	-13.3772	for | test
0.182609	-13.3773	however | a
0.8	-13.3773	expressions | emotions
0.8	-13.3773	rare | retrieval
0.8	-13.3773	ask | judge
0.8	-13.3773	moses: | statistical
0.0743338	-13.3774	the | the | sets.
0.409091	-13.3774	which | interest
0.409091	-13.3774	is | answer.
0.153374	-13.3775	of | punctuation
0.0174868	-13.3775	the | of | with
0.538462	-13.3775	order | analyze
0.354839	-13.3777	vast majority | the
0.354839	-13.3777	of | vowels
0.124498	-13.3778	global | and
0.115646	-13.3778	create | and
0.206522	-13.3778	in | auxiliary
0.37037	-13.3779	constituent | by
0.223684	-13.3779	the | terminology
0.6	-13.3779	system | f-measure.
0.470588	-13.378	setting | learning
0.470588	-13.378	shallow | which
0.0679157	-13.378	most | is
0.235294	-13.378	all | being
0.8	-13.378	informatics | edinburgh,
0.8	-13.378	partial | chart.
0.238806	-13.3783	all | verb
0.470588	-13.3784	size | smaller
0.122137	-13.3784	we define | a
0.6	-13.3784	algorithm | generates
0.307692	-13.3785	additional | their
0.27451	-13.3785	these | automatic
0.220779	-13.3785	can | efficiently
0.0458667	-13.3785	is not | the
0.185841	-13.3786	second, | for
0.182609	-13.3786	and | instances.
0.5	-13.3786	so | just
0.8	-13.3787	0.05) | positions.
0.8	-13.3787	opinions | official
0.470588	-13.3788	renal | disease
0.470588	-13.3788	using | application
0.107143	-13.3788	the | yields
0.161074	-13.3788	that | change
0.173228	-13.3788	we train | and
0.242424	-13.3789	participants | or
0.354839	-13.3789	conclude | future
0.181034	-13.379	categories | are
0.226667	-13.379	subjectivity | for
0.110063	-13.379	and | 2002;
0.0659218	-13.3792	which | or
0.428571	-13.3792	equation | function
0.162162	-13.3793	although | can
0.315789	-13.3793	evaluates | and
0.185841	-13.3794	the | def=
0.09447	-13.3794	labels | and
0.8	-13.3794	specific | contexts.
0.206522	-13.3795	value | if
0.470588	-13.3796	true | 1
0.470588	-13.3796	normalized | length
0.333333	-13.3796	of | tx
0.288889	-13.3796	is | bit
0.28	-13.3797	i) | a
0.282609	-13.3797	this | rare
0.258621	-13.3797	is | 6.
0.184211	-13.3798	we | relations.
0.0911017	-13.3798	two | each
0.282609	-13.3799	• | queries
0.11985	-13.3799	many | with
0.0919037	-13.3799	accuracy | is
0.6	-13.3799	naive | streams
0.470588	-13.38	manual | quality
0.8	-13.38	cross-entity | entities
0.302326	-13.3801	survey | in
0.209302	-13.3801	training | evaluation
0.28	-13.3802	evaluate | results
0.282609	-13.3802	unlabeled | which
0.194175	-13.3802	these | similar
0.171875	-13.3803	because | words
0.1875	-13.3804	example | where
0.0894309	-13.3804	the | 2011
0.164384	-13.3806	of | (figure
0.444444	-13.3807	information | attributes
0.8	-13.3807	context | co-citation.
0.8	-13.3807	subsection | describe
0.8	-13.3807	jaku | jaos
0.8	-13.3807	scores | geometric
0.188679	-13.3808	produces | a
0.6	-13.3809	using | fold
0.428571	-13.381	how | relative
0.214286	-13.381	v | set
0.302326	-13.3811	two | classifiers
0.138614	-13.3812	in | there | no
0.0562249	-13.3812	is | when
0.0992366	-13.3813	the | occurs
0.238806	-13.3814	but | since
0.185841	-13.3817	on | initial
0.197917	-13.3818	however, | work
0.182609	-13.3818	both | information
0.8	-13.3821	years | had
0.8	-13.3821	formulate | generate
0.8	-13.3821	predicted | input.
0.384615	-13.3822	be | user’s
0.444444	-13.3822	parameter, | is
0.714286	-13.3824	methods | grid
0.135514	-13.3825	results | only
0.254237	-13.3825	the | intention
0.538462	-13.3825	increment | of
0.126531	-13.3826	from | corpora
0.324324	-13.3826	but | include
0.164286	-13.3826	the | carry
0.0888889	-13.3827	the | performs
0.142857	-13.3827	to | structure.
0.8	-13.3827	5-gram | 5-gram
0.8	-13.3827	also | sadness,
0.8	-13.3827	mert | scale
0.242424	-13.383	because | these
0.226667	-13.383	using | grammar
0.269231	-13.383	entity | types
0.188679	-13.383	system | score
0.223684	-13.383	annotated | two
0.324324	-13.3831	their | frequency
0.105714	-13.3831	to | mt
0.37037	-13.3832	i.e., | same
0.428571	-13.3832	feature | give
0.073509	-13.3832	a | label
0.184211	-13.3833	the | summarization.
0.17037	-13.3833	second | an
0.0858779	-13.3833	expression | the
0.151515	-13.3834	then | two
0.157895	-13.3834	denotes | and
0.8	-13.3834	crowdsourcing | collection
0.8	-13.3834	michael | chris
0.8	-13.3834	blast | user
0.25	-13.3835	over | pairs
0.428571	-13.3835	although | improvements
0.190476	-13.3835	derive | of
0.223684	-13.3836	to | fluency
0.5	-13.3837	now | if
0.5	-13.3837	ranker | using
0.0917031	-13.3838	to | label
0.188679	-13.3838	calculate | for
0.538462	-13.3838	set | sentence pairs.
0.151515	-13.3839	of | reduced
0.428571	-13.3839	it | meaning.
0.106017	-13.384	to | parameters
0.0605214	-13.384	a | score
0.354839	-13.384	nlu | nlu
0.354839	-13.384	as | encode
0.0783699	-13.384	some | that
0.333333	-13.3841	this | reason
0.444444	-13.3841	baseline | quality
0.8	-13.3841	alternative | entries.
0.8	-13.3841	pairs | compressions.
0.5	-13.3841	no | states
0.5	-13.3841	although | among
0.5	-13.3841	event | f-score
0.0907173	-13.3843	many | as
0.470588	-13.3844	• | query
0.333333	-13.3844	editor | the
0.37037	-13.3845	to | words).
0.315789	-13.3846	annotations | training
0.315789	-13.3846	improved | more
0.117857	-13.3847	have | they
0.220779	-13.3847	and | return
0.27451	-13.3848	brought | the
0.8	-13.3848	high-level | generation
0.15625	-13.3849	direct | in
0.0921053	-13.3849	algorithms | and
0.6	-13.385	variable | function
0.235294	-13.3851	of | sl
0.307692	-13.3852	(2009) | features
0.209302	-13.3852	has been | it
0.088	-13.3852	in | correct
0.269231	-13.3853	parsing | syntactic
0.333333	-13.3853	scores | 0
0.0805369	-13.3854	experimental | and
0.354839	-13.3855	lists | which
0.6	-13.3855	measures, | correlation
0.6	-13.3855	will | acceptable
0.8	-13.3855	thus | dna
0.8	-13.3855	when | employed.
0.214286	-13.3855	measures | which
0.0957944	-13.3856	extracted | a
0.123506	-13.3856	of | turn
0.142105	-13.3856	pronoun | the
0.106017	-13.3859	for | label
0.126531	-13.3859	the | showing
0.357143	-13.3859	for | reviews.
0.714286	-13.386	exact | found
0.097561	-13.386	might | a
0.409091	-13.386	statement | that
0.444444	-13.386	improvements | up
0.185841	-13.3861	ace | the
0.061127	-13.3861	are | this
0.324324	-13.3861	with | letter
0.245902	-13.3861	that | available.
0.242424	-13.3863	to | notion
0.161074	-13.3863	so, | of
0.0407125	-13.3863	models | and
0.409091	-13.3864	techniques | 2010).
0.190476	-13.3864	and | (with
0.6	-13.3865	in | eugenio
0.538462	-13.3866	because | nature
0.140625	-13.3867	in | turn
0.0888889	-13.3868	in this paper, | to
0.226667	-13.3868	into | when
0.428571	-13.3868	et | alia).
0.231884	-13.3868	are | accurate
0.0694789	-13.3868	end | the
0.0963855	-13.3869	previous | that
0.34375	-13.3869	whereas | model
0.37037	-13.387	= | wt
0.37037	-13.387	l | url
0.188679	-13.387	the | definite
0.307692	-13.3871	use | wordnet
0.28	-13.3871	figure | gives
0.27451	-13.3871	methods | extraction
0.384615	-13.3871	8 | per
0.470588	-13.3872	trained | lm
0.167883	-13.3872	to | contribution
0.333333	-13.3872	output | gold
0.315789	-13.3873	applied | text
0.0973236	-13.3875	estimate | of
0.6	-13.3875	wordnet: | electronic
0.8	-13.3875	afp | news
0.8	-13.3875	engineering | sciences
0.470588	-13.3876	conclude | paper
0.258621	-13.3876	and | concepts.
0.177419	-13.3876	is | necessary
0.126582	-13.3877	annotations | a
0.220779	-13.3877	individual | with
0.324324	-13.3878	relations | pairs
0.384615	-13.388	et | huang,
0.288889	-13.388	version | used
0.0813559	-13.3881	and | (e.g.,
0.34375	-13.3881	essence, | the
0.109375	-13.3882	as | standard
0.130631	-13.3882	(2010) | that
0.146067	-13.3882	further | with
0.206897	-13.3882	languages | which
0.8	-13.3882	on | 10–18,
0.428571	-13.3883	important | nlp
0.282609	-13.3886	the | scenarios.
0.0921053	-13.3886	from | two
0.245902	-13.3886	considering | with
0.444444	-13.3886	many | during
0.2	-13.3887	subjects | in
0.307692	-13.3889	under | using
0.238806	-13.3889	computer | university,
0.288889	-13.3889	each | (i.e.
0.8	-13.3889	(bodrumlu | al.,
0.8	-13.3889	pem | bilingual
0.8	-13.3889	nlu | dm
0.8	-13.3889	eisner, | eisner,
0.444444	-13.389	the | “the | “the
0.470588	-13.3892	that | attributes.
0.0544061	-13.3892	to | given
0.245902	-13.3893	ground | of
0.538462	-13.3893	for | english-german
0.269231	-13.3894	given | query
0.37037	-13.3895	add | information
0.176	-13.3895	under | are
0.8	-13.3895	trained | tracer
0.8	-13.3895	learn | threshold
0.8	-13.3895	e | j=1
0.140625	-13.3896	1 | n
0.714286	-13.3896	10% | accuracy
0.6	-13.3896	system | out-of-domain
0.235294	-13.3897	and | ccg
0.061284	-13.3897	– | of
0.209302	-13.3898	from | resources
0.231884	-13.3898	results | selection
0.538462	-13.3898	humans | than
0.295455	-13.3898	for | weights.
0.254237	-13.3901	use | n-gram
0.117438	-13.3901	and | times
0.37037	-13.3902	features | parser.
0.714286	-13.3902	fertility | word
0.714286	-13.3902	definition | confinement
0.357143	-13.3902	baseline | significant
0.8	-13.3902	literal | usage
0.16	-13.3904	to | state-of-the-art
0.0326385	-13.3904	this | by
0.37037	-13.3905	description | corpus
0.269231	-13.3905	in | reviews.
0.220779	-13.3905	the | positive.
0.143617	-13.3905	successfully | the
0.409091	-13.3905	(2007) | approach
0.0162137	-13.3906	and | and | and
0.0685096	-13.3906	an | that | a
0.150602	-13.3906	world | to
0.185841	-13.3907	results | development
0.142105	-13.3908	every | that
0.409091	-13.3909	corpora | languages.
0.8	-13.3909	(artstein | 2008).
0.8	-13.3909	errors | splitting
0.8	-13.3909	regina | massachusetts
0.8	-13.3909	impute | chinese
0.202128	-13.391	this | relatively
0.131222	-13.391	a | weighted
0.0881764	-13.391	and | .
0.169118	-13.3911	issue | to
0.185841	-13.3911	our | uses | and
0.409091	-13.3912	selects | with
0.444444	-13.3913	vary | number
0.5	-13.3913	and | correspondences
0.282609	-13.3915	our | accuracy.
0.333333	-13.3915	conditional | fields | al.,
0.231884	-13.3916	simple | one
0.538462	-13.3916	several | significant
0.8	-13.3916	event extraction | events,
0.8	-13.3916	across | grouped
0.163121	-13.3917	that | string
0.226667	-13.3917	heuristic | for
0.165468	-13.3917	the | rating
0.101877	-13.3917	to | discourse
0.0905263	-13.3917	evidence | the
0.0814558	-13.3918	the | partial
0.409091	-13.3919	here, | all
0.470588	-13.392	is | form:
0.074928	-13.392	overall | of
0.162162	-13.3921	most | these
0.177966	-13.3921	show | different
0.269231	-13.3921	of | generation,
0.128205	-13.3921	negative | is
0.194175	-13.3922	by | supervised
0.184211	-13.3923	a | analysis.
0.8	-13.3923	trigger | edges.
0.0936073	-13.3923	a | several
0.0408076	-13.3924	our | with
0.258621	-13.3924	and | moves
0.117021	-13.3925	a | h
0.0734072	-13.3925	the | means
0.062753	-13.3927	the | trees
0.138614	-13.3929	this | corpus.
0.124498	-13.3929	used | system
0.254237	-13.3929	words | translations
0.8	-13.3929	modeling | conversational
0.8	-13.3929	gap-fill | key
0.0576096	-13.393	results | on
0.27451	-13.393	source | reference
0.079288	-13.3931	however, | not | the
0.444444	-13.3932	patterns | rules
0.269231	-13.3933	factors | on
0.324324	-13.3933	words | good
0.0217737	-13.3933	the | words | the
0.179487	-13.3934	which | less
0.157895	-13.3934	algorithm | all
0.295455	-13.3934	to | hits
0.357143	-13.3935	context | without
0.538462	-13.3935	combined | several
0.307692	-13.3936	performance | comparable
0.282609	-13.3936	research | foundation
0.118959	-13.3937	(which | the
0.161074	-13.3937	the | cognitive
0.409091	-13.3937	important | opinions
0.245902	-13.3937	must | one
0.245902	-13.3937	of | probabilities.
0.6	-13.3937	by | evaluators
0.190476	-13.3937	model | reordering
0.315789	-13.3938	the | specified.
0.315789	-13.3938	created | corpus
0.254237	-13.3938	algorithm | other
0.428571	-13.3938	are | unless
0.0975	-13.3938	content | in
0.714286	-13.3939	name | lda
0.714286	-13.3939	heavily | which
0.714286	-13.3939	seem | require
0.714286	-13.3939	alternatively, | may
0.470588	-13.394	as | maxent
0.444444	-13.394	apply | methods
0.264151	-13.394	truth | of
0.223684	-13.3941	ordering | a
0.115385	-13.3942	type | as
0.324324	-13.3942	evidence | or
0.6	-13.3942	classes | appropriate
0.384615	-13.3943	that | obtains
0.15894	-13.3943	while | used
0.114094	-13.3943	with | first
0.8	-13.3943	within | management
0.391304	-13.3944	other | sources
0.226667	-13.3944	translation | input
0.0950119	-13.3945	in | experiments, | the
0.714286	-13.3945	baseline | obtained using
0.357143	-13.3945	it | interesting | note
0.315789	-13.3946	task | many
0.444444	-13.3947	we | pattern,
0.197917	-13.3948	any | such
0.0739687	-13.3948	we | standard
0.135266	-13.3948	to | degree
0.37037	-13.3949	outperforms | bleu
0.333333	-13.3949	meant | in
0.214286	-13.395	stimulus | the
0.0411739	-13.3951	2 | and
0.714286	-13.3951	complement | other
0.124	-13.3951	sentences | which
0.288889	-13.3951	to | state,
0.0992366	-13.3951	propose | for
0.34375	-13.3952	and | 2008a).
0.118081	-13.3953	9 | of
0.197917	-13.3953	it | known
0.384615	-13.3953	then | path
0.3	-13.3955	plural | in
0.333333	-13.3955	refer | reader
0.179487	-13.3956	the | goes
0.8	-13.3957	necessary | experiments
0.8	-13.3957	inventory | polysemous
0.111842	-13.3958	we observe | the
0.282609	-13.3958	combination | systems
0.288889	-13.3959	to | pathway
0.470588	-13.396	each | 500
0.34375	-13.396	what | information
0.235294	-13.3962	parsers | with
0.384615	-13.3963	builds | by
0.0429389	-13.3964	a | translation
0.8	-13.3964	distance metric | distance
0.197917	-13.3965	containing | with
0.245902	-13.3966	while | high
0.3	-13.3968	100.0 | 100.0
0.3	-13.3968	johnson | and
0.184211	-13.3968	content | from
0.25	-13.3969	baseline | model.
0.142857	-13.3969	r | c
0.5	-13.3969	cost | can be
0.714286	-13.397	cast | binary
0.714286	-13.397	could | efficiently
0.8	-13.3971	proposed | adequacy
0.8	-13.3971	terms | changed
0.333333	-13.3972	both | frequency
0.409091	-13.3972	twitter | their
0.0850662	-13.3972	3.1 | of
0.214286	-13.3972	items | are
0.192308	-13.3973	the | bag-of-words
0.0661345	-13.3973	have | be
0.226667	-13.3973	to | applications.
0.5	-13.3973	el | la
0.0982368	-13.3975	report | and
0.315789	-13.3975	the | theory.
0.409091	-13.3975	collect | set
0.259259	-13.3975	of | fillers
0.137255	-13.3976	sets | are
0.127119	-13.3976	to | decision
0.114583	-13.3977	49th annual | 19-24,
0.188679	-13.3977	tags | words
0.295455	-13.3977	meaning | their
0.8	-13.3977	track | track.
0.323529	-13.3978	an | arc
0.133028	-13.3978	reason | is
0.0895833	-13.3979	for | types
0.130631	-13.398	the | gains
0.269231	-13.3981	oracle | is
0.444444	-13.3981	# | 100
0.161972	-13.3981	of | matched
0.204545	-13.3983	of | vice
0.170543	-13.3983	on | performance.
0.254237	-13.3983	function | used
0.0989848	-13.3983	current | to
0.391304	-13.3984	then | cost
0.34375	-13.3984	10,000 | and
0.211765	-13.3985	position | word
0.161972	-13.3985	top | as
0.192308	-13.3986	baseline | set
0.384615	-13.3986	parse | rule
0.235294	-13.3986	also | very
0.108025	-13.3988	first | each
0.6	-13.3988	within | particular
0.259259	-13.3988	a | paradigm
0.264151	-13.3988	results | comparison
0.384615	-13.3989	science, | university,
0.295455	-13.3989	using | achieve
0.37037	-13.399	of | importance.
0.192308	-13.399	from | level
0.254237	-13.399	75% | the
0.5	-13.399	c | c)
0.391304	-13.3991	model outperforms | in
0.357143	-13.3991	usually | information
0.8	-13.3991	accuracy | partials
0.307692	-13.3992	presumably | of
0.204545	-13.3992	a | read
0.206897	-13.3992	similarly, | is
0.27451	-13.3993	degree | by
0.165468	-13.3993	of | tested
0.188679	-13.3993	since | were
0.0754438	-13.3994	objective | the
0.254237	-13.3994	is | sparse
0.116608	-13.3995	we | prior
0.34375	-13.3996	here | its
0.0700637	-13.3997	tree | a
0.307692	-13.3997	in | public
0.315789	-13.3997	is | superior
0.226667	-13.3998	to | email
0.171875	-13.3998	to | robust
0.288889	-13.3998	computation | that
0.8	-13.3998	elaborate | intended
0.8	-13.3998	evaluated | 10-fold
0.282609	-13.3999	expensive | and
0.384615	-13.3999	part-of-speech | pos
0.323529	-13.4	and | fourth
0.117647	-13.4	a | 2007).
0.714286	-13.4	provide | context.
0.714286	-13.4	players | our
0.428571	-13.4	synchronous | rules
0.104225	-13.4	we | parameters
0.25	-13.4002	but | work
0.264151	-13.4002	that | promising
0.16	-13.4003	prior | on
0.0821678	-13.4004	there is | a
0.122047	-13.4004	words | same
0.8	-13.4005	novel | projection
0.8	-13.4005	report | setting.
0.8	-13.4005	superiority | over
0.188679	-13.4006	that | typically
0.223684	-13.4008	that | objects
0.0834846	-13.4009	a | within
0.0616915	-13.4009	new | to
0.214286	-13.4009	40% | the
0.190476	-13.401	who | with
0.8	-13.4012	our | (dcs
0.161074	-13.4013	the | denoting
0.114187	-13.4016	a | concept
0.470588	-13.4017	have | contexts.
0.0762195	-13.4017	2007; | et
0.714286	-13.4019	evaluations | metrics
0.714286	-13.4019	much | etc.
0.062	-13.4019	take | the
0.323529	-13.402	drawn | distribution
0.103352	-13.402	wordnet | and
0.0236987	-13.4021	/ | / | /
0.391304	-13.4021	⊆ | is
0.357143	-13.4021	to | bipartite
0.37037	-13.4022	dialogue | tutorial
0.34375	-13.4022	synonyms | are
0.103064	-13.4023	articles | and
0.245902	-13.4023	of | activation
0.0987342	-13.4023	of | e.g.,
0.295455	-13.4023	with | score,
0.135922	-13.4024	is | existing
0.714286	-13.4025	woz | wizard
0.217949	-13.4026	indeed | in
0.103641	-13.4026	in | full
0.8	-13.4026	me | data
0.8	-13.4026	including | fragments
0.538462	-13.4027	translations | language,
0.101928	-13.4028	= | t
0.357143	-13.4028	is | modular
0.357143	-13.4028	13 | 15
0.0209639	-13.4028	this | in
0.37037	-13.4029	reverse | with
0.470588	-13.4029	class-based | of
0.470588	-13.4029	columbia | york,
0.384615	-13.4029	for | necessary.
0.384615	-13.4029	approach: | and
0.153374	-13.403	and | participants
0.295455	-13.403	largest | is
0.714286	-13.4031	3.3 | algorithms
0.124	-13.4031	lexical | from
0.235294	-13.4032	compare | other
0.538462	-13.4032	showed | very
0.315789	-13.4033	latter | which
0.0810345	-13.4033	these | can
0.428571	-13.4033	knowledge | human
0.8	-13.4033	produces | labeled
0.8	-13.4033	seven | final
0.5	-13.4033	graphs | k-nn
0.391304	-13.4034	of | los
0.254237	-13.4034	this | smaller
0.193878	-13.4034	with | results.
0.114583	-13.4035	there | one
0.6	-13.4035	uses | lm
0.173228	-13.4035	and | edit
0.245902	-13.4036	by | tagger
0.428571	-13.4037	c(a | b)
0.288889	-13.4038	corpus-based | and
0.5	-13.4038	nlp | part-of-speech
0.269231	-13.4039	phrase-based | using
0.110063	-13.4039	only | at
0.0347201	-13.4039	words | and
0.181034	-13.404	several | or
0.195876	-13.404	work | syntactic
0.211765	-13.404	unlabeled | data
0.245902	-13.404	a | input,
0.202128	-13.4041	our | improves
0.0708661	-13.4041	in | topic
0.470588	-13.4041	principal | is
0.217949	-13.4041	of | software
0.37037	-13.4042	furthermore, | all
0.37037	-13.4042	relation | corpus
0.101333	-13.4042	natural | to
0.5	-13.4042	the | §3.
0.5	-13.4042	easy | but
0.27451	-13.4043	in | algorithms.
0.170543	-13.4043	collection of | a
0.166667	-13.4044	a | structured
0.238806	-13.4044	taken | be
0.6	-13.4045	each | checked
0.6	-13.4045	graph | vertices.
0.6	-13.4045	describes | where
0.174603	-13.4045	improve | with
0.109718	-13.4045	feature | which
0.0764526	-13.4046	of | development
0.8	-13.4046	size | shape
0.5	-13.4046	a | deduction
0.231884	-13.4047	claim | of
0.37037	-13.4048	parasense | the
0.276596	-13.4048	domain | language
0.295455	-13.4048	as | construct
0.179487	-13.405	our | reordering
0.384615	-13.4052	from | approaches.
0.223684	-13.4052	is | variation
0.0718157	-13.4053	relation | a
0.238806	-13.4053	sentences | any
0.0994764	-13.4053	applied | in
0.8	-13.4053	measures: | measures
0.8	-13.4053	potentially | applications
0.391304	-13.4055	our | preference
0.5	-13.4055	not | resource.
0.202128	-13.4056	motivation | in
0.209302	-13.4056	stop | and
0.231884	-13.4059	use | while
0.126582	-13.406	we | cluster
0.220779	-13.406	i) | to
0.409091	-13.406	keeping | that
0.6	-13.406	participle | verb
0.8	-13.406	density | relevant
0.130631	-13.4061	problem | by
0.288889	-13.4061	as | resources.
0.714286	-13.4062	languages, | english,
0.444444	-13.4062	available | data,
0.129464	-13.4063	generally | of
0.193878	-13.4063	the | seeds
0.295455	-13.4063	use | decoding
0.129464	-13.4064	role | is
0.0671378	-13.4064	for | the same
0.127119	-13.4064	match | and
0.5	-13.4064	need | important
0.384615	-13.4065	(houvardas | stamatatos,
0.146893	-13.4066	problem, | and
0.241935	-13.4066	for | respective
0.169231	-13.4066	it | often
0.288889	-13.4066	the | mrs
0.0824373	-13.4066	given | =
0.6	-13.4066	text | improved
0.714286	-13.4068	unsupervised | induction,
0.288889	-13.4068	would | them
0.245902	-13.4068	although | were
0.538462	-13.4069	author | paper
0.114983	-13.407	to | segmentation
0.0754148	-13.4071	several | to
0.153374	-13.4071	for | nouns
0.139175	-13.4071	chunk | the
0.37037	-13.4074	results | complete
0.470588	-13.4074	grounding | by
0.714286	-13.4074	supervised | exploits
0.211765	-13.4074	part | research
0.315789	-13.4074	models | 2007).
0.220779	-13.4074	this | assumption
0.8	-13.4074	posterior | top
0.204545	-13.4075	unsupervised | model
0.391304	-13.4075	swap | of
0.259259	-13.4075	and | manner
0.220779	-13.4076	for | (as
0.6	-13.4076	also | shorter
0.21519	-13.4076	the | are,
0.269231	-13.4077	v | t
0.185185	-13.4078	paraphrase | from
0.0640608	-13.4079	terms | and
0.538462	-13.4083	these | methods,
0.269231	-13.4084	statistical | features
0.391304	-13.4085	different | contexts
0.228571	-13.4085	which | subject
0.139175	-13.4085	ranking | is
0.444444	-13.4085	• | within
0.133333	-13.4088	motivation | the
0.538462	-13.4088	obtain | representation
0.8	-13.4088	informatics, | crichton
0.112583	-13.4089	these | only
0.188679	-13.4089	the | judgements
0.231884	-13.4089	of | family
0.333333	-13.409	for example, | corresponding
0.155844	-13.4092	this | instances
0.409091	-13.4092	a | reading.
0.37037	-13.4093	different | forms
0.100796	-13.4093	results | all
0.0425733	-13.4096	in | same
0.282609	-13.4097	does | improve
0.538462	-13.4097	– | structure
0.470588	-13.4098	stereotype | a
0.118519	-13.4099	for example | the
0.153846	-13.41	our | structure
0.142857	-13.41	working | the
0.0952381	-13.4101	also | our
0.137755	-13.4101	an | function
0.444444	-13.4101	sequence | labels
0.146893	-13.4102	in | having
0.089172	-13.4102	for | verb
0.6	-13.4102	advantage | performance
0.538462	-13.4102	requires | given
0.538462	-13.4102	from | source,
0.8	-13.4102	within | systems.
0.114583	-13.4103	of | written
0.135266	-13.4103	the | indicated
0.0729313	-13.4103	the | cases
0.0934579	-13.4104	the | (as
0.185185	-13.4105	that | evidence
0.118519	-13.4106	idea | in
0.538462	-13.4107	decoding | finding
0.231884	-13.4108	decide | or
0.444444	-13.4108	variable | over
0.0516039	-13.4109	however, | in
0.323529	-13.4109	common | were
0.259259	-13.4109	challenging | in
0.8	-13.4109	pre-processing | extraction
0.8	-13.4109	ambiguous | assigning
0.8	-13.4109	pairs | nlf
0.241935	-13.411	this | intended
0.276596	-13.411	are | assigned to
0.231884	-13.4112	constituent | is
0.259259	-13.4112	for | constructions
0.2	-13.4112	2006), | a
0.333333	-13.4113	messages | or
0.6	-13.4113	linguistic | languages.
0.6	-13.4113	goal | correctly
0.149701	-13.4114	a | setting
0.136585	-13.4115	collected | and
0.0547185	-13.4116	of | while
0.3	-13.4116	of | matching.
0.8	-13.4116	imt | online
0.5	-13.4116	will | methods.
0.217949	-13.4117	to | something
0.714286	-13.4118	closed | 2011
0.714286	-13.4118	trend | results
0.6	-13.4118	follows: | introduces
0.259259	-13.4118	since | algorithm
0.117216	-13.4118	add | and
0.152439	-13.4119	that | just
0.183486	-13.4119	of | commonly
0.5	-13.412	based | morphology
0.409091	-13.4121	projection, | the
0.231884	-13.4122	structure | sentence
0.391304	-13.4123	particular | semantic
0.315789	-13.4123	of | 10)
0.6	-13.4123	in | personout
0.8	-13.4123	intervening | are
0.714286	-13.4124	5 5 | .
0.3	-13.4124	of | enabling
0.165468	-13.4125	s | ,
0.159722	-13.4125	are | sentences.
0.538462	-13.4125	encode | into
0.3	-13.4126	model | pairs.
0.104046	-13.4126	term | in
0.6	-13.4128	(em) | (dempster
0.103448	-13.4129	we | labels
0.151515	-13.4131	the | introducing
0.145251	-13.4132	the | extremely
0.152439	-13.4132	as | step
0.444444	-13.4132	translation | (koehn,
0.177966	-13.4135	the | considerable
0.0765625	-13.4136	user | and
0.323529	-13.4137	differences | performance
0.323529	-13.4137	but | overall
0.276596	-13.4137	argmax | summationdisplay
0.8	-13.4137	(% | recall
0.175	-13.4138	first | other
0.470588	-13.4139	dialogue, | that
0.225352	-13.4139	each | had
0.333333	-13.4139	that | phrases,
0.444444	-13.4139	productdisplay | probability
0.538462	-13.4139	words | pos tags,
0.34375	-13.4139	these | known
0.104348	-13.414	will | this
0.21519	-13.414	with | base
0.315789	-13.4142	the | enough.
0.315789	-13.4142	supervised | trained
0.333333	-13.4142	existing | into
0.470588	-13.4143	three | sets,
0.176471	-13.4143	to | handle
0.160839	-13.4144	message | a
0.225352	-13.4144	never | a
0.357143	-13.4145	annotation | corpus,
0.245902	-13.4145	other | less
0.171875	-13.4147	pair | each
0.169231	-13.4148	we aim | the
0.34375	-13.4148	name, | and
0.167939	-13.4149	of | game
0.714286	-13.4149	conditions, | models
0.357143	-13.4149	in: | conference
0.357143	-13.4149	(also | a
0.6	-13.4149	each | ...,
0.0710456	-13.4149	the | clusters
0.3	-13.415	on | absolute
0.206897	-13.4151	schemes | and
0.323529	-13.4152	chris | for
0.545455	-13.4152	a | sign.
0.195876	-13.4154	extracted | each
0.220779	-13.4154	writing | to
0.2	-13.4154	rst | the
0.538462	-13.4154	large | examples.
0.8	-13.4158	bayesian | topic
0.8	-13.4158	remaining | assigned
0.104956	-13.4159	sample | of
0.0994764	-13.4159	if | have
0.307692	-13.416	pair | relation
0.6	-13.416	knowledge, | study
0.6	-13.416	maryland | usa
0.164286	-13.416	for | purpose
0.241935	-13.4161	5 | proposed
0.0830258	-13.4161	its | as
0.714286	-13.4162	extraction | whole
0.25	-13.4162	more | high
0.223684	-13.4162	last | on
0.181034	-13.4163	to | occurrences
0.34375	-13.4163	2. | also
0.391304	-13.4164	approach | approaches.
0.220779	-13.4164	distributions | as
0.176471	-13.4165	cluster | the | cluster
0.8	-13.4165	relations | primitives.
0.8	-13.4165	main | modal
0.8	-13.4165	french | connectives
0.282609	-13.4166	da | a
0.220779	-13.4166	where | type
0.0915179	-13.4166	since | with
0.107362	-13.4166	to | n-gram
0.142077	-13.4167	when | set
0.128889	-13.4167	with | gold
0.223684	-13.4167	offset | the
0.545455	-13.4167	communicative | such
0.235294	-13.4168	such | high
0.151515	-13.4169	program | the
0.0854369	-13.4169	in | –
0.143646	-13.417	the | stored
0.444444	-13.417	identify | entities
0.37037	-13.4171	whether | current
0.391304	-13.4171	that | significantly different
0.116788	-13.4171	expressions | in
0.8	-13.4172	barzilay | elhadad
0.181818	-13.4173	japanese | in
0.714286	-13.4174	agenda | decoding
0.714286	-13.4174	expert | median
0.714286	-13.4174	as expected, | models
0.0899358	-13.4176	are | over
0.545455	-13.4176	nus | learner
0.391304	-13.4177	solved | using
0.197917	-13.4177	a | relatedness
0.8	-13.4179	in | disability
0.8	-13.4179	smallest | aligned
0.333333	-13.418	second | uses
0.409091	-13.4181	feature | added
0.5	-13.4181	highest | 3
0.254545	-13.4183	dialogue | task
0.193878	-13.4183	each | random
0.139896	-13.4184	can | very
0.0984456	-13.4186	of | mentions
0.0677146	-13.4186	described | and
0.0976864	-13.4186	of | certain
0.333333	-13.4186	our | promising
0.6	-13.4186	following | detail.
0.545455	-13.4186	against | majority
0.206897	-13.4186	coreference | that
0.8	-13.4186	requires | time,
0.8	-13.4186	modeling | morphological
0.37037	-13.4187	acoustic | from
0.714286	-13.4187	events | done
0.409091	-13.4188	all | components
0.292683	-13.4189	while | approaches
0.259259	-13.4189	per | each
0.391304	-13.4191	vectors | obtained
0.545455	-13.4191	last | including
0.545455	-13.4191	ˆe | =
0.333333	-13.4192	we plan | use
0.714286	-13.4193	make | mistakes
0.241935	-13.4193	the | repository
0.188679	-13.4193	instead | using
0.241935	-13.4195	sections, | the
0.470588	-13.4197	one | back
0.231884	-13.4197	second, | with
0.6	-13.4197	variance | =
0.714286	-13.4199	way | them.
0.0732759	-13.4199	subset | the
0.0989583	-13.42	in | considered
0.323529	-13.42	probably | that
0.259259	-13.42	and | target.
0.391304	-13.4201	one | lack
0.228571	-13.4203	argued | of
0.0512465	-13.4205	with | which
0.113793	-13.4206	specifically, we | the
0.225352	-13.4206	for | 12
0.169231	-13.4206	interpretation | a
0.185185	-13.4206	features | speech
0.409091	-13.4206	acl-hlt | 19-24
0.8	-13.4207	verb | historical
0.225352	-13.4208	we | parameters.
0.152866	-13.4208	for | comments.
0.282609	-13.4208	translation | comparable
0.193878	-13.4208	them | or
0.470588	-13.4209	getting | with
0.470588	-13.4209	4 | details
0.444444	-13.4209	for | all | including
0.147929	-13.421	uses | from
0.193878	-13.421	can | observed
0.153846	-13.4212	the | y.
0.333333	-13.4212	single | semantic
0.288889	-13.4212	and | (2011)
0.21519	-13.4212	for | go
0.5	-13.4212	are | functions,
0.470588	-13.4213	σ∗ | is
0.315789	-13.4214	limitation | to
0.169231	-13.4214	defined | set
0.197917	-13.4214	to establish | the
0.292683	-13.4215	is | throughout
0.193878	-13.4215	can be | text
0.264151	-13.4215	goals | of
0.167939	-13.4216	table 2, | the
0.3	-13.4216	very | many
0.209302	-13.4217	this | vectors
0.34375	-13.4217	an | contextual
0.470588	-13.4218	special | other
0.2	-13.4218	with | (i.e.,
0.103152	-13.4219	first | an
0.0632369	-13.4219	of | specific
0.391304	-13.4219	another | method
0.090708	-13.4219	are | some
0.0655926	-13.4221	target | a
0.21519	-13.4221	systems | one
0.102857	-13.4223	the | unigram
0.21519	-13.4223	our | monolingual
0.144444	-13.4224	scores | from
0.175	-13.4225	p | by
0.5	-13.4225	(raghavan | al.,
0.2	-13.4226	books | the
0.102564	-13.4227	precision | for
0.409091	-13.4228	factorial | of
0.8	-13.4228	also | proposed.
0.202247	-13.4229	if | them
0.3	-13.4229	besides | we
0.124481	-13.4229	the | pruning
0.19	-13.4229	decisions | to
0.538462	-13.4229	such | weighted
0.714286	-13.423	|a|
0.120623	-13.423	all | are | as
0.173554	-13.4232	the | leaf
0.0893617	-13.4233	it | also
0.421053	-13.4233	research | domain
0.421053	-13.4233	absolute | points
0.6	-13.4233	formed | other
0.444444	-13.4233	of | arc.
0.470588	-13.4234	approach | 2004).
0.470588	-13.4234	however, | assume
0.156863	-13.4234	(2006) | to
0.538462	-13.4234	analysis | tweets
0.25	-13.4235	turns | a
0.288889	-13.4235	acknowledgments | national science
0.195876	-13.4236	in | biomedical
0.714286	-13.4237	coarse-grained | relations
0.714286	-13.4237	training | sgd
0.14881	-13.4237	f: | f:
0.444444	-13.4237	below: | =
0.125523	-13.4238	the | robust
0.470588	-13.4238	is: | +
0.195876	-13.4238	basic | with
0.235294	-13.4238	the | connections
0.167939	-13.4239	behind | is
0.391304	-13.4239	150 | number
0.6	-13.4239	of | varies.
0.288889	-13.424	output | different
0.217949	-13.4241	and | japan
0.142077	-13.4242	we | equation
0.0979381	-13.4242	α | the
0.136364	-13.4242	still | for
0.0434343	-13.4243	models | of
0.5	-13.4243	cases | (e.g.
0.144444	-13.4244	3 | section
0.0693196	-13.4244	other | that
0.0731707	-13.4245	produced | the
0.333333	-13.4245	the | contain.
0.5	-13.4247	needed | set
0.166667	-13.4248	5 | system
0.538462	-13.4248	2008), | 2008),
0.307692	-13.4249	target | opinion
0.111864	-13.4249	of | edge
0.0523256	-13.4249	the | the | words.
0.714286	-13.4249	sequence | letters
0.444444	-13.4249	senses | disambiguation
0.259259	-13.425	summationdisplay | (7)
0.421053	-13.4251	only | matches
0.421053	-13.4251	ranked | according
0.0701987	-13.4252	3 | a | of
0.259259	-13.4252	model, | feature
0.409091	-13.4253	levels | or
0.228571	-13.4254	morpheme | and
0.130841	-13.4255	the | observations
0.132075	-13.4256	with | specific
0.185185	-13.4256	the | permutation
0.235294	-13.4256	on | vs.
0.307692	-13.4257	approach | across
0.470588	-13.4259	from | 5).
0.197917	-13.4259	is | allowing
0.101093	-13.4259	paraphrase | and
0.159722	-13.426	n-gram | with
0.6	-13.426	university | sweden
0.0923788	-13.4261	generation | and
0.202247	-13.4261	the | actions.
0.166667	-13.4262	n-gram | as
0.259259	-13.4262	they | words.
0.276596	-13.4263	recognition | an
0.545455	-13.4263	simple | nlp
0.545455	-13.4263	is | false.
0.8	-13.4263	ordered | ranking
0.8	-13.4263	number | reduced,
0.8	-13.4263	categorical | distributional
0.333333	-13.4265	of | useful.
0.6	-13.4265	terms | lexicon
0.270833	-13.4266	signal | to
0.0934579	-13.4266	evaluation | with
0.292683	-13.4267	tune | in
0.391304	-13.4267	= | −1
0.470588	-13.4267	they | typically
0.538462	-13.4267	ml | ml
0.538462	-13.4267	kohli | kohli
0.714286	-13.4268	including | dialog
0.217949	-13.4268	namely, | the
0.444444	-13.4268	only | (i.e.
0.545455	-13.4268	one | annotate
0.235294	-13.4269	for | asr
0.264151	-13.4269	of | disagreement
0.391304	-13.427	approach | limited
0.8	-13.427	nonstandard | into
0.344828	-13.4271	still | performance
0.333333	-13.4271	r | its
0.444444	-13.4272	is | damping
0.165414	-13.4273	input, | the
0.183486	-13.4273	upon | of
0.391304	-13.4274	small | no
0.21519	-13.4274	performed | than
0.0783034	-13.4275	can | more
0.470588	-13.4276	using | produces
0.6	-13.4276	term | computed
0.391304	-13.4277	hidden | state
0.545455	-13.4277	and | keyphrases.
0.217949	-13.4278	a | performance,
0.235294	-13.4279	results | metrics
0.241935	-13.428	the | preserving
0.0747384	-13.4281	in | time
0.714286	-13.4281	different | constructions
0.25	-13.4281	for | genia
0.231884	-13.4281	is | largest
0.209877	-13.4282	we think | the
0.175	-13.4283	recently, | et
0.314286	-13.4284	experiments | obtained
0.3	-13.4284	as | once
0.3	-13.4284	tree | grammars
0.238095	-13.4285	learner | that
0.8	-13.4285	overall, | relations
0.344828	-13.4286	distinguish | different
0.333333	-13.4286	while | contain
0.409091	-13.4286	brand | a
0.292683	-13.4287	to | inputs
0.220779	-13.4287	as | except
0.270833	-13.4288	on | 40
0.132701	-13.4288	of | trees.
0.444444	-13.4288	in | optimization.
0.163121	-13.429	we need | and
0.25	-13.429	dependency | sentences
0.21519	-13.429	framework | an
0.150943	-13.4291	our | several
0.37037	-13.4291	likewise, | to
0.135	-13.4291	mwe | the
0.135	-13.4291	of | much | than
0.197917	-13.4291	this | observed
0.8	-13.4292	recall | 85%
0.8	-13.4292	detecting | negation
0.314286	-13.4293	model | domain.
0.18018	-13.4293	data | then
0.166667	-13.4294	in | (and
0.714286	-13.4294	state-of-the-art | tasks.
0.18018	-13.4294	(2004) | and
0.288889	-13.4294	rank | from
0.292683	-13.4295	certain | than
0.333333	-13.4295	if | it.
0.193878	-13.4295	the | consequence
0.5	-13.4296	tagset | pos
0.0198822	-13.4297	by | the | of
0.6	-13.4297	hmm | probabilities
0.545455	-13.4297	500 | sentences
0.0849421	-13.4299	reason | the
0.357143	-13.4299	first, we | all
0.8	-13.4299	evaluation | insights
0.8	-13.4299	linear | variables.
0.202247	-13.43	a | 200
0.714286	-13.43	project | ontology
0.37037	-13.4301	research | computational
0.103448	-13.4301	experiments | our
0.545455	-13.4301	when | (table
0.241935	-13.4304	highest | all
0.333333	-13.4304	sentences | from | corpus
0.158621	-13.4306	we | labeling
0.421053	-13.4307	in | discourse,
0.292683	-13.4308	sentences | manually
0.228571	-13.4308	d, | a
0.409091	-13.4308	dialect | into
0.6	-13.4308	figure | shows | difference
0.444444	-13.4308	we | ready
0.106061	-13.4308	in | mt
0.117424	-13.4309	= | j
0.333333	-13.431	methods | 2007;
0.141304	-13.4311	from | input
0.714286	-13.4313	combined | achieves
0.0831793	-13.4315	machine translation | the
0.0442571	-13.4315	from | with
0.108626	-13.4317	dictionary | and
0.231884	-13.4318	since | corpus
0.109325	-13.4319	sum | of
0.344828	-13.432	of | intensity
0.8	-13.432	proposition | respectively.
0.270833	-13.4321	was | up
0.206897	-13.4321	not | general
0.157534	-13.4322	of | usage
0.0828729	-13.4323	we used | to
0.270833	-13.4324	feature | type
0.625	-13.4324	solution | guaranteed
0.625	-13.4324	rule-based | lucy
0.625	-13.4324	lemmas | lemmas
0.625	-13.4324	jane | rwth
0.6	-13.4324	score, | value
0.714286	-13.4325	these | coded
0.186916	-13.4325	therefore | with
0.176471	-13.4325	for | threshold
0.0768	-13.4326	candidate | and
0.421053	-13.4326	the | formats.
0.545455	-13.4326	tasks | various
0.127193	-13.4326	of | states
0.120623	-13.4327	scheme | of
0.234375	-13.4327	which | uses
0.8	-13.4327	as: | 1n
0.307692	-13.4328	result | where
0.197802	-13.4329	an | graph
0.245902	-13.4329	semantic | (e.g.,
0.625	-13.433	0-1 | likelihood
0.104046	-13.433	reading | the
0.545455	-13.433	then | list.
0.37037	-13.4331	alignment | ibm
0.15	-13.4331	reported in | and
0.391304	-13.4333	parser | produce
0.0220264	-13.4333	these | the
0.225352	-13.4333	speech | using
0.234375	-13.4333	extraction | events
0.0180924	-13.4334	with | and
0.6	-13.4334	good | being
0.8	-13.4334	1993; | 1997;
0.8	-13.4334	mountain | 94043
0.625	-13.4335	extracted | fragment
0.625	-13.4335	gender | genre
0.0214022	-13.4335	the | on | and
0.0305688	-13.4336	the | results
0.149701	-13.4336	the | system combination
0.37037	-13.4337	◦ | →
0.292683	-13.4337	popular | with
0.276596	-13.4337	show | high
0.202247	-13.4338	that | sufficient
0.714286	-13.4338	could | aid
0.209877	-13.4338	with | state-of-the-art
0.0511204	-13.4339	for | training
0.122951	-13.4339	product | and
0.6	-13.434	(the | 2
0.545455	-13.434	school | salt
0.545455	-13.434	graph-based | models
0.5	-13.434	often | low
0.37037	-13.4341	an | 2003).
0.254545	-13.4341	of | l2
0.8	-13.4341	knowledge | enhance
0.0816697	-13.4342	us | a
0.276596	-13.4342	significant, | the
0.323529	-13.4343	in | development.
0.0734214	-13.4343	second | to
0.409091	-13.4344	rank | score
0.225352	-13.4345	restricted | in
0.228571	-13.4345	it | performance.
0.314286	-13.4346	phrasal | are
0.106707	-13.4346	their | or
0.0863454	-13.4347	however, | with
0.144509	-13.4347	best | system
0.0939759	-13.4347	the | taking
0.209877	-13.4347	and | attachment
0.156463	-13.4348	showing | of
0.0964467	-13.4349	for | input
0.231884	-13.4349	for | filtering
0.209877	-13.4349	features | them
0.8	-13.4349	range | comments
0.204545	-13.435	follows | in
0.186916	-13.435	on | subset
0.714286	-13.4351	object, | object
0.146199	-13.4351	range | for
0.0813743	-13.4351	2005; | et
0.344828	-13.4351	although | accuracy
0.625	-13.4352	weighting | query-comp
0.086	-13.4355	and | error
0.37037	-13.4357	automatically | rules
0.175	-13.4357	data | large
0.333333	-13.4357	2 | very
0.357143	-13.4357	this | phonetic
0.121457	-13.4358	that | patterns
0.625	-13.4358	48.5 | on
0.625	-13.4358	output | next
0.225352	-13.4359	space, | and
0.090708	-13.4359	this | any
0.176471	-13.436	alignment | source
0.245902	-13.436	grant | research
0.158621	-13.436	direction | to
0.095122	-13.4361	a | often
0.116364	-13.4361	a | languages.
0.04694	-13.4362	a | feature
0.119691	-13.4362	providing | the
0.204545	-13.4363	the | (not
0.421053	-13.4363	approach | manual
0.136364	-13.4363	is | supervised
0.5	-13.4363	(2003) | from
0.714286	-13.4364	is | till
0.202247	-13.4365	of | propagation
0.264151	-13.4365	we | verify
0.178571	-13.4367	for | description
0.444444	-13.4367	turns | claims
0.2125	-13.4368	of | agreement.
0.14881	-13.4369	is | unlabeled
0.625	-13.4369	mbr | byrne,
0.357143	-13.4369	weight | vector
0.714286	-13.437	made | automatically
0.2	-13.437	exhibit | of
0.151899	-13.437	a | relations.
0.8	-13.437	spatial | preposition
0.8	-13.437	proposed | enhance
0.175	-13.4371	following | our
0.391304	-13.4371	human | conference
0.245902	-13.4371	bilingual | word
0.6	-13.4372	issues | issues
0.6	-13.4372	four | rules
0.181818	-13.4374	large | training
0.391304	-13.4375	in | files.
0.333333	-13.4375	semantic | disambiguation
0.375	-13.4375	phrase-based | but
0.137056	-13.4376	maximum | is
0.714286	-13.4376	recording | it
0.5	-13.4376	process, | used
0.307692	-13.4377	which | created
0.152866	-13.4377	to | predictions
0.6	-13.4377	synchronous | parsing
0.8	-13.4377	tau | score
0.8	-13.4377	crf | resources.
0.444444	-13.4379	voting | that
0.444444	-13.4379	for | parsers,
0.625	-13.438	many | short
0.231884	-13.438	unigram | is
0.136364	-13.438	the | is | 3.
0.095122	-13.4381	the | actual
0.119231	-13.4382	m | in
0.421053	-13.4382	for | hypothesis,
0.165414	-13.4382	acquired | the
0.231884	-13.4382	kind | that
0.147929	-13.4382	the | illustrated
0.375	-13.4382	performance | noun
0.264151	-13.4382	orientation | is
0.166667	-13.4383	for | spoken
0.6	-13.4383	model | cfg
0.130233	-13.4383	estimate | a
0.466667	-13.4384	experiments | assess
0.466667	-13.4384	organize | a
0.375	-13.4385	so | every
0.241935	-13.4388	file | in
0.714286	-13.4389	item | .
0.545455	-13.4389	preprocessing | sentence
0.545455	-13.4389	improvement | precision.
0.545455	-13.4389	both | in-domain
0.5	-13.439	approach | outputs
0.137755	-13.4391	voice | the
0.059387	-13.4391	four | the
0.625	-13.4391	about | day
0.344828	-13.4391	hyperedge | a
0.444444	-13.4391	evaluation | models.
0.8	-13.4391	correspondence | bullets
0.0605159	-13.4391	the | weight
0.153846	-13.4392	combination | on
0.0668287	-13.4392	better | of
0.193878	-13.4392	of | spatial
0.112628	-13.4394	or | one
0.225352	-13.4394	translation | some
0.545455	-13.4394	string | kb
0.545455	-13.4394	distribution | topics.
0.0815217	-13.4395	the | alignments
0.121457	-13.4395	argument | is
0.357143	-13.4395	phrases | words.
0.264151	-13.4396	alternative | from
0.173554	-13.4398	types, | and
0.6	-13.4399	age | classification
0.375	-13.4399	2008), | which
0.152866	-13.44	compute | =
0.323529	-13.4401	italian, | and
0.0713306	-13.4401	methods | in
0.159722	-13.4402	two | when
0.121951	-13.4405	wikipedia | a
0.466667	-13.4405	took | took
0.323529	-13.4406	we | going
0.157534	-13.4406	this | full
0.8	-13.4406	therefore | needed
0.8	-13.4406	t. | c.
0.3	-13.4407	automatically | these
0.145349	-13.4407	is | variable
0.421053	-13.4408	details | how
0.276596	-13.4408	features | relevant
0.5	-13.4408	j., | m.
0.254545	-13.4409	votes | the
0.3	-13.441	from | europarl
0.25	-13.441	3.1 | data
0.270833	-13.4411	for | test.
0.109635	-13.4411	4 | 2
0.357143	-13.4411	hence, | have
0.162963	-13.4411	tool | a
0.2125	-13.4412	since | would
0.3	-13.4413	that | learners
0.234375	-13.4413	before | on
0.133005	-13.4414	performance | data
0.333333	-13.4414	example, | “the
0.155405	-13.4414	1) | is
0.714286	-13.4415	transformations | syntactic
0.176991	-13.4416	a | behavior
0.0994624	-13.4417	in | examples
0.254545	-13.4418	a | is:
0.545455	-13.4418	try | identify
0.625	-13.4419	linguistic | twitter
0.625	-13.4419	transfer | transferring
0.625	-13.4419	accuracies | significantly
0.344828	-13.4419	from | terms.
0.264151	-13.442	discovery | of
0.8	-13.442	however, | inferior
0.8	-13.442	examples | selected.
0.12844	-13.4421	the | consistently
0.225352	-13.4422	feature | weight
0.375	-13.4422	whole | it
0.151899	-13.4423	from | possible
0.0739645	-13.4425	this paper, | a
0.143678	-13.4425	in | basic
0.625	-13.4425	step | systems.
0.625	-13.4425	against | rouge
0.409091	-13.4425	the | emotions.
0.37037	-13.4427	made | available
0.444444	-13.4427	their | senses
0.444444	-13.4427	in | sports
0.264151	-13.4427	method | proposed
0.714286	-13.4428	labeling | information.
0.714286	-13.4428	pass | any
0.25	-13.4428	has | computational
0.159722	-13.4428	z | a
0.188119	-13.4429	with | complex
0.282609	-13.4429	only | events
0.195652	-13.4429	4.2 | our
0.307692	-13.443	experiments | previous
0.625	-13.443	true | distribution
0.5	-13.443	2010b) | a
0.6	-13.4431	online | multiple
0.6	-13.4431	overall | distribution
0.125	-13.4431	in | highest
0.123457	-13.4432	since | is | we
0.375	-13.4432	from | fit
0.264151	-13.4432	compounds, | the
0.064117	-13.4433	large | and
0.421053	-13.4435	show | full
0.0862423	-13.4436	for | no
0.375	-13.4436	train | sets
0.344828	-13.4437	must | its
0.147059	-13.4437	numbers | in
0.391304	-13.4438	distribution | observed
0.12844	-13.4438	gain | of
0.082397	-13.4438	additional | and
0.314286	-13.4439	summary | word
0.421053	-13.4439	given | seed
0.149068	-13.444	5 | our | and
0.0848126	-13.444	is | input
0.228571	-13.4441	test | development
0.8	-13.4442	greedily | best
0.37037	-13.4443	matrix | space
0.0392157	-13.4443	sentence | and
0.307692	-13.4444	one | consider
0.3	-13.4445	and | matrices
0.197802	-13.4445	near | the
0.222222	-13.4447	was | annotation
0.625	-13.4447	entries | table.
0.409091	-13.4447	that | users’
0.259259	-13.4447	new | then
0.0686528	-13.4448	very | and
0.292683	-13.4449	observed | at
0.391304	-13.4449	further | its
0.18018	-13.4449	general | which
0.8	-13.4449	transition | next
0.114695	-13.4451	texts | in
0.466667	-13.4451	for | conversation,
0.138298	-13.4451	the | signature
0.357143	-13.4453	show | patterns
0.149068	-13.4454	furthermore, | is
0.0312087	-13.4454	one | the | of
0.409091	-13.4454	the | robust.
0.126638	-13.4454	has | at
0.391304	-13.4456	filtering | data
0.8	-13.4456	level | correctness
0.8	-13.4456	works | latent
0.254545	-13.4457	common | words
0.545455	-13.4457	involve | number
0.5	-13.4457	in | antecedents.
0.314286	-13.4459	used | 2008),
0.029502	-13.4459	the | sentences
0.095	-13.446	application | the
0.234375	-13.4461	perceptron | a
0.143678	-13.4462	parsing | by
0.228571	-13.4462	a | baseline,
0.195652	-13.4462	which | meaning
0.0547833	-13.4463	the | query
0.157534	-13.4463	task | data
0.8	-13.4463	correct, | percentage
0.3	-13.4464	is | flexible
0.170732	-13.4465	a | map
0.421053	-13.4465	of | provided.
0.333333	-13.4465	two | information.
0.375	-13.4466	her | her
0.0752688	-13.4467	of | content
0.0830189	-13.4468	the | temporal
0.466667	-13.4468	identified, | are
0.307692	-13.4469	for | machine translation,
0.625	-13.447	studies | portuguese
0.625	-13.447	readers, | with
0.234375	-13.447	is | negative.
0.444444	-13.4471	reverb | precision
0.409091	-13.4472	reordering | 2010).
0.466667	-13.4472	for | coreference.
0.6	-13.4474	automatically | novel
0.625	-13.4475	parsing | fan-out
0.625	-13.4475	predictions | those
0.219178	-13.4476	names | are
0.123932	-13.4476	and | expressions
0.259259	-13.4477	weights | feature
0.545455	-13.4477	automated | scores
0.344828	-13.4478	simple | while
0.8	-13.4478	goal | retrieve
0.8	-13.4478	these | exclusive
0.147059	-13.4479	an | model.
0.124464	-13.4479	the | orientation
0.6	-13.448	while | decrease
0.6	-13.448	p(q|d)
0.466667	-13.448	grammatical | sentence.
0.5	-13.448	allows | apply
0.202247	-13.4481	probabilities | from
0.130841	-13.4481	denotes the | of
0.0791738	-13.4481	the | (or
0.545455	-13.4482	low, | al.,
0.265306	-13.4484	each | added
0.409091	-13.4484	extraction | structured
0.209877	-13.4484	to | summary.
0.5	-13.4484	developed | new
0.714286	-13.4485	across | overall
0.714286	-13.4485	augmenting | set
0.6	-13.4485	with | adult
0.8	-13.4485	gamedate | gamewinner
0.143678	-13.4486	labeled | with
0.10355	-13.4486	useful | and
0.375	-13.4486	performance | local
0.391304	-13.4487	hierarchical | syntactic
0.391304	-13.4487	even | order
0.183486	-13.4487	of | alignment.
0.142045	-13.4489	how | it
0.5	-13.4489	that | tools.
0.0330455	-13.449	two | and | the
0.323529	-13.4491	where | means
0.234375	-13.4491	most | about
0.130841	-13.4492	as | original
0.714286	-13.4492	pinyin | english
0.714286	-13.4492	rules | detecting
0.254545	-13.4492	all | reported
0.625	-13.4492	domain | assigned
0.142857	-13.4492	posts | the
0.207317	-13.4494	the | gloss
0.165414	-13.4494	bayesian | a
0.197802	-13.4495	the | (de
0.357143	-13.4495	and | insights
0.545455	-13.4497	domain | jst
0.625	-13.4498	with | customers
0.625	-13.4498	p(t|s)
0.333333	-13.4498	to | reviews,
0.8	-13.4499	span | 20
0.285714	-13.45	that | conversational
0.314286	-13.4501	this | limit
0.0647399	-13.4501	that | semantic
0.241935	-13.4501	projecting | the
0.127854	-13.4502	making | to
0.5	-13.4502	we collect | from
0.0784983	-13.4504	unsupervised | and
0.625	-13.4504	morphological | grammatical
0.184466	-13.4505	of | crucial
0.186275	-13.4505	of | brown
0.157534	-13.4506	would | if
0.209877	-13.4507	while | sentences
0.5	-13.4507	take | following
0.357143	-13.4508	then | sum
0.625	-13.4509	90 | comparison
0.314286	-13.451	at | thus
0.122449	-13.451	a | trigger
0.3	-13.451	w | vector
0.155405	-13.451	the | body
0.323529	-13.4511	we | tokenization
0.6	-13.4512	probabilities | emission
0.285714	-13.4513	increase | which
0.0864198	-13.4514	of | four
0.0924171	-13.4514	detection | of
0.117424	-13.4515	product | of
0.292683	-13.4515	these | probabilities
0.225352	-13.4515	confirms | of
0.3	-13.4516	comparable | pairs
0.0952381	-13.4518	novel | of
0.6	-13.4518	verbs, | adjectives
0.162963	-13.4518	(the | for
0.466667	-13.4518	grammar | syntax
0.254545	-13.4519	for | phrase.
0.209877	-13.4519	also | baseline
0.5	-13.452	its | processing
0.241935	-13.4521	interactive | and
0.625	-13.4521	topic | attribute
0.625	-13.4521	head | it.
0.116981	-13.4521	annotated | as
0.245614	-13.4522	li, | and
0.131707	-13.4523	models, | a
0.276596	-13.4523	firstly, | of
0.121951	-13.4523	phrases | as
0.714286	-13.4524	l-lda | topic
0.333333	-13.4525	to | issues.
0.181818	-13.4525	web | that
0.323529	-13.4526	in | regulation
0.01083	-13.4526	a | a
0.230769	-13.4526	(2009) | al.
0.314286	-13.4527	bottleneck | the
0.117871	-13.4527	the | hmm
0.121457	-13.4527	of | real
0.238095	-13.4528	can | certain
0.25	-13.4528	evaluation | metrics
0.0981432	-13.4528	it | its
0.158621	-13.4528	all | ∈
0.8	-13.4528	effectiveness | identification.
0.6	-13.4529	graph-based | 2010).
0.391304	-13.453	followed by | or
0.265306	-13.453	a | vertical
0.207317	-13.453	x2 | the
0.170732	-13.4531	a | actually
0.714286	-13.4531	2.2 | prior
0.109635	-13.4531	could | that
0.466667	-13.4531	training | takes
0.178571	-13.4532	the | receive
0.625	-13.4532	useful | evaluating
0.109272	-13.4532	there | no | of
0.08742	-13.4532	pair | is
0.545455	-13.4532	3 | 5,
0.109272	-13.4533	is | frequency
0.0834951	-13.4535	advantage | the
0.197802	-13.4535	between | words.
0.8	-13.4536	whole | individual
0.8	-13.4536	senses | disambiguated
0.8	-13.4536	might | something
0.12963	-13.4537	article | and
0.188119	-13.4537	is | nature
0.421053	-13.4538	combination | development
0.150943	-13.4539	no | than
0.292683	-13.4539	also | help
0.0315132	-13.4539	also | of
0.333333	-13.454	a | segment.
0.0638298	-13.454	we then | the
0.391304	-13.4541	in | challenge.
0.135678	-13.4542	to | summarization
0.204819	-13.4542	model | phrase-based
0.204819	-13.4542	type, | and
0.160584	-13.4543	from | baseline
0.8	-13.4543	without | supervision.
0.8	-13.4543	× | given
0.8	-13.4543	frontier | rule
0.8	-13.4543	office | director
0.5	-13.4543	capturing | are
0.714286	-13.4544	3.1 | categories
0.186275	-13.4544	3 | approach
0.6	-13.4545	which | sequences.
0.225352	-13.4546	the | switch
0.357143	-13.4546	show | example
0.314286	-13.4547	within | other
0.216216	-13.4547	example | 1.
0.625	-13.4549	can | easier.
0.714286	-13.455	synonymy | lexical
0.344828	-13.455	quality | terms
0.8	-13.455	incoherent | relation
0.0707351	-13.4551	included | the
0.375	-13.4551	each | rest
0.545455	-13.4551	∂l | =
0.421053	-13.4553	(culotta | et
0.444444	-13.4555	how | multiple
0.466667	-13.4556	may | others
0.421053	-13.4557	strategy | other
0.305556	-13.4557	siblings | the
0.234375	-13.4557	the | actions,
0.0728863	-13.4558	sentence | as
0.202247	-13.4558	collocation | and
0.144509	-13.4558	in | always
0.8	-13.4558	lipitor, | azithromycin,
0.170732	-13.4559	parallel | data
0.323529	-13.4559	a | effort.
0.444444	-13.4559	this | properly
0.150943	-13.456	we have | this
0.146199	-13.456	computation | of
0.625	-13.456	approach | 1999).
0.0631229	-13.4561	linear | the
0.238095	-13.4561	performance | features.
0.421053	-13.4561	for | semantics,
0.421053	-13.4561	combined | new
0.545455	-13.4561	trigram | smoothing
0.219178	-13.4563	from | times
0.265306	-13.4563	included | or
0.0486697	-13.4563	the | parse
0.0376553	-13.4564	our | on
0.0529141	-13.4564	possible | of
0.466667	-13.4565	output: | set
0.391304	-13.4566	aspects | aspect
0.166667	-13.4567	each | values
0.161765	-13.4567	total | by
0.120968	-13.4567	to | context.
0.421053	-13.4568	over | time.
0.466667	-13.4569	39 | 1
0.714286	-13.457	partial | (moschitti,
0.0790378	-13.457	patterns | and
0.0322857	-13.457	is | or
0.0503849	-13.4571	two | for
0.6	-13.4572	however, | poor
0.207317	-13.4575	the | preserve
0.714286	-13.4576	algorithms | solving
0.0581614	-13.4576	being | the
0.545455	-13.4576	is | prominently
0.123932	-13.4577	reflect | of
0.150943	-13.4578	interest | of
0.15	-13.4578	practice | the
0.0555556	-13.4579	we | that | by
0.0742424	-13.4579	clusters | the
0.8	-13.4579	with | gigabytes
0.126697	-13.4581	greater | of
0.104167	-13.4581	that | linguistic
0.109635	-13.4582	that | make
0.115242	-13.4582	binary | a
0.156463	-13.4582	complete | in
0.391304	-13.4583	4.2.1 | of
0.714286	-13.4583	reflects | fact
0.625	-13.4583	likely | become
0.259259	-13.4583	and | graphs.
0.162963	-13.4583	attention | and
0.186275	-13.4583	to | running
0.186275	-13.4584	simple | model
0.238095	-13.4585	results | system.
0.207317	-13.4585	this | line
0.545455	-13.4586	• | choose
0.0834951	-13.4587	for example, | that
0.0627706	-13.4587	corpus, | the
0.8	-13.4587	token’s | to
0.444444	-13.4588	that | units.
0.714286	-13.4589	contain | elements
0.344828	-13.459	represents | using
0.292683	-13.4593	are | label.
0.292683	-13.4593	the | (4).
0.168	-13.4593	experiments | test
0.5	-13.4593	p(e|f)
0.5	-13.4593	community | it
0.142045	-13.4594	when | same
0.625	-13.4595	5000 | avenue
0.241935	-13.4596	and | judged
0.265306	-13.4597	4.2 | each
0.0968586	-13.4597	it is | it
0.5	-13.4598	algorithms | 2010;
0.305556	-13.4599	are | 7.
0.108911	-13.4599	5 | 0
0.241935	-13.46	whereas | with
0.344828	-13.46	have | whereas
0.0681234	-13.4601	rules | a
0.156463	-13.4601	it seems | the
0.8	-13.4601	use | (m
0.8	-13.4601	learners | second
0.259259	-13.4602	using | local
0.276596	-13.4604	are | groups
0.15942	-13.4604	would | have
0.625	-13.4606	found | written
0.19	-13.4606	results | similar
0.172131	-13.4607	set | based on
0.207317	-13.4607	complexity, | the
0.127273	-13.4608	system | bleu
0.230769	-13.4608	corpus | articles
0.181818	-13.4608	allows | by
0.265306	-13.4609	pattern | noun
0.276596	-13.4609	rules, | as
0.139037	-13.4609	studies | et
0.254545	-13.461	and | heuristics
0.2125	-13.461	are | (for
0.545455	-13.4611	10%, | of
0.545455	-13.4611	problem | ordering
0.444444	-13.4612	distributional | semantically
0.444444	-13.4612	our | builds
0.0730253	-13.4613	the | of | grammar
0.0814815	-13.4615	of | web
0.444444	-13.4616	here, | model
0.8	-13.4616	were | asd
0.234375	-13.4617	a | distortion
0.305556	-13.4618	are | nature
0.0171092	-13.462	a | that
0.714286	-13.4622	all | simultaneously.
0.0834951	-13.4623	this paper | for
0.625	-13.4623	sample | label
0.161765	-13.4623	studies | that
0.106918	-13.4624	made | to
0.292683	-13.4625	of | constitutes
0.466667	-13.4625	most | ours
0.5	-13.4625	they | generation
0.15942	-13.4626	resource | in
0.136842	-13.4626	assessment | the
0.0376618	-13.4627	of | semantic
0.305556	-13.4627	an | stage
0.276596	-13.4629	triple | and
0.625	-13.4629	weak | agreement
0.0757098	-13.4629	we will | to
0.197802	-13.463	have | (e.g.,
0.265306	-13.4631	and | compositionality
0.146341	-13.4631	learns | the
0.8	-13.4631	various | techniques.
0.444444	-13.4632	as | community.
0.270833	-13.4634	preferred | and
0.5	-13.4634	much | without
0.141243	-13.4635	a | utterance
0.545455	-13.4636	that | under-resourced
0.545455	-13.4636	performs | selection
0.075	-13.4637	also | from
0.222222	-13.4637	performance | proposed
0.245614	-13.4637	complex | which
0.0587103	-13.4637	research | and
0.357143	-13.4638	mentions | only
0.357143	-13.4638	we | modeling.
0.3	-13.4639	these | resources
0.113971	-13.4639	to | concept
0.102639	-13.4639	benefit | the
0.175439	-13.464	that | achieved
0.00913673	-13.464	the | for
0.202381	-13.464	during | on
0.0727003	-13.464	the | matching
0.5	-13.4643	naacl | on
0.5	-13.4643	alignments | pairs.
0.6	-13.4644	(where | or
0.112281	-13.4645	of | million
0.8	-13.4645	duc | summaries
0.222222	-13.4647	of | central
0.276596	-13.4647	combinations | that
0.195652	-13.4647	syntactic | words
0.186275	-13.4647	for | n-grams
0.344828	-13.465	the | 4.4.
0.122363	-13.4651	moses | the
0.545455	-13.4651	thus, | complexity
0.165354	-13.4652	the | trend
0.0314724	-13.4652	an | is
0.182692	-13.4653	in | little
0.265306	-13.4653	reverse | in
0.0122168	-13.4653	the | by
0.8	-13.4653	proposed | jiang
0.8	-13.4653	features | assessment.
0.8	-13.4653	resulting | extended
0.191489	-13.4654	the | factored
0.714286	-13.4655	grammatical | (e.g.
0.0442865	-13.4655	the | part
0.466667	-13.4655	best | threshold
0.323529	-13.4656	phosphorylation | protein
0.254545	-13.4656	are | sources
0.125	-13.4656	2006), | and
0.344828	-13.4657	in | compiled
0.5	-13.4657	a | these.
0.0926366	-13.4658	contexts | the
0.314286	-13.4659	darpa | and
0.165354	-13.4659	to | suitable
0.466667	-13.4659	supervised | techniques
0.25	-13.466	had | than
0.0486202	-13.466	us | the
0.172131	-13.4661	some | features
0.292683	-13.4662	lines | with
0.391304	-13.4662	from | trees,
0.391304	-13.4662	using | might
0.204819	-13.4662	the | localization
0.270833	-13.4664	inputs | a
0.154362	-13.4664	contribution | a
0.314286	-13.4665	can | you
0.213333	-13.4665	the | characterize
0.230769	-13.4665	domain-specific | and
0.466667	-13.4667	pyp | a
0.444444	-13.4669	other | phrases.
0.146341	-13.467	we | introduced
0.10559	-13.4671	the | compound
0.6	-13.4671	is | accurately.
0.0820896	-13.4671	of | events
0.111888	-13.4672	tagging | and
0.0989011	-13.4672	types | are
0.147239	-13.4673	automatically | by
0.238095	-13.4674	need | data
0.254545	-13.4675	hence, | as
0.8	-13.4675	filtered | original
0.225352	-13.4676	the | abstractness
0.323529	-13.4677	segmentation | pos
0.3	-13.4677	transliteration | as
0.6	-13.4677	methods | adapt
0.391304	-13.468	further | additional
0.207317	-13.468	on | real
0.2	-13.468	our | version
0.714286	-13.4681	selects | similar
0.625	-13.4681	demonstrated | statistical
0.444444	-13.4681	several | measures
0.270833	-13.4682	(2011) | in
0.8	-13.4682	(e.g., | popular
0.129187	-13.4684	find | with
0.125561	-13.4685	the | know
0.213333	-13.4685	that | consistently
0.625	-13.4686	containing | names
0.625	-13.4686	indicate | strategy
0.625	-13.4686	make | hard
0.191489	-13.4686	here | an
0.444444	-13.4686	conversational | are
0.545455	-13.4686	use | source-side
0.333333	-13.4687	best | result
0.357143	-13.4687	nn | nmod
0.27907	-13.4688	5.2.2 | the
0.6	-13.4688	dictionary | created
0.102639	-13.4688	and | discourse
0.137566	-13.469	the | poor
0.0398406	-13.469	the | rules
0.357143	-13.469	method | languages.
0.207317	-13.4691	the | reliably
0.285714	-13.4693	such | automatically
0.0542317	-13.4693	to | text
0.156463	-13.4693	abstract this paper | the
0.714286	-13.4694	number | exceed
0.714286	-13.4694	|g
0.238095	-13.4694	details | our
0.122881	-13.4696	(2007) | of
0.357143	-13.4697	the | chunk.
0.625	-13.4698	o | onull
0.466667	-13.4698	such | resource
0.117188	-13.4701	is | an | or
0.545455	-13.4701	(smith | 2009;
0.545455	-13.4701	narrative | document
0.5	-13.4703	words | phrases,
0.6	-13.4704	imt | translation
0.314286	-13.4705	feature | example
0.545455	-13.4706	dataset | authors
0.545455	-13.4706	(1996) | with
0.466667	-13.4706	the | extractors.
0.714286	-13.4707	expected | loss
0.164062	-13.4708	by | who
0.333333	-13.4708	the | line.
0.166667	-13.4709	dimensions | and
0.131707	-13.471	the | coefficient
0.0810313	-13.471	corpus, | of
0.153333	-13.471	to | noise
0.184466	-13.4711	from | like
0.245614	-13.4711	measure | all
0.344828	-13.4711	80 | accuracy
0.333333	-13.4711	analysis | annotation
0.191489	-13.4711	lexical | our
0.26	-13.4712	features | particular
0.153333	-13.4712	this | parser
0.8	-13.4712	yi, | training
0.8	-13.4712	s. | y.
0.138298	-13.4713	group | to
0.285714	-13.4714	from | taking
0.285714	-13.4714	although | no
0.245614	-13.4714	summationdisplay | log
0.164062	-13.4716	a | exact
0.155405	-13.4716	for | already
0.545455	-13.4716	has | communicative
0.357143	-13.4717	they | arguments
0.197802	-13.4718	comments | of
0.444444	-13.4718	other | types.
0.138889	-13.4719	orientation | the
0.421053	-13.4719	(2010), | which
0.8	-13.4719	architecture | includes
0.8	-13.4719	textual | (rte).
0.141243	-13.472	three | each
0.039823	-13.472	the | of | into
0.276596	-13.4721	only | thus
0.0453502	-13.4721	task | and
0.0921986	-13.4722	test | on
0.238095	-13.4723	mr | the
0.124444	-13.4723	work on | of
0.3	-13.4724	surface | or
0.8	-13.4726	help | smt.
0.5	-13.4726	problem | probabilistic
0.421053	-13.4727	used | trigger
0.357143	-13.4727	is | discrete
0.6	-13.4727	phosphorylation | vav
0.0783972	-13.4728	approaches | to
0.466667	-13.4728	distributions, | for
0.0519187	-13.4729	and | models
0.222222	-13.4729	measure | semantic
0.12844	-13.4729	a | computational
0.195652	-13.473	which | 2009).
0.234375	-13.4731	it | relatively
0.125	-13.4731	distinguish | and
0.175439	-13.4732	that | polarity
0.202381	-13.4732	group | with
0.344828	-13.4733	where | phrases
0.714286	-13.4734	most | empty
0.4	-13.4734	for | days
0.118577	-13.4734	of | interaction
0.245614	-13.4734	system | high
0.204819	-13.4734	our | provided
0.8	-13.4734	weak | metric
0.285714	-13.4735	understood | a
0.124444	-13.4735	also | language
0.26	-13.4736	text | given
0.112727	-13.4736	2011 | pages
0.545455	-13.4737	a | court
0.545455	-13.4737	therefore, | possible
0.188119	-13.4738	may | if
0.0948718	-13.4738	to | another
0.178571	-13.4739	this | binary
0.184466	-13.4739	correctly | as
0.333333	-13.4739	if | nps
0.146341	-13.4739	each | context
0.714286	-13.474	once | automatically
0.147239	-13.4741	examine | and
0.466667	-13.4741	(8) | each
0.305556	-13.4742	of | headlines
0.305556	-13.4742	and | (2001)
0.625	-13.4744	ˆy | argmax
0.0416465	-13.4744	to | translation
0.193548	-13.4745	by | relative
0.466667	-13.4745	during | generation
0.4	-13.4746	model | estimates
0.227273	-13.4746	second column | the
0.27907	-13.4748	to capture | between
0.142045	-13.4748	lda | and
0.333333	-13.4748	in | decreasing
0.193548	-13.4748	for | 24
0.230769	-13.4748	to | value.
0.146341	-13.4749	in | alternative
0.150327	-13.4749	with | current
0.121849	-13.475	in | base
0.136126	-13.475	of | 50
0.323529	-13.4752	of | data).
0.333333	-13.4752	parallel | sentences.
0.36	-13.4754	requires | set
0.466667	-13.4754	continuing | of
0.333333	-13.4755	process | text
0.391304	-13.4756	a | problem:
0.238095	-13.4756	use | without
0.227273	-13.4756	given | following
0.8	-13.4756	behavior | available
0.8	-13.4756	credit | not
0.0877193	-13.4757	2.2 | the
0.102719	-13.4758	our | model | a
0.175439	-13.4758	ordering | in
0.197802	-13.4758	experts | of
0.207317	-13.4758	to | techniques.
0.125	-13.4758	fact, | of
0.0634921	-13.4759	document | and
0.5	-13.4759	with | links.
0.5	-13.4759	makes | good
0.292683	-13.476	shortest | in
0.117188	-13.476	now | a
0.357143	-13.476	the | spanish-english
0.219178	-13.4761	generally | that
0.241379	-13.4761	low | which
0.333333	-13.4761	information | make
0.545455	-13.4762	association | ratio
0.154362	-13.4763	in | 2007)
0.227273	-13.4763	reviews | on
0.166667	-13.4764	a | objects
0.209877	-13.4765	affect | in
0.172414	-13.4767	will | their
0.714286	-13.4767	extracted | table.
0.625	-13.4767	network | web
0.625	-13.4767	using | stopword
0.170732	-13.4768	a | ambiguous
0.292683	-13.4768	(y | |x
0.4	-13.4768	instances | labeled
0.144578	-13.4768	is: | the
0.0870488	-13.477	the | global
0.265306	-13.4771	applied | words
0.0762852	-13.4771	and | scores | the
0.113553	-13.4772	to | future
0.323529	-13.4773	= | j)
0.5	-13.4773	sentence | redundancy
0.166667	-13.4775	distribution | words
0.227273	-13.4775	john | to
0.188119	-13.4776	on | topics
0.138122	-13.4776	means | for
0.158273	-13.4776	pairwise | of
0.164062	-13.4777	in | increases
0.545455	-13.4777	velikovich | al.
0.25	-13.4778	further | two
0.227273	-13.4778	(3), | the
0.11583	-13.4778	conditions | the
0.8	-13.4778	summary | summary,
0.625	-13.4779	likely | results.
0.209877	-13.4779	where | may
0.285714	-13.478	above, | are
0.714286	-13.478	projectable | translation
0.357143	-13.478	and | effect.
0.2	-13.478	and | frames
0.466667	-13.478	causality | between
0.126126	-13.4781	this | unsupervised
0.186275	-13.4783	with | made
0.270833	-13.4784	for | reasons, | the
0.466667	-13.4784	which | observations
0.107492	-13.4784	involves | the
0.391304	-13.4785	module | into
0.241379	-13.4786	solid | the
0.8	-13.4786	usage | within
0.8	-13.4786	2007) | discontinuous
0.8	-13.4786	could | exploited
0.714286	-13.4787	morphologically | using
0.245614	-13.4787	the | subset.
0.444444	-13.4789	into | parts:
0.4	-13.479	or | online
0.162791	-13.4791	groups | in
0.625	-13.4791	have | medium
0.144578	-13.4791	view | a
0.5	-13.4791	approach | acquisition
0.128571	-13.4793	may | an
0.238095	-13.4793	are | types.
0.8	-13.4793	1, | 9,
0.26	-13.4794	the | ontology,
0.234375	-13.4794	binary | each
0.0369657	-13.4795	two | in
0.122363	-13.4795	to | problem.
0.204819	-13.4795	model | state
0.0946292	-13.4796	the | entropy
0.110727	-13.4796	since | it
0.195652	-13.4796	we | employ
0.0997151	-13.4796	of | e.g.
0.26	-13.4797	approach | relation
0.168	-13.4797	to | phrases.
0.0628507	-13.4797	tag | the
0.134715	-13.4797	the | user’s
0.625	-13.4797	multiple-beam | clark
0.227273	-13.4797	a | search.
0.444444	-13.4797	0.15 | 0.3
0.466667	-13.4797	2.1 | over
0.210526	-13.4797	with | coverage
0.285714	-13.4799	in | judgment
0.305556	-13.4799	that | superior
0.223881	-13.4799	canonical | of
0.178571	-13.48	global | for
0.113971	-13.48	trained | data
0.158273	-13.48	approach | words
0.25	-13.4801	we | specify
0.333333	-13.4801	detection | task
0.191489	-13.4801	median | the
0.444444	-13.4801	7th | and
0.5	-13.4801	find | useful
0.5	-13.4801	β | 0
0.1	-13.4802	of | clusters
0.466667	-13.4802	contains | nodes
0.545455	-13.4803	method | way.
0.285714	-13.4804	contexts | or
0.216216	-13.4804	that | strongly
0.104938	-13.4804	this | no
0.391304	-13.4807	relation | existing
0.391304	-13.4807	but | selection
0.714286	-13.4807	state, | then
0.714286	-13.4807	0.84 | -0.00
0.714286	-13.4807	she | nexavar
0.714286	-13.4807	return | pseudocode
0.714286	-13.4807	sys4 | live
0.714286	-13.4807	0.61 | 0.56
0.714286	-13.4807	lp | lp
0.714286	-13.4807	0.65 | 0.50
0.714286	-13.4807	genome | managed
0.714286	-13.4807	medlda | medlda
0.714286	-13.4807	browser | wiki
0.714286	-13.4807	.4 | .7
0.714286	-13.4807	ptop | pdr
0.0529644	-13.4807	relations | the
0.36	-13.4808	set | describe
0.265306	-13.4808	recall | only
0.545455	-13.4808	wordnet | wikipedia.
0.8	-13.4808	currently | standard
0.193548	-13.4809	the | uniformly
0.270833	-13.481	substring | the
0.175439	-13.481	training | sets
0.113139	-13.481	of | representations
0.130435	-13.4811	the | hierarchy | the
0.129808	-13.4811	first | sentence
0.117647	-13.4812	consider | that
0.207317	-13.4812	4.2 | from
0.106796	-13.4812	out | for
0.4	-13.4813	whereas | system
0.173913	-13.4813	the | null.
0.219178	-13.4814	train | each
0.106452	-13.4815	the | noisy
0.314286	-13.4816	very | large
0.254902	-13.4817	date | and
0.421053	-13.4817	following | equation
0.123894	-13.4817	that | through
0.344828	-13.482	surprisingly, | in
0.195652	-13.482	were | so
0.323529	-13.4821	comma | of
0.391304	-13.4822	is | person.
0.285714	-13.4823	different | sources
0.285714	-13.4823	university | edinburgh
0.545455	-13.4823	posted | for
0.8	-13.4823	discussed | sec.
0.8	-13.4823	positions | through
0.106109	-13.4824	which | sentence
0.129187	-13.4824	constructed | and
0.0771812	-13.4824	one | from
0.5	-13.4824	set | i.
0.444444	-13.4826	briefly | related
0.165354	-13.4828	changes | to
0.219178	-13.4828	necessarily | or
0.265306	-13.4828	metric | correlation
0.234375	-13.4829	present | it
0.5	-13.4829	all | category,
0.4	-13.4831	used | fixed
0.4	-13.4831	instances | class.
0.8	-13.4831	threshold | validation
0.8	-13.4831	tool | 2001).
0.545455	-13.4833	pair | equivalent
0.197802	-13.4836	detailed | for
0.625	-13.4837	statistical | evidence
0.625	-13.4837	associating | lexical
0.087146	-13.4838	first, | and
0.216216	-13.4838	it is | likely
0.234375	-13.4838	the | right,
0.0626398	-13.484	n-gram | the
0.195652	-13.484	the | concept.
0.4	-13.4843	could | made
0.0771812	-13.4843	to | candidate
0.545455	-13.4843	variables | those
0.292683	-13.4844	data | low
0.0732824	-13.4846	example | is
0.8	-13.4846	domain | estimated
0.8	-13.4846	people | every
0.292683	-13.4847	permits | of
0.444444	-13.4847	null | nonterminals
0.26	-13.4848	proportion | by
0.139665	-13.4848	related | or
0.333333	-13.4848	uses | text
0.4	-13.485	concordance | the
0.186275	-13.485	the | chapter
0.123348	-13.485	is | errors
0.391304	-13.4851	learner | it
0.444444	-13.4851	nw | nw
0.444444	-13.4851	100000 | gigaword
0.285714	-13.4852	so | should
0.204819	-13.4853	clause | in
0.202381	-13.4854	our | manually
0.180952	-13.4854	rest of | paper
0.114943	-13.4855	4.1 | the | in
0.314286	-13.4857	grammatical | which
0.0924574	-13.4857	then | on
0.142857	-13.4857	3 | 8
0.2	-13.4857	observed | by
0.5	-13.4857	performance | around
0.270833	-13.4858	however | are
0.182692	-13.4858	previously | for
0.391304	-13.4859	in | states.
0.344828	-13.4859	lf | uf
0.15493	-13.486	then, we | the
0.164062	-13.4861	of | s,
0.084507	-13.4862	a | argument
0.180952	-13.4862	this | decoding
0.0737327	-13.4864	could | in
0.545455	-13.4864	node | auxiliary
0.545455	-13.4864	translation | nonterminal
0.5	-13.4866	calculated | both
0.121339	-13.4868	subjects | the
0.204819	-13.4868	this | produced
0.8	-13.4868	detected | document
0.8	-13.4868	nature | difficult
0.4	-13.4869	in | triple
0.1	-13.487	three | with
0.0913462	-13.487	model | all
0.131068	-13.487	for | per
0.0612245	-13.4871	two | with
0.466667	-13.4871	by | strengths
0.26	-13.4872	one | annotated
0.421053	-13.4872	compression | an
0.204819	-13.4872	we | below
0.4	-13.4873	method | performs
0.041011	-13.4873	so | the
0.157143	-13.4875	that | included
0.227273	-13.4875	matrix | from
0.223881	-13.4875	this | access
0.25	-13.4876	using | value
0.421053	-13.4876	with | constraints,
0.180952	-13.4876	word-level | the
0.8	-13.4876	utterances | still
0.8	-13.4876	expressing | like
0.149351	-13.4878	sequences | a
0.241379	-13.4879	occur | than
0.166667	-13.488	partially | by
0.1875	-13.4881	experimental results | to
0.375	-13.4881	analyzing | dependency
0.285714	-13.4882	representation | used
0.272727	-13.4882	errors | more
0.189474	-13.4882	several | this
0.109589	-13.4882	approach, | of
0.305556	-13.4882	permutation | and
0.322581	-13.4883	this | constructing
0.8	-13.4883	projects | hr0011-08-c-0110.
0.625	-13.4884	both | compatible
0.285714	-13.4885	this | sometimes
0.0983607	-13.4886	and | case
0.0997151	-13.4886	must | in
0.285714	-13.4887	by | unknown
0.0663329	-13.4887	i | to
0.216216	-13.4887	(2005) | that
0.265306	-13.4888	we | solving
0.245614	-13.4888	evaluation | new
0.202381	-13.4888	is | particularly
0.213333	-13.4889	such | many
0.161538	-13.4889	a | review
0.314286	-13.489	in | restaurant
0.117188	-13.489	units | the
0.245614	-13.489	to | feedback.
0.0338321	-13.4892	• | the
0.322581	-13.4893	8 | two
0.15493	-13.4895	= | summationtext
0.545455	-13.4895	proved | useful
0.123894	-13.4896	that | current
0.305556	-13.4897	s | 〉
0.36	-13.4899	in | session
0.172414	-13.4901	by | frequency
0.333333	-13.4901	women | to
0.230769	-13.4901	queries | by
0.153846	-13.4902	method | training
0.625	-13.4902	corpus | format
0.466667	-13.4902	words | impression
0.270833	-13.4904	to | study.
0.0587659	-13.4904	tags | the
0.421053	-13.4904	with | κ
0.207317	-13.4904	a | japanese
0.305556	-13.4905	instance | target
0.545455	-13.4905	tuned | model
0.270833	-13.4906	then | .
0.0941476	-13.4906	set | used
0.161538	-13.4906	standard | model
0.204819	-13.4906	for | connectives
0.26	-13.4907	svm | from
0.169492	-13.4908	) | where
0.158273	-13.4908	an | problem
0.213333	-13.4909	scores | were
0.391304	-13.491	acl-hlt | 2011,
0.344828	-13.491	are | humans
0.545455	-13.491	density | its
0.175439	-13.4911	for | limited
0.15493	-13.4911	of | input.
0.305556	-13.4911	specific | its
0.143713	-13.4911	popular | and
0.204819	-13.4912	and | official
0.0967742	-13.4913	the | α
0.254902	-13.4913	parallel | corpus.
0.8	-13.4913	class | reversible
0.8	-13.4913	dictionaries | translated
0.147436	-13.4914	were | different
0.625	-13.4914	used | disambiguation.
0.0728376	-13.4914	the | achieved
0.0951157	-13.4915	as | any
0.213333	-13.4915	joint | using
0.545455	-13.4915	increase | rate
0.466667	-13.4915	in | careful
0.137363	-13.4916	the | mark
0.162791	-13.4918	using | approach
0.4	-13.4918	corpus | evaluation.
0.197674	-13.4918	the | item.
0.292683	-13.4919	models | unsupervised
0.129187	-13.4919	first, | for
0.230769	-13.4919	such | available
0.27907	-13.492	and | g.
0.0953608	-13.4921	using | set
0.0913462	-13.4922	not | they
0.0482804	-13.4922	an | on
0.0957447	-13.4923	our | is | from
0.189474	-13.4924	and | dialog
0.0421748	-13.4924	the | topic
0.051149	-13.4925	t | the
0.027089	-13.4925	model | and
0.165354	-13.4926	looked | the
0.220588	-13.4927	model | like
0.344828	-13.4927	experiments | original
0.146341	-13.4927	only | they
0.0836653	-13.4927	both | with
0.297297	-13.4929	which | fixed
0.8	-13.4929	metrics | muc
0.115385	-13.493	degree | of
0.375	-13.493	in | following sections.
0.078125	-13.493	we | event
0.272727	-13.4931	observation, | the
0.0258113	-13.4931	the | is | with
0.625	-13.4932	sources | twitter
0.625	-13.4932	link | concepts
0.4	-13.4933	over | range
0.161538	-13.4933	approaches | from
0.103774	-13.4933	above | to
0.322581	-13.4935	relative | different
0.8	-13.4936	when | maximizing
0.8	-13.4936	[ilp] | are
0.466667	-13.4937	input | linear
0.145455	-13.4938	it | even
0.166667	-13.4939	section | describe | the
0.36	-13.494	approaches | automatic
0.105096	-13.494	(2009) | the | of
0.2	-13.4942	dimension | and
0.36	-13.4943	these | sentences,
0.36	-13.4943	all | greater
0.219178	-13.4944	a | senses.
0.322581	-13.4944	done | or
0.270833	-13.4945	base | all
0.204819	-13.4945	and | program
0.219178	-13.4946	document | its
0.216216	-13.4946	the paper | follows.
0.216216	-13.4946	produced | with
0.0717423	-13.4947	a | x
0.191489	-13.4947	common | which
0.0906921	-13.4948	a | 2009).
0.625	-13.4949	in | 2(b),
0.305556	-13.4949	was | among
0.153846	-13.495	results | presented
0.344828	-13.495	phrases, | are
0.204819	-13.4951	was | did
0.210526	-13.4951	this | position
0.25	-13.4952	potential | are
0.173913	-13.4952	using | model, | the
0.13198	-13.4953	specifically, | to
0.25	-13.4954	where | query
0.25	-13.4954	each | combination
0.625	-13.4955	hedges | expression
0.130653	-13.4956	the | remains
0.27907	-13.4957	user | new
0.134021	-13.4958	(for | and
0.230769	-13.4958	in | queries.
0.237288	-13.4959	of | state,
0.4	-13.4959	develop | system
0.4	-13.4959	used | (i.e.
0.25	-13.4959	in | verbs.
0.193548	-13.4959	more | human
0.223881	-13.4959	et | 2011),
0.0957447	-13.496	to | every
0.241379	-13.4961	and | brown
0.5	-13.4961	as | available,
0.153846	-13.4962	that | seen
0.160305	-13.4963	personal | and
0.36	-13.4964	only | used.
0.466667	-13.4964	shallow | or
0.191489	-13.4965	encode | and
0.254902	-13.4966	data | knowledge
0.8	-13.4966	collapsing | side
0.164062	-13.4968	speakers | to
0.322581	-13.4969	combined | models
0.344828	-13.4969	information | extract
0.230769	-13.4969	in | interactive
0.0640569	-13.497	many | in
0.139665	-13.497	this | original
0.204819	-13.497	code | and
0.25	-13.4971	following | given
0.0917874	-13.4973	is | which | are
0.625	-13.4973	expressed | verbs
0.466667	-13.4973	accuracies | accuracy
0.0638051	-13.4974	we | following
0.137363	-13.4974	2) | of
0.197674	-13.4974	by | probabilities
0.121739	-13.4975	simple | as
0.17094	-13.4975	total | with
0.116279	-13.4976	observe | of
0.245614	-13.4976	the | lexicons.
0.421053	-13.4976	annotation | created
0.11583	-13.4976	scheme | and
0.545455	-13.4977	semantic | disambiguation.
0.265306	-13.4978	possible | any
0.25	-13.4978	action, | the
0.344828	-13.4979	the | characteristics.
0.5	-13.498	information | semantics
0.223881	-13.4981	a | voice
0.0528846	-13.4981	state | the
0.4	-13.4982	future, | more
0.117409	-13.4982	of | construction
0.0615551	-13.4982	not | not
0.466667	-13.4982	by | constituency
0.8	-13.4982	does | dominate
0.0904762	-13.4983	of | features,
0.0915663	-13.4983	our | models
0.36	-13.4984	0.81 | 0.80
0.5	-13.4984	describe | details
0.182692	-13.4985	task | information
0.173913	-13.4985	to | changes
0.180952	-13.4985	works | al.,
0.143713	-13.4985	take | for
0.0860215	-13.4986	these | an
0.0506236	-13.4986	semantic | in
0.126168	-13.4986	a | character
0.36	-13.4987	visually | to
0.545455	-13.4988	best | items
0.126761	-13.4989	in | utterances
0.227273	-13.4989	marked | as
0.107383	-13.499	parameters | to
0.314286	-13.4991	the | helper
0.625	-13.4991	class | modifier
0.0401477	-13.4992	we | our | and
0.421053	-13.4992	and | 2006a).
0.213333	-13.4992	and | reranking
0.10089	-13.4993	thank | and | for
0.0954907	-13.4993	we | 1.
0.15493	-13.4994	on | automatically
0.297297	-13.4994	worse | with
0.375	-13.4994	the | padt
0.466667	-13.4995	of | task).
0.169492	-13.4996	2.3 | to
0.625	-13.4997	edges, | edge
0.8	-13.4997	shorter | often
0.141176	-13.4997	considering | in
0.0825147	-13.4998	the | shown | figure
0.322581	-13.4999	relations | concepts
0.421053	-13.5	graph | matrix
0.142012	-13.5	primary | to
0.172414	-13.5002	translation | only
0.145455	-13.5004	that | classes
0.8	-13.5004	pp@2null1] | ]
0.172414	-13.5006	svm | for
0.166667	-13.5006	while | were
0.17094	-13.5006	corpus | only
0.152778	-13.5007	gradient | the
0.254902	-13.5007	the | reversible
0.0791367	-13.5008	these | or
0.5	-13.5008	army | contract/grant
0.625	-13.5009	fail | capture
0.625	-13.5009	dp | algorithm
0.625	-13.5009	tree | pst
0.139665	-13.5009	global | in
0.195402	-13.501	aligned | with
0.113636	-13.5011	those | not
0.344828	-13.5012	examples | labeled
0.8	-13.5012	effectively | labeling
0.8	-13.5012	detection | extraction.
0.135135	-13.5013	do not | are
0.545455	-13.5013	for | narration
0.179245	-13.5014	this | individual
0.421053	-13.5016	again, | be
0.4375	-13.5017	candidate | score.
0.466667	-13.5017	distributional | represent
0.138889	-13.5019	made | for
0.375	-13.5019	picked | in
0.8	-13.502	identified | parser.
0.625	-13.5021	boolean | mentioned?
0.625	-13.5021	selection | categories
0.625	-13.5021	global | entailment
0.285714	-13.5023	and | products
0.0811321	-13.5023	the | level.
0.129187	-13.5023	to | accurate
0.180952	-13.5023	the | levels.
0.375	-13.5023	similarity | author
0.0744337	-13.5023	find | and
0.4	-13.5024	relevant | task
0.421053	-13.5024	frequencies | all
0.0784314	-13.5024	to | (e.g.,
0.545455	-13.5024	number | post
0.545455	-13.5024	scfg | scfg
0.285714	-13.5025	the | branches
0.344828	-13.5025	getting | to
0.4375	-13.5026	results | runs
0.4375	-13.5026	sa | system:
0.625	-13.5026	ud | search
0.625	-13.5026	document | distributions.
0.314286	-13.5027	these | annotators
0.5	-13.5027	in | pass,
0.118852	-13.5028	use | translation
0.195402	-13.5028	in | phrasal
0.227273	-13.5028	trying | of
0.36	-13.5029	weights | model.
0.0690608	-13.5029	a | rule
0.0625	-13.503	alignment | and
0.625	-13.5032	languages | out
0.5	-13.5032	with | complements
0.5	-13.5032	head | non-head
0.5	-13.5032	applied | reduce
0.0631458	-13.5034	this is | in
0.175926	-13.5035	vp | to
0.189474	-13.5035	is | 5.
0.8	-13.5035	motivation | significantly
0.314286	-13.5036	2 | results.
0.25	-13.5037	as | filter
0.625	-13.5038	are | hardly
0.223881	-13.5038	translations | language
0.0359259	-13.5039	the | such as
0.189474	-13.5039	sentences | then
0.26	-13.504	shall | to
0.10596	-13.504	their | not
0.173913	-13.5041	model | without
0.27907	-13.5042	table | reports | of
0.25	-13.5042	used | version
0.4	-13.5043	small | very
0.265306	-13.5044	of | agreed
0.421053	-13.5044	lf | for
0.375	-13.5044	base | =
0.191489	-13.5046	other | any
0.0799257	-13.5048	for | pair | the
0.15493	-13.5048	algorithms | as
0.202381	-13.5048	is | boundary
0.322581	-13.5049	with | phrase.
0.241379	-13.505	the | 1993;
0.625	-13.505	words | glosses
0.545455	-13.505	output | aligned
0.8	-13.505	(chang | kernel
0.129353	-13.5051	common | as
0.272727	-13.5053	for | chen
0.0689655	-13.5054	et | 2008;
0.213333	-13.5055	again | a
0.111111	-13.5055	follows: | =
0.0184658	-13.5056	this | a
0.195402	-13.5056	that | rules.
0.625	-13.5056	indicates | slot
0.5	-13.5056	english | affect
0.14557	-13.5057	4. | is
0.0684932	-13.5058	the | of | classification
0.0942408	-13.5059	he | the
0.227273	-13.5059	this | weighted
0.227273	-13.5059	literature, | the
0.204819	-13.506	of, | of,
0.545455	-13.506	it | usable
0.272727	-13.5061	mt | language
0.105096	-13.5061	step, | the
0.117886	-13.5061	with | method
0.0499638	-13.5063	and | over
0.27907	-13.5063	abstract | propose
0.0934343	-13.5063	(2010) | to
0.207792	-13.5064	abstractness | the
0.0972222	-13.5065	which | only
0.104101	-13.5065	our | were
0.36	-13.5066	identify | relations
0.375	-13.5066	use | assess
0.545455	-13.5066	resulting | distribution
0.8	-13.5066	pr(t|w)
0.8	-13.5066	store | memory
0.0620155	-13.5067	as | have
0.4375	-13.5067	to | workers,
0.162791	-13.5068	the | ne
0.285714	-13.5069	sentence | generated
0.233333	-13.5069	statistical | used
0.114943	-13.5069	of | distributions
0.0874439	-13.5071	used | an
0.241379	-13.5071	models | large
0.205128	-13.5071	the | sec.
0.545455	-13.5071	mt | still
0.314286	-13.5072	understanding | dialogue
0.166667	-13.5073	a | if | its
0.161538	-13.5073	words | similar
0.297297	-13.5073	segments | by
0.120172	-13.5073	approach | it
0.8	-13.5073	mi | model.
0.102102	-13.5074	experiment, | the
0.272727	-13.5074	that | listed
0.625	-13.5074	approximation | only
0.625	-13.5074	contiguous | phrases
0.149351	-13.5075	case | not
0.466667	-13.5075	some | explicitly
0.4375	-13.5076	we | variable.
0.4375	-13.5076	 | merges
0.421053	-13.5076	these | follow
0.27907	-13.5077	computed | set
0.4	-13.5078	aligned | both
0.344828	-13.5078	into | before
0.189474	-13.5079	while | its
0.223881	-13.5079	no | these
0.4375	-13.508	u. | office
0.25	-13.508	nodes | other
0.120172	-13.508	the | predictive
0.344828	-13.5081	7.1 | and
0.1875	-13.5082	classifying | of
0.132653	-13.5082	features | both
0.297297	-13.5082	with | topics.
0.13587	-13.5083	abstract | as
0.241379	-13.5083	that | english.
0.223881	-13.5083	and | never
0.272727	-13.5084	for | personal
0.179245	-13.5084	root | is
0.375	-13.5084	syntactic | shallow
0.13369	-13.5084	factors | to
0.25	-13.5085	is | tokens.
0.5	-13.5085	which | recursively
0.625	-13.5086	time | minutes,
0.131313	-13.5087	semantics | in
0.174312	-13.5088	a | bayes
0.0849257	-13.5089	of | resources
0.285714	-13.5091	where | relative
0.0386285	-13.5091	the | evaluation
0.26	-13.5092	4.1 | using
0.185567	-13.5092	very | which
0.220588	-13.5093	tokenized | the
0.4	-13.5093	we | lexemes
0.344828	-13.5094	decoding | training
0.0122646	-13.5096	the | on
0.305556	-13.5097	explicit | between
0.117886	-13.5097	t, | the
0.230769	-13.5097	when | even
0.101493	-13.5098	the | syntax
0.289474	-13.5098	if | answer
0.14557	-13.5098	of | visual
0.0566388	-13.5099	parsing | and
0.241379	-13.5099	errors, | a
0.223881	-13.51	all | significantly
0.0733229	-13.5101	(1) | a
0.205128	-13.5102	all | top
0.168067	-13.5103	and | groups
0.195402	-13.5103	record | of
0.153846	-13.5103	involve | of
0.237288	-13.5104	of | thread
0.625	-13.5104	users | aac
0.0929648	-13.5104	the | strong
0.272727	-13.5105	fk | the
0.183673	-13.5107	best | at
0.333333	-13.5107	four | their
0.193182	-13.5107	an | automatically
0.0870536	-13.5107	problem | is
0.314286	-13.5108	an | research.
0.207792	-13.5108	we also | performance
0.289474	-13.5109	model | fixed
0.137363	-13.511	fraction | the
0.216216	-13.5111	evaluation, | in
0.0713224	-13.5111	who | the
0.265306	-13.5112	e | th
0.17094	-13.5112	difference | two
0.8	-13.5112	dataset | produce
0.314286	-13.5114	each | character
0.36	-13.5115	a | government
0.0609358	-13.5115	of | noun
0.421053	-13.5117	bing | and
0.104918	-13.5118	rules | from
0.4375	-13.5118	of | priori.
0.289474	-13.5118	given | containing
0.5	-13.5118	such | via
0.107383	-13.5119	a | position
0.115385	-13.5119	of | temporal
0.2	-13.5121	of | newspaper
0.0790441	-13.5121	to | system.
0.254902	-13.5122	mi | a
0.4375	-13.5122	is | biggest
0.0699001	-13.5122	is | new
0.625	-13.5122	or | digital
0.233333	-13.5123	are | part-of-speech
0.5	-13.5123	this | continues
0.297297	-13.5125	using | process
0.217391	-13.5126	and | utterances.
0.0954907	-13.5126	for | rule
0.333333	-13.5127	then, | using
0.333333	-13.5127	of | turns.
0.097493	-13.5128	of | structures
0.0716418	-13.5128	rule | and
0.13198	-13.5128	of | minimum
0.5	-13.5128	quite | large
0.297297	-13.513	them | features
0.0505952	-13.5131	the | sequence
0.121212	-13.5131	and | performed
0.123853	-13.5132	and | entailment
0.289474	-13.5132	while | small
0.0852878	-13.5132	every | a
0.120172	-13.5133	which | target
0.625	-13.5134	trainer | the
0.0949868	-13.5134	and | web
0.545455	-13.5134	existing | corpora
0.545455	-13.5134	annotation | relatively
0.285714	-13.5135	pt | of
0.8	-13.5135	expected | distributional
0.122807	-13.5137	results | task
0.36	-13.5139	match | gold
0.346154	-13.5139	discussion | two
0.625	-13.514	10% | 30%
0.0782918	-13.514	the | experiments.
0.266667	-13.5141	of | recognition.
0.0977654	-13.5141	best | with
0.27907	-13.5141	only | nodes
0.344828	-13.5141	we | in sec.
0.466667	-13.5142	free | not
0.136612	-13.5143	such | all
0.17757	-13.5144	achieve | on
0.217391	-13.5145	that | follow
0.266667	-13.5146	this | exactly
0.104101	-13.5146	question | a
0.625	-13.5146	distributional | baroni
0.27907	-13.5149	we | re
0.0894118	-13.5149	in | including
0.36	-13.515	their | thus
0.119149	-13.515	for | sentences.
0.136612	-13.515	by | relation
0.297297	-13.515	algorithm | clusters
0.8	-13.515	state: | slot3],
0.8	-13.515	computation | equation
0.0596234	-13.5152	the | of | algorithm
0.5	-13.5152	(du | al.,
0.344828	-13.5154	will | discussed
0.11194	-13.5155	alignments | and
0.175926	-13.5155	of | treebank.
0.157895	-13.5155	feature | its
0.266667	-13.5156	w | p
0.172414	-13.5156	more | performance
0.265306	-13.5156	probability | pair
0.195402	-13.5156	derivations | and
0.207792	-13.5156	that | n-grams
0.137363	-13.5156	that | full
0.466667	-13.5156	parameters | during
0.172727	-13.5156	here | are
0.227273	-13.5157	find | set
0.8	-13.5158	across | demonstrating
0.333333	-13.5159	university | italy
0.150685	-13.516	construct | in
0.122271	-13.516	ranking | in
0.272727	-13.5161	after | will
0.227273	-13.5161	that | adjective
0.193182	-13.5161	each | p
0.346154	-13.5163	identification | semantic
0.254902	-13.5164	the | oil
0.4375	-13.5164	learn | labeled
0.466667	-13.5165	robot | its
0.4	-13.5166	corpus | newswire
0.0745543	-13.5166	second | a
0.179245	-13.5167	outputs | a
0.0909091	-13.5168	in | additional
0.149351	-13.5169	a | start
0.17757	-13.5171	have | possible
0.625	-13.5171	model, | length.
0.161538	-13.5171	the | morphemes
0.131313	-13.5172	in | called
0.0949868	-13.5172	candidate | is
0.346154	-13.5173	translation | nist
0.241379	-13.5173	and | union
0.421053	-13.5173	has | growing
0.115079	-13.5173	an | algorithm
0.210526	-13.5173	find | most
0.36	-13.5174	identify | them
0.36	-13.5174	of | 4000
0.220588	-13.5174	multimodal | and
0.123288	-13.5174	polarity | and
0.056444	-13.5175	with | =
0.128079	-13.5176	select | for
0.625	-13.5177	candidate | abstractive
0.0563771	-13.5178	the | domain | the
0.220588	-13.5179	features | generated
0.165289	-13.5181	is | consistent
0.14966	-13.5181	data | labeled
0.13198	-13.5181	random | for
0.8	-13.5181	streaming | all
0.266667	-13.5182	combinations | are
0.150685	-13.5182	previous | or
0.272727	-13.5182	was partially | by
0.111524	-13.5182	and | language.
0.0688202	-13.5183	one | on
0.128079	-13.5183	training | data.
0.122271	-13.5183	empirical methods | language
0.0746753	-13.5184	for | three
0.116466	-13.5185	how | from
0.0797774	-13.5185	group | the
0.375	-13.5185	membership | a
0.0711111	-13.5186	is defined | the
0.5	-13.5186	in-domain | test
0.147436	-13.5187	it | shown
0.123288	-13.5187	architecture | the
0.0562731	-13.5187	the | step
0.189474	-13.5188	ie | and
0.466667	-13.5188	50 | pairs
0.466667	-13.5188	content-only | features
0.4	-13.5189	corpus | training.
0.0900474	-13.519	tables | the
0.1875	-13.5191	σ | and
0.139535	-13.5193	as: | and
0.181818	-13.5193	in | express
0.148387	-13.5194	now | for
0.169492	-13.5195	that | distributions
0.625	-13.5195	studies, | only
0.0318168	-13.5196	4 | the
0.346154	-13.5196	some | features.
0.213333	-13.5197	score | 0
0.545455	-13.5197	be | losing
0.8	-13.5197	signature | associated
0.5	-13.52	trained | learner
0.3125	-13.5201	different | annotated
0.117409	-13.5203	the | involving
0.14557	-13.5203	we | complete
0.305556	-13.5203	the | keywords.
0.148387	-13.5204	2. | from
0.8	-13.5204	variable | four
0.297297	-13.5205	information | better
0.5	-13.5205	word | counts.
0.466667	-13.5206	identify | user
0.165289	-13.5207	by | translations
0.625	-13.5207	journal | 1993).
0.375	-13.5207	using | tags.
0.189474	-13.5208	assessing | the
0.297297	-13.5208	but | low
0.159091	-13.5209	in | numbers
0.131313	-13.521	equation | to
0.322581	-13.521	continuous | as
0.25	-13.5211	discourse | relations
0.333333	-13.5211	are | arcs
0.162791	-13.5212	the | 20%
0.0818363	-13.5212	of | relative
0.233333	-13.5213	we | encode
0.150685	-13.5214	in | international
0.333333	-13.5214	all | classifiers
0.375	-13.5214	for | out-of-domain
0.0637019	-13.5214	target | in
0.466667	-13.5215	of | graphic’s
0.272727	-13.5216	50 | 50 | 70
0.147436	-13.5217	this | across
0.0988372	-13.5219	experiments, we | the
0.625	-13.5219	strategic | from
0.625	-13.5219	weights | costs
0.466667	-13.5219	analogous | that
0.150685	-13.522	4.2 | is
0.107639	-13.522	model | probability
0.174312	-13.5221	model | similarity
0.193182	-13.5221	recall | our
0.17094	-13.5221	in | hence
0.163934	-13.5223	reports | and
0.322581	-13.5223	not | annotation.
0.289474	-13.5224	findings | are
0.2	-13.5224	corpus | available
0.107266	-13.5224	the | analyses
0.625	-13.5225	contains | frame
0.227273	-13.5225	it | various
0.102804	-13.5226	five | of
0.016702	-13.5226	as | the | of
0.333333	-13.5227	over | relations
0.4375	-13.5228	weighting | ranking
0.8	-13.5228	null(null3|null1null2)
0.217391	-13.5229	if | token
0.545455	-13.5229	supervised | comparable
0.185567	-13.5229	contains | each
0.237288	-13.523	labels | our
0.237288	-13.523	purposes, | the
0.097493	-13.5232	is | labeled
0.105611	-13.5232	are | data.
0.06	-13.5232	such | as
0.36	-13.5233	problem | probability
0.0625	-13.5233	which | an
0.17757	-13.5233	the | thread
0.305556	-13.5233	when | present
0.466667	-13.5233	above | certain
0.297297	-13.5234	it | last
0.0411706	-13.5236	in | translation
0.128079	-13.5237	and | developed
0.3125	-13.5238	conclusion | described
0.2	-13.5238	information | document
0.466667	-13.5238	between | adjacent
0.36	-13.524	to | turn.
0.0701754	-13.5241	the | always
0.4375	-13.5241	parser | beam
0.380952	-13.5241	causing | to
0.289474	-13.5241	they | phrases
0.144654	-13.5242	for | reported
0.466667	-13.5242	system | slot
0.375	-13.5243	these | get
0.8	-13.5243	figure 3 shows | sentence
0.8	-13.5243	comment | post
0.163934	-13.5244	this | include
0.0715373	-13.5244	translation | with
0.245283	-13.5244	ontology | that
0.121739	-13.5247	choosing | the
0.237288	-13.5247	user | their
0.421053	-13.5247	evaluation | nist
0.380952	-13.5248	method | need
0.305556	-13.5248	methods | summarization
0.0780399	-13.5248	we | simple
0.4375	-13.5249	with | measure,
0.346154	-13.525	using | log
0.322581	-13.5251	due | errors
0.380952	-13.5252	improves | only
0.289474	-13.5252	discounting | and
0.25	-13.5252	process | two
0.0786106	-13.5252	other | with
0.126829	-13.5254	some | these
0.5	-13.5254	a | title.
0.5	-13.5254	are | inter
0.191011	-13.5255	also | no
0.545455	-13.5255	using | engineering
0.4	-13.5256	reveals | are
0.625	-13.5256	for | population.
0.245283	-13.5256	the | innovation
0.305556	-13.5257	as | occurrence
0.36	-13.5258	positives | false
0.0601504	-13.5258	proposed | of
0.5	-13.5259	model | items,
0.163934	-13.526	are | nouns
0.183673	-13.526	probability | its
0.237288	-13.5261	accuracy | other
0.117409	-13.5261	resource | of
0.322581	-13.5261	on | cultural
0.229508	-13.5261	diversity | to
0.127451	-13.5261	tuned | the
0.4375	-13.5262	quality | improved
0.12069	-13.5262	of | remaining
0.163934	-13.5263	architecture | of
0.0793951	-13.5263	the | coverage
0.25	-13.5264	we | 0.5
0.322581	-13.5264	forms | an
0.3125	-13.5265	moses | system
0.4375	-13.5266	translations | correct
0.4375	-13.5266	2.2 | algorithm
0.091358	-13.5267	clustering | and
0.4	-13.5267	corresponding | event
0.197674	-13.5267	keyword | and
0.8	-13.5267	concatenating | data
0.150685	-13.5268	the | examples,
0.297297	-13.5269	can | relations.
0.466667	-13.5269	syntactic | information.
0.625	-13.5274	sentence, | tree.
0.5	-13.5274	size | always
0.245283	-13.5275	request | and
0.8	-13.5275	algorithm | pagerank
0.346154	-13.5277	states | no
0.346154	-13.5277	testing | corpus
0.214286	-13.5277	find | word
0.380952	-13.5278	at | early
0.0788991	-13.5278	that | best
0.112782	-13.5278	the | parsing.
0.5	-13.5278	galley | 2006;
0.147436	-13.5279	here | with
0.4	-13.5279	count | its
0.625	-13.528	30% | 50%
0.128713	-13.5281	for | trees
0.0857143	-13.5281	many | for
0.132979	-13.5282	and | count
0.380952	-13.5282	represent | terms
0.0340136	-13.5282	the | while
0.179245	-13.5282	with | range
0.163934	-13.5283	sentiment | from
0.5	-13.5283	table 1: | section
0.305556	-13.5284	run | test
0.254902	-13.5285	we | restrict
0.161538	-13.5285	decoding, | the
0.122172	-13.5285	each | associated
0.322581	-13.5286	2 | tasks
0.0740132	-13.5286	often | to
0.0378151	-13.5286	one | in
0.346154	-13.5287	automatically | given
0.126214	-13.5287	opinions | the
0.8	-13.529	al., | carreras,
0.0237423	-13.5291	it | the | of
0.153846	-13.5291	operators | the
0.130653	-13.5291	it | many
0.4375	-13.5292	cluster | highest
0.466667	-13.5292	abstract | texts
0.272727	-13.5293	mechanisms | of
0.545455	-13.5293	alignment | pairs,
0.346154	-13.5294	specification | in
0.126214	-13.5294	social | in
0.0784672	-13.5295	thank | for
0.254902	-13.5295	with | real
0.4	-13.5295	set | 2008
0.333333	-13.5295	we | balance
0.3125	-13.5296	the | sessions.
0.107266	-13.5296	in | resources
0.175926	-13.5297	lemma | of
0.161538	-13.5297	strategies | in
0.466667	-13.5297	number | modifiers
0.11194	-13.5298	we | sentences.
0.625	-13.5298	in | (regardless
0.0953678	-13.5298	predicates | the
0.223881	-13.5298	n-gram | features
0.220588	-13.5299	texts | which
0.26	-13.5299	queries | or
0.210526	-13.53	simple | more
0.346154	-13.5301	specified | that
0.254902	-13.5303	which | aligned
0.4	-13.5303	uses | linguistic
0.08	-13.5303	the | highly
0.346154	-13.5304	lexicons | with
0.380952	-13.5304	and | author.
0.163934	-13.5305	to | computation
0.421053	-13.5305	elements | those
0.421053	-13.5305	this | result.
0.103226	-13.5305	using | words
0.0741351	-13.5306	of | state
0.36	-13.5307	annotation | 2
0.26	-13.5307	national | science
0.0349928	-13.5307	the | is | than
0.207792	-13.5308	two | alignment
0.305556	-13.5308	including | their
0.233333	-13.5309	these | values
0.17757	-13.5309	pairs | one
0.4	-13.531	3.2.2 | and
0.4	-13.531	way | model.
0.4	-13.531	k | 1.
0.322581	-13.5311	type | annotation
0.191011	-13.5311	from | unlabeled
0.305556	-13.5311	created | different
0.210526	-13.5313	is | 1).
0.144654	-13.5314	more | also
0.207792	-13.5314	on | pp.
0.545455	-13.5314	input | e.g.
0.8	-13.5314	bigrams | similar
0.8	-13.5314	french, | french
0.152778	-13.5316	we | split
0.380952	-13.5316	also | application
0.0966851	-13.5316	following | =
0.2	-13.5316	to | 4).
0.375	-13.5317	performance | improves
0.197674	-13.5318	of | intelligence
0.5	-13.5318	2 | 36
0.109155	-13.5321	λ | the
0.18	-13.5321	action | is
0.0473965	-13.5322	the | found | the
0.0801527	-13.5322	and | classification
0.0416452	-13.5324	first | to
0.297297	-13.5324	lms | on
0.421053	-13.5325	can | search.
0.0904645	-13.5326	system, | and
0.146497	-13.5327	defined | where
0.223881	-13.5327	for | d.
0.5	-13.5327	after | numbers
0.166667	-13.5328	bayes | and
0.375	-13.5328	roles | role
0.102894	-13.5329	7 | to
0.159091	-13.5329	scope | and
0.0675676	-13.5329	aligned | the
0.466667	-13.5329	documents, | sentences
0.210526	-13.5329	use | statistical
0.14557	-13.533	not | many
0.8	-13.533	any | dictionary,
0.380952	-13.5331	models | combine
0.272727	-13.5331	by | edit
0.375	-13.5331	the | reviewers.
0.0777577	-13.5332	distribution | in
0.156716	-13.5332	of | segments
0.5	-13.5332	hypothesis | decoding
0.183673	-13.5333	that | taken
0.380952	-13.5334	static | dynamic
0.346154	-13.5335	order | assess
0.4375	-13.5335	may | estimate
0.168067	-13.5335	our | out
0.104235	-13.5335	each | then
0.101852	-13.5335	use | words
0.272727	-13.5336	as | 1998).
0.0918367	-13.5337	on | english
0.5	-13.5337	shaded | are
0.333333	-13.5338	two | questions
0.333333	-13.5338	data | version
0.0446163	-13.5339	approach | a
0.297297	-13.5339	manually | were
0.168067	-13.534	for | translation,
0.166667	-13.5341	analyses | in
0.171171	-13.5341	of | numerical
0.625	-13.5341	reordering | long-range
0.380952	-13.5342	three | provide
0.223881	-13.5342	is | composition
0.0682451	-13.5342	could | a
0.36	-13.5343	row | shows
0.0773381	-13.5343	the | paper.
0.207792	-13.5343	translation | alignment
0.466667	-13.5343	a | not)
0.220588	-13.5344	for | design
0.132979	-13.5344	this | within
0.171171	-13.5344	following | two
0.183673	-13.5344	each | k
0.8	-13.5345	conclusions | sponsor.
0.421053	-13.5346	optimized | training
0.421053	-13.5346	because | frequent
0.088993	-13.5346	to | further
0.229508	-13.5346	in | groups.
0.178218	-13.5346	index | and
0.117647	-13.5347	and | corpora.
0.625	-13.5348	observe that | any
0.289474	-13.5349	mean | at
0.111524	-13.5349	of | complete
0.0573689	-13.535	we | new
0.142857	-13.535	mean | is
0.162602	-13.5352	specifically, | is
0.36	-13.5353	lexical | elements
0.114173	-13.5353	was | all
0.8	-13.5353	verbs, | tense
0.123288	-13.5354	usually | a
0.150685	-13.5355	approach | two
0.108392	-13.5355	having | of
0.322581	-13.5356	more | easier
0.0906863	-13.5356	is | search
0.545455	-13.5357	therefore, | necessary
0.545455	-13.5357	phrases | phrase-based
0.5	-13.5357	what | did
0.5	-13.5357	goal | action
0.4	-13.5358	which | (cf.
0.089372	-13.5358	all | this
0.14375	-13.5359	same | one
0.0677732	-13.5359	in | best
0.156716	-13.5359	a | 2007)
0.0646766	-13.536	4.1 | the
0.188889	-13.536	in | come
0.4375	-13.5361	as | query,
0.122172	-13.5361	the | adds
0.160305	-13.5362	hard | to
0.0534805	-13.5362	annotation | the | of
0.0979827	-13.5363	constituent | the
0.211268	-13.5365	used | available
0.154412	-13.5365	distinction | and
0.233333	-13.5366	then, | that
0.466667	-13.5366	16 | 21
0.3125	-13.5367	in particular, | our
0.5	-13.5367	onset | word
0.545455	-13.5368	both | agree
0.8	-13.5369	speaking | non-native
0.380952	-13.5372	they | larger
0.625	-13.5372	entailment | various
0.375	-13.5372	ip@null | vp@2null2
0.132979	-13.5373	in | reported
0.36	-13.5375	4 | now
0.466667	-13.5375	contains | empty
0.245283	-13.5376	because | because
0.195402	-13.5377	system | gold
0.333333	-13.5377	trees, | that
0.5	-13.5377	pair | f
0.0353818	-13.5378	at | and
0.4375	-13.5378	this | abstractive
0.102167	-13.5378	scores | on
0.102894	-13.5379	a | ranking
0.545455	-13.5379	share | meaning
0.0869565	-13.5379	of | aligned
0.233333	-13.538	a | images
0.205128	-13.538	is | composed
0.141104	-13.5382	also | where
0.266667	-13.5383	transitive | is
0.0311419	-13.5383	it | that
0.36	-13.5385	clustering | features
0.0372517	-13.5385	for | =
0.202532	-13.5385	this | strong
0.152778	-13.5385	1998; | and
0.625	-13.5385	for | constructive
0.191011	-13.5385	of | assessment
0.8	-13.5385	methods | acquire
0.109091	-13.5387	are | about
0.25	-13.5387	however, | similar
0.5	-13.5387	potential | user
0.0763889	-13.5388	let | be | the
0.115538	-13.5389	or | but
0.545455	-13.5389	method | representations
0.210526	-13.5389	proposed | used
0.101852	-13.539	as | while
0.3125	-13.5391	thus, | feature
0.4375	-13.5391	pairs | documents.
0.104575	-13.5392	best | by
0.11194	-13.5393	that | result
0.11194	-13.5393	1 | where
0.0801527	-13.5394	main | in
0.545455	-13.5395	colleagues | their
0.254902	-13.5396	3 | types
0.197674	-13.5398	denoted | and
0.135593	-13.5399	these | models
0.333333	-13.54	features | latent
0.205128	-13.54	total | words
0.191011	-13.54	required | for
0.545455	-13.54	and | 1979)
0.8	-13.5401	languages | family.
0.0906863	-13.5401	is | structure
0.346154	-13.5403	order | allow
0.254902	-13.5403	set | tags
0.197674	-13.5403	the | offset
0.148649	-13.5404	word | document
0.14375	-13.5404	the | management
0.205128	-13.5404	paragraphs | the
0.110701	-13.5404	with | so
0.0913706	-13.5404	latter | the
0.18	-13.5406	specifies | the
0.26087	-13.5406	evaluating | with
0.0685315	-13.5407	is | text
0.135593	-13.5407	as | linguistic
0.223881	-13.5408	vector | context
0.202532	-13.5409	algorithms | are
0.380952	-13.5409	of | confidence.
0.380952	-13.5409	recent | approaches
0.163934	-13.5409	random | from
0.4	-13.5409	features | entities.
0.146667	-13.5409	is | making
0.193182	-13.5409	pseudo | the
0.154412	-13.541	testing | in
0.104575	-13.541	or | other
0.254902	-13.5411	of | transformed
0.305556	-13.5413	task | very
0.193182	-13.5414	the | cut
0.116183	-13.5415	that | overall
0.8	-13.5416	annotators | cohen’s
0.217391	-13.5417	this | noisy
0.4375	-13.5417	describe, | the
0.27907	-13.5417	a | lm.
0.0522876	-13.5418	we | method
0.245283	-13.5418	this | proper
0.171171	-13.542	a | 2).
0.27907	-13.542	also | errors
0.346154	-13.5421	the | setup.
0.0860927	-13.5421	of | provided
0.225806	-13.5422	have | words.
0.545455	-13.5422	reading | were
0.211268	-13.5423	and | mi
0.8	-13.5424	metrics | synonym
0.175926	-13.5425	compares | and
0.4375	-13.5426	retrieve | or
0.106529	-13.5426	rank | of
0.220588	-13.5427	concatenated | the
0.0780399	-13.5428	can | it
0.0865604	-13.5429	for | we | an
0.0528967	-13.5429	for | system
0.195402	-13.543	europarl | and
0.229508	-13.543	we | sensitive
0.159091	-13.5431	information | text
0.237288	-13.5431	the | proceeds
0.3125	-13.5432	clustering | used
0.322581	-13.5432	motivate | of
0.144654	-13.5432	scoring | a
0.211268	-13.5432	in | down
0.545455	-13.5433	probability, | word
0.4	-13.5437	number | person
0.116183	-13.5437	modeling | a
0.25	-13.5437	translating | language
0.346154	-13.5438	good | ,
0.545455	-13.5438	recall | calculated
0.545455	-13.5438	words | twice
0.197531	-13.5439	closed | the
0.013756	-13.5439	a | is
0.217391	-13.5441	these | likely
0.272727	-13.5444	describes | system
0.136364	-13.5444	and | types.
0.2	-13.5444	university | abstract
0.4	-13.5445	arabic | data
0.145695	-13.5445	± | ±
0.065051	-13.5446	in | evaluation
0.0859729	-13.5446	that | less
0.5	-13.5446	about | categories
0.625	-13.5447	these, | were
0.0766488	-13.5447	in | words,
0.127451	-13.5447	illustrate | the
0.4375	-13.5448	specify | one
0.132979	-13.5449	for | highest
0.205128	-13.5449	examples | an
0.193182	-13.5449	are | identified
0.545455	-13.5449	increase | they
0.3125	-13.545	documents | terms
0.16	-13.5452	of | japan
0.197531	-13.5453	− | +
0.625	-13.5453	5a | is
0.625	-13.5453	incorporated | word
0.16129	-13.5453	works | for
0.171171	-13.5454	and | outperform
0.174312	-13.5454	use | search
0.185567	-13.5455	was | development
0.122172	-13.5456	is | edge
0.8	-13.5456	early | development
0.8	-13.5456	rule-based | knowledge
0.36	-13.5457	sentences | classified
0.0717629	-13.5457	equation | the
0.134078	-13.5457	of | community
0.202532	-13.5458	a | popular
0.137931	-13.5458	we | properties
0.266667	-13.546	using | improves
0.141104	-13.546	one | then
0.0271045	-13.546	the | the same
0.0658762	-13.546	the | b
0.4375	-13.5461	thus | automatic
0.0817996	-13.5461	into | with
0.346154	-13.5462	there | variation
0.178218	-13.5462	4.3 | for
0.297297	-13.5463	is | inference.
0.10061	-13.5464	we | bilingual
0.8	-13.5464	news | contentious
0.155556	-13.5465	to | n-best
0.0968661	-13.5465	in | see
0.545455	-13.5465	shot | in
0.297297	-13.5466	model | automatic
0.156716	-13.5467	exists | and
0.0424144	-13.5468	results | a
0.106529	-13.5468	the | listed
0.466667	-13.5468	from | definitions
0.4375	-13.5469	that | gained
0.137931	-13.5469	the | items.
0.5	-13.5471	at | allophonic
0.181818	-13.5471	to | end,
0.0935829	-13.5472	a | selected
0.115079	-13.5472	boundaries | the
0.8	-13.5472	are | norms,
0.333333	-13.5473	sa | [slot1,
0.36	-13.5475	been | years.
0.36	-13.5475	to | equivalent.
0.207792	-13.5475	in | mainly
0.207792	-13.5475	can | generate
0.12	-13.5476	as | sense
0.282051	-13.5476	we | sampler
0.346154	-13.5476	remaining | used
0.346154	-13.5476	quality | automatic
0.16	-13.5476	of | 2001;
0.119469	-13.5477	last | a
0.0955056	-13.5478	which | different
0.176471	-13.5478	below, | of
0.132979	-13.5481	examined | the
0.237288	-13.5481	relations, | as
0.4375	-13.5482	labeling | data.
0.4375	-13.5482	different | graphs
0.233333	-13.5482	used | like
0.086758	-13.5482	experiment | and
0.346154	-13.5483	np | advp
0.0615735	-13.5483	which is | of
0.124402	-13.5483	our | previous
0.146667	-13.5484	further | on
0.131579	-13.5484	here, | a
0.237288	-13.5486	achieved | which
0.333333	-13.5486	10: | =
0.229508	-13.5488	from | document.
0.8	-13.5488	corpus | 2008.
0.0911392	-13.5489	the | proper
0.289474	-13.549	2.1 | used
0.0909091	-13.549	makes | to
0.0977011	-13.5491	such | such
0.289474	-13.5492	of | judgments.
0.380952	-13.5493	active | instances
0.380952	-13.5493	criteria | which
0.132979	-13.5494	abstract | an
0.466667	-13.5495	composed | three
0.346154	-13.5496	is | positive.
0.174312	-13.5496	is | ranked
0.333333	-13.5496	speel | ‘play’
0.5	-13.5496	6.1 | corpus
0.156716	-13.5498	(and | to
0.114173	-13.5498	highest | and
0.222222	-13.5499	which | another
0.171171	-13.5499	human | an
0.0405539	-13.5499	when | to
0.266667	-13.55	we | genia
0.333333	-13.55	with | 10%
0.2	-13.55	document | information
0.466667	-13.55	preparation, | the
0.5	-13.5501	enough | data.
0.237288	-13.5502	more | what
0.555556	-13.5502	tweets | bullish
0.555556	-13.5502	connotative | positive
0.3125	-13.5503	by | automated
0.115702	-13.5503	reviews | the
0.0493736	-13.5503	we | such
0.089372	-13.5503	the overall | of
0.666667	-13.5504	1.2 | 1.3
0.666667	-13.5504	john | bill
0.666667	-13.5504	kappa | value
0.666667	-13.5504	object | landmark
0.666667	-13.5504	texts | following
0.666667	-13.5504	through | assignment
0.142857	-13.5504	the | keyword
0.8	-13.5504	formulate | generative
0.466667	-13.5505	dependency | capture
0.152174	-13.5506	2 | one
0.36	-13.5507	≈ | is
0.555556	-13.5508	weighting | importance
0.108696	-13.5508	clusters | and
0.207792	-13.5508	setting | =
0.10989	-13.5509	only | word
0.625	-13.5509	type | character
0.0963173	-13.5509	the | explicitly
0.123288	-13.5509	our | large
0.545455	-13.5509	at | leaf
0.545455	-13.5509	introducing | into
0.121622	-13.5509	it | about
0.666667	-13.551	pro-best | exm
0.666667	-13.551	supervised | maxent,
0.666667	-13.551	backtracking | deterministic
0.666667	-13.551	conll | closed
0.666667	-13.551	bahuta | banaahuaakottaige
0.16	-13.551	in | noisy
0.125	-13.551	was | two
0.162602	-13.5512	we | 8
0.8	-13.5512	issues | issue
0.289474	-13.5513	salient | is
0.3125	-13.5516	the | split.
0.333333	-13.5516	nodes | semantic
0.245283	-13.5516	human | technology
0.666667	-13.5517	liu | tree-to-string
0.30303	-13.5518	to | probabilities,
0.0634146	-13.5519	to | whether
0.545455	-13.552	similarly | other
0.115226	-13.5521	and | appear
0.0379913	-13.5521	that | more
0.150685	-13.5523	a | accurate
0.18	-13.5523	the | and | summary.
0.109489	-13.5523	2003) | the
0.245283	-13.5523	graphs | that
0.666667	-13.5524	unary | production
0.666667	-13.5524	items | student’s
0.666667	-13.5524	text, | corresponding
0.666667	-13.5524	scope | focus
0.555556	-13.5524	matrix | projection
0.229508	-13.5524	features | defined
0.083691	-13.5525	type | is
0.178218	-13.5525	a | spelling
0.4375	-13.5526	newstest2010 | ter
0.5	-13.5526	identify | opinions
0.118644	-13.5527	the | hierarchy
0.3125	-13.5528	4 | examples
0.245283	-13.5528	attachment | is
0.8	-13.5528	where | higher.
0.188889	-13.5529	corpora | or
0.322581	-13.5529	appearance | the
0.0641509	-13.5529	differences | the
0.0938338	-13.5529	adaptation | the
0.666667	-13.553	constraints, | coreference
0.666667	-13.553	scope | betwen
0.555556	-13.553	propbank | verb.
0.36	-13.5532	4.1.1 | and
0.229508	-13.5533	text | also
0.466667	-13.5533	a | word).
0.466667	-13.5533	from | eight
0.0797665	-13.5534	it | all
0.266667	-13.5534	3 | parsing
0.0547945	-13.5534	have | on
0.322581	-13.5535	extracted | web
0.166667	-13.5537	the | n-gram language
0.666667	-13.5537	accuracy | cross-validation.
0.666667	-13.5537	0.55 | selection
0.666667	-13.5537	node | incident
0.0844156	-13.5538	in | 2009).
0.466667	-13.5538	verb | pronoun
0.128866	-13.5539	semantic | words
0.555556	-13.5541	systems | address
0.555556	-13.5541	half | them
0.555556	-13.5541	e1 | relation
0.211268	-13.5541	use | speech
0.2	-13.5542	nonterminal | is
0.162602	-13.5543	0 0 | .
0.297297	-13.5543	obviously | of
0.297297	-13.5543	presenting | in
0.666667	-13.5544	v|ja01
0.113725	-13.5544	temporal | and
0.157895	-13.5544	final | on
0.8	-13.5544	subjects | what
0.168142	-13.5544	each | selected
0.153285	-13.5546	of | depth
0.14966	-13.5547	but | were
0.333333	-13.5547	which | predictions
0.289474	-13.5548	given | v
0.322581	-13.5548	shown | high
0.108696	-13.5548	sequence | is
0.107527	-13.5548	pairs | with
0.169643	-13.5549	can | knowledge
0.4	-13.5549	appropriate | these
0.25	-13.5549	language | 2002).
0.266667	-13.555	given | 3.
0.666667	-13.555	proposed | widely
0.666667	-13.555	on | (wmt)
0.666667	-13.555	mt | problem.
0.666667	-13.555	yago, | wikipedia
0.666667	-13.555	different | factoids.
0.333333	-13.555	information | e.g.
0.555556	-13.5552	k | kth
0.555556	-13.5552	bleu | hypothesis
0.545455	-13.5553	or | candidates.
0.545455	-13.5553	deterministic | dependency
0.380952	-13.5554	were | tested
0.346154	-13.5555	1989; | and
0.087886	-13.5555	for | standard
0.297297	-13.5555	data | training.
0.297297	-13.5555	array | in
0.0668486	-13.5555	key | the
0.145695	-13.5556	a | frame
0.0549954	-13.5557	each | from
0.333333	-13.5557	trigger | its
0.555556	-13.5558	tokens | aligned
0.214286	-13.556	class | word
0.15873	-13.556	to | classes.
0.145695	-13.5561	the | identity
0.5	-13.5561	useful | statistical
0.289474	-13.5562	the | week.
0.15873	-13.5563	problem | not
0.666667	-13.5564	differences | perception
0.119469	-13.5564	regarding | the
0.217391	-13.5565	although | features
0.0735786	-13.5565	and | time
0.110294	-13.5565	the | clause
0.211268	-13.5565	test | three
0.346154	-13.5566	tagger | trained
0.233333	-13.5566	the | hebrew
0.202532	-13.5567	approximately | in
0.322581	-13.5568	consensus | a
0.040724	-13.5568	shows | of
0.8	-13.5568	t | !
0.8	-13.5568	entity | expansion.
0.4	-13.5569	words, | similar
0.4	-13.5569	rule | original
0.666667	-13.557	resolve | anaphoric
0.100946	-13.557	the | half | the
0.4375	-13.557	et | 2008a;
0.0918367	-13.557	is | available
0.266667	-13.5572	to | articles,
0.625	-13.5572	4 | lexicalized
0.625	-13.5572	can | (at
0.183673	-13.5575	predict | for
0.229508	-13.5575	corpora | different
0.8	-13.5576	scored(r|s)
0.8	-13.5576	taking | containing
0.165289	-13.5577	in | 2010
0.166667	-13.5577	for | property
0.666667	-13.5577	weights | matrices
0.172727	-13.5578	of | strongly
0.4375	-13.5579	translations | them
0.154412	-13.5579	patterns | be
0.555556	-13.558	we first | review
0.120536	-13.558	of | reading
0.0938338	-13.558	two | an
0.154412	-13.558	to | link
0.152174	-13.5581	as | simple
0.174757	-13.5582	rather than | for
0.289474	-13.5583	files | for
0.240741	-13.5583	equal to | in
0.666667	-13.5584	2009) | (huang
0.666667	-13.5584	project | collaboration
0.666667	-13.5584	tokens; | is
0.666667	-13.5584	algorithm | steps.
0.666667	-13.5584	aligned | interleaving
0.0863636	-13.5584	for | research
0.134078	-13.5584	of | objects
0.8	-13.5584	analyzing | about
0.8	-13.5584	motivation | measure
0.104027	-13.5585	his | the
0.282051	-13.5586	listed | for
0.36	-13.5587	model, | additional
0.3125	-13.5588	lastly, | the
0.0519373	-13.5588	them | the
0.0739496	-13.5589	with | different
0.135593	-13.5589	are | generated
0.466667	-13.5589	0.9 | 100
0.666667	-13.5591	perform | searches
0.666667	-13.5591	problem | segmentation,
0.666667	-13.5591	25 | their
0.666667	-13.5591	sharm | sheikh
0.666667	-13.5591	æ | ø
0.666667	-13.5591	n-gram | efficiently
0.0491924	-13.5591	and | its
0.155556	-13.5592	suggested | of
0.217391	-13.5593	bigram | as
0.289474	-13.5594	the | sounds
0.193182	-13.5594	for | statistics
0.145695	-13.5595	two | (1)
0.380952	-13.5596	what | dialogue
0.380952	-13.5596	systemt | a
0.0844156	-13.5596	far | the
0.162602	-13.5597	scope | a
0.0610169	-13.5597	of | found
0.5	-13.5597	at | 5,
0.129534	-13.5598	counts | a
0.141975	-13.5599	for | i.e.
0.26087	-13.5599	and | incorporated
0.2	-13.5601	semantics | that
0.8	-13.5601	linear | svms
0.5	-13.5602	opinionated | opinion
0.666667	-13.5604	analysis | fixation
0.666667	-13.5604	outperforms | tfidf
0.666667	-13.5604	accuracy | efficiency.
0.333333	-13.5604	classifier | svm
0.333333	-13.5604	able | them
0.25	-13.5605	a | right-hand
0.214286	-13.5605	also | relations
0.0753065	-13.5606	not | our
0.555556	-13.5608	collected | utterances
0.30303	-13.5609	with | control
0.141104	-13.561	the | users.
0.4375	-13.561	sie:you | es:it
0.117904	-13.561	mentions | a
0.625	-13.561	1 | his
0.0475862	-13.561	the | associated
0.16129	-13.561	new | using
0.666667	-13.5611	robust | entailment
0.13089	-13.5611	implementation | a
0.266667	-13.5612	phase | for
0.322581	-13.5613	important | how
0.545455	-13.5613	apply | framework
0.168142	-13.5613	density | the
0.282051	-13.5614	predictive | for
0.555556	-13.5614	results | 15%
0.0558184	-13.5615	2 | for
0.202532	-13.5616	to | emotional
0.380952	-13.5616	by | relation,
0.625	-13.5616	path | decoding
0.211268	-13.5616	used | including
0.121076	-13.5616	the | wmt
0.0551918	-13.5616	baseline | and
0.10473	-13.5617	candidate | for
0.0923483	-13.5617	of | 2006).
0.18	-13.5617	as | set,
0.0906801	-13.5617	how | can
0.172727	-13.5617	data | experiments
0.666667	-13.5618	newswire | arabic
0.13089	-13.5618	for | lexicon
0.191011	-13.5618	the | understood
0.555556	-13.562	such | expressions,
0.0973451	-13.562	being | a
0.25	-13.562	b | y
0.0651042	-13.5621	by | it
0.154412	-13.5621	model | document
0.222222	-13.5622	is | topics.
0.625	-13.5622	some | formal
0.0687135	-13.5622	source | is
0.5	-13.5622	using | computation
0.36	-13.5623	* | <
0.289474	-13.5623	use | metrics
0.666667	-13.5624	efficiently | tree
0.057	-13.5624	the | highest
0.545455	-13.5624	scores | 261
0.8	-13.5625	pair | w.
0.8	-13.5625	ivan | and
0.380952	-13.5627	summary | given
0.15	-13.5627	performance | test
0.0824524	-13.5627	a | w
0.5	-13.5627	all | annotator
0.4375	-13.5628	all | involve
0.666667	-13.5631	per | pair.
0.666667	-13.5631	word–emotion | by
0.555556	-13.5631	word | giza++.
0.8	-13.5633	p. | k.
0.117904	-13.5634	the | parsed
0.166667	-13.5635	in | unigram
0.380952	-13.5635	however, | questions
0.176471	-13.5635	sentences, | with
0.0386715	-13.5636	baseline | the
0.272727	-13.5636	be | independently
0.466667	-13.5636	not | all,
0.202532	-13.5637	often | such
0.113821	-13.5637	examples | as
0.666667	-13.5638	exhibit | variation
0.666667	-13.5638	ravichandran | relation
0.4	-13.5638	tests | our
0.0466045	-13.564	their | in
0.222222	-13.564	simple | our
0.4375	-13.5641	only | minimal
0.625	-13.5641	orientations | or
0.8	-13.5641	stronger | human
0.8	-13.5641	cape | sao
0.4	-13.5642	only | rank
0.5	-13.5642	significantly | results.
0.5	-13.5642	tweet | objective
0.188889	-13.5643	however, | any
0.30303	-13.5643	separated | a
0.0744501	-13.5644	is | dependency
0.26087	-13.5644	but | out
0.0960452	-13.5644	we | words.
0.36	-13.5645	search | hypotheses
0.666667	-13.5645	parasense | translation
0.666667	-13.5645	graphs | vertex
0.666667	-13.5645	unlike | models,
0.282051	-13.5646	hypothesis | word
0.121622	-13.5646	in | pattern
0.266667	-13.5647	system, | it
0.555556	-13.5648	cs | sr
0.411765	-13.5648	this | refined
0.625	-13.5648	performed | classification.
0.0593952	-13.5648	the | various
0.0957746	-13.5649	rate | of
0.134831	-13.5649	selecting | a
0.666667	-13.5651	prove | useful
0.666667	-13.5651	fl | transliteration:
0.225806	-13.5651	for | papers
0.411765	-13.5652	that | both.
0.346154	-13.5653	particular | (e.g.,
0.126214	-13.5653	in | learned
0.0885167	-13.5654	slightly | the
0.380952	-13.5654	type | similarity
0.625	-13.5654	design, | with
0.3125	-13.5655	search | its
0.11336	-13.5655	up | for
0.322581	-13.5656	2 | basic
0.322581	-13.5656	dependencies | words
0.0190997	-13.5656	the | are | and
0.229508	-13.5656	first | list
0.411765	-13.5657	developed | parallel
0.205479	-13.5657	all | considered
0.545455	-13.5657	: | →r
0.666667	-13.5658	provides | api
0.666667	-13.5658	button | task
0.666667	-13.5658	implicit | explicitly
0.4	-13.5658	translating | which
0.197531	-13.5661	tree | node
0.411765	-13.5661	of | penalized
0.137725	-13.5662	the | established
0.2	-13.5662	this | equivalent
0.297297	-13.5663	english | available
0.211268	-13.5663	where | node
0.255319	-13.5663	for | reverse
0.346154	-13.5664	their | across
0.130208	-13.5664	for | summarization
0.666667	-13.5665	view, | usa
0.141975	-13.5666	for | distance
0.625	-13.5667	0.56 | 0.86
0.083691	-13.5667	the | problems
0.211268	-13.5667	for example, | first
0.346154	-13.5668	for | phenomena.
0.25	-13.5668	when | domain
0.5	-13.5668	min | .
0.5	-13.5668	therefore, | small
0.555556	-13.567	form | allowed
0.0923483	-13.567	baseline | on
0.255319	-13.567	present | task
0.322581	-13.5672	with | alignments.
0.297297	-13.5672	of | studies.
0.0864486	-13.5673	is | whether
0.129534	-13.5673	and | therefore
0.5	-13.5673	we | proposition.
0.4	-13.5674	appear | document
0.545455	-13.5674	clearly | other
0.466667	-13.5674	given | respective
0.8	-13.5674	applications | tree
0.166667	-13.5675	with | systems.
0.322581	-13.5676	mapping | words
0.11284	-13.5677	can | given
0.186813	-13.5677	score | sentence
0.0474227	-13.5678	important | the
0.202532	-13.5678	grammar | language
0.222222	-13.5678	randomly | sentences
0.4	-13.5678	in | produced.
0.087886	-13.5679	sentences | by
0.208333	-13.5679	combination | which
0.282051	-13.568	study | study
0.545455	-13.568	j | pi,j)
0.4375	-13.5681	passive | active
0.36	-13.5682	4 | several
0.122066	-13.5682	typically | to
0.211268	-13.5682	machine | techniques
0.8	-13.5682	systems | 2007)).
0.8	-13.5682	language | team,
0.184783	-13.5683	and | for | models.
0.062201	-13.5684	entities | the
0.363636	-13.5684	wilson | al.
0.363636	-13.5684	characters | ¡
0.666667	-13.5685	hundreds | millions
0.666667	-13.5685	personality | native
0.411765	-13.5686	phrases | lexical
0.555556	-13.5687	subj | comps
0.36	-13.5689	taken | this
0.275	-13.5689	separately | for
0.275	-13.5689	rightmost | of
0.128205	-13.569	that | take
0.8	-13.569	look | evidence
0.188889	-13.5691	as | learn
0.137725	-13.5691	all | models
0.211268	-13.5691	such | approach
0.666667	-13.5692	ellen | computing
0.666667	-13.5692	collins | koo
0.625	-13.5692	labels | features.
0.625	-13.5692	approach | discovering
0.21875	-13.5692	examples | such
0.168142	-13.5692	is | previously
0.151079	-13.5694	support | for
0.0990712	-13.5695	through | to
0.148936	-13.5695	query | with
0.275	-13.5698	we | images
0.466667	-13.5698	’7 | ’7
0.5	-13.5698	because | share
0.3125	-13.5699	but | necessarily
0.166667	-13.5699	• | using
0.4375	-13.5699	5% | binary
0.4375	-13.5699	contrast, | over
0.168142	-13.5699	internal | and
0.0387812	-13.57	to | one
0.0970588	-13.57	reduce | of
0.163793	-13.5701	exponential | the
0.162602	-13.5701	using | text
0.363636	-13.5702	breakdown | in
0.4	-13.5703	using | measure.
0.625	-13.5705	within | already
0.333333	-13.5705	of | 7%
0.2	-13.5706	4 | approach
0.380952	-13.5708	and | iii)
0.211268	-13.5708	convergence | and
0.333333	-13.5709	as | zhang
0.555556	-13.571	path | tree.
0.625	-13.5711	entry | name
0.3125	-13.5712	as | similarly
0.4375	-13.5712	alternative | would
0.0828025	-13.5712	the | outperforms
0.666667	-13.5713	defined | previously
0.289474	-13.5715	corpus | aligned
0.176471	-13.5715	200 | and
0.411765	-13.5715	errors | could
0.0687135	-13.5715	to | annotation
0.214286	-13.5716	implemented | as
0.16129	-13.5716	for | factor
0.0709779	-13.5717	and | methods
0.36	-13.5718	signature module | a
0.346154	-13.572	if | correctly
0.169643	-13.5721	setup | in
0.36	-13.5722	allows | any
0.144737	-13.5722	that | generate
0.8	-13.5723	department | communication
0.25	-13.5724	they | pairs
0.134831	-13.5724	set | target
0.131148	-13.5724	build | for
0.240741	-13.5725	not | seen
0.4	-13.5727	results | wsj
0.146667	-13.5727	functional | the
0.380952	-13.5728	syntactic | 2007).
0.153285	-13.5731	update | to
0.4	-13.5731	of | 0)
0.8	-13.5731	relevant | subsequent
0.411765	-13.5732	sense | polysemous
0.118943	-13.5732	good | is
0.118943	-13.5732	have | number
0.333333	-13.5733	ordering | the | ordering
0.100629	-13.5733	the | t.
0.0698027	-13.5734	user | to
0.118943	-13.5734	detailed | and
0.128866	-13.5735	use | models
0.06375	-13.5735	in | present
0.380952	-13.5736	german | number
0.188889	-13.5736	5 | test
0.363636	-13.5736	of | hearing
0.411765	-13.5736	and | versa,
0.142857	-13.5736	within | are
0.555556	-13.5738	subtle | between
0.168142	-13.5738	in | regression
0.8	-13.5739	pr(d|ep,f)
0.266667	-13.574	incorrectly | to
0.666667	-13.574	correct | decisions
0.4	-13.574	existing | learning
0.255319	-13.5743	original | sentence
0.178218	-13.5743	with | judgments
0.16129	-13.5743	that | equal
0.4375	-13.5744	both | (in
0.174757	-13.5745	the | iteratively
0.333333	-13.5746	without | same
0.129534	-13.5746	generated | for
0.666667	-13.5747	with | deaf
0.666667	-13.5747	tooth | xylitol
0.282051	-13.5747	over | top
0.8	-13.5747	topics | interpretable
0.36	-13.5748	for | forms.
0.26087	-13.5748	rather, | to
0.101639	-13.5749	chose | the
0.146667	-13.5751	cases | are
0.26087	-13.5751	show that | well
0.4375	-13.5752	so | labels
0.118421	-13.5752	under | is
0.126904	-13.5752	that | automatically
0.666667	-13.5753	robust | unseen
0.21875	-13.5754	frequency | corpus
0.36	-13.5755	so | cannot
0.0574949	-13.5755	this | used
0.555556	-13.5756	classifier | classified
0.625	-13.5756	additional | improves
0.625	-13.5756	line | while
0.3125	-13.5757	finally, | were
0.195122	-13.5757	has | accuracy
0.208333	-13.5758	model | generate
0.346154	-13.5759	after | before
0.0717703	-13.5759	distribution | and
0.666667	-13.576	local | coherence
0.666667	-13.576	chunking | other
0.666667	-13.576	sentiment | (sa)
0.208333	-13.576	paraphrase | which
0.555556	-13.5761	mert | experiment
0.555556	-13.5761	large | spaces
0.555556	-13.5761	modeling | linguistic
0.363636	-13.5762	parser | 2007).
0.176471	-13.5762	2007), | to
0.625	-13.5763	report | f-measure
0.0934066	-13.5765	that | standard
0.12093	-13.5765	fragment | the
0.116379	-13.5766	these | also
0.024689	-13.5766	the | into
0.666667	-13.5767	inside | phrase.
0.666667	-13.5767	specifically, | hypothesis
0.346154	-13.5767	trees | grammar
0.555556	-13.5767	interand | agreement
0.16	-13.5768	(2004) | of
0.289474	-13.5768	model | conditioned
0.625	-13.5769	verb | medieval
0.240741	-13.5769	better | compared
0.282051	-13.577	applied | parsing
0.346154	-13.577	range | 1
0.4375	-13.577	vectors | space.
0.30303	-13.577	with | phrases.
0.0675105	-13.5771	of | values
0.180851	-13.5771	clusters | with
0.074266	-13.5771	the | questions
0.266667	-13.5773	(akkaya | al.,
0.0795229	-13.5773	sentences | as
0.0692771	-13.5773	is | possible
0.333333	-13.5774	learned | patterns
0.141026	-13.5774	conference | computational
0.202532	-13.5775	this | real
0.380952	-13.5775	had | impact
0.208333	-13.5775	given | would
0.0895522	-13.5775	the | svm
0.188889	-13.5776	with | binary
0.134831	-13.5776	of | mwe
0.0503145	-13.5777	the | additional
0.346154	-13.5777	get | but
0.266667	-13.5778	possible | when
0.140244	-13.5778	of | optimization
0.178218	-13.5778	an | label
0.11588	-13.5779	taking | and
0.0693816	-13.578	of | examples
0.8	-13.578	listed | tables
0.8	-13.578	about | query.
0.666667	-13.5781	studying | one
0.30303	-13.5782	combination | information
0.346154	-13.5784	towards | more
0.555556	-13.5784	acl | papers
0.322581	-13.5784	considered | their
0.4	-13.5785	vbd | vbn
0.30303	-13.5785	while | across
0.25	-13.5786	or | additional
0.130435	-13.5787	data | performance
0.202703	-13.5787	automatic | semantic
0.666667	-13.5788	veselin | and
0.666667	-13.5788	kappa | (κ)
0.666667	-13.5788	objective | summary
0.666667	-13.5788	condition, | has
0.225806	-13.5789	good | which
0.116883	-13.5789	understand | the
0.4	-13.5789	do | share
0.289474	-13.5789	used, | are
0.147887	-13.579	in | plan
0.107527	-13.579	represents | a
0.5	-13.579	4.2 | framework
0.275	-13.5791	with | institute
0.141026	-13.5791	applying | in
0.0421988	-13.5793	out | the
0.4	-13.5793	impact | have
0.4	-13.5793	without | method
0.188889	-13.5794	phrasal | and
0.625	-13.5795	multiple | translations.
0.245283	-13.5795	13 | for
0.555556	-13.5796	plan | integrate
0.555556	-13.5796	d. | y.
0.0909091	-13.5796	had | and
0.188889	-13.5796	of | paraphrasing
0.363636	-13.5796	and | sparsity.
0.5	-13.5796	initial | improvement
0.411765	-13.58	model; | is
0.666667	-13.5801	algorithm | crammer
0.555556	-13.5801	drop | quality
0.146667	-13.5801	for | linear
0.625	-13.5801	limitations | may
0.625	-13.5801	like | above
0.5	-13.5801	other | not.
0.380952	-13.5802	the | speaking.
0.232143	-13.5802	by | methods.
0.0895522	-13.5802	we | make
0.225806	-13.5803	these | user
0.363636	-13.5803	correlated | human
0.0734558	-13.5804	the | objective
0.0921053	-13.5804	is | words.
0.4	-13.5805	with | ith
0.232143	-13.5805	novel | based
0.0520571	-13.5805	since | in
0.266667	-13.5806	the | prefers
0.153285	-13.5806	5.2 | a
0.5	-13.5806	grouped | their
0.16	-13.5807	a | english.
0.555556	-13.5807	capture | words,
0.289474	-13.5807	qualitative | that
0.138554	-13.5807	is performed | the
0.186813	-13.5807	d | m
0.666667	-13.5808	names | english.
0.666667	-13.5808	new | segmentation.
0.666667	-13.5808	database | facts
0.666667	-13.5808	2006 | text
0.21875	-13.5808	slot | in
0.21875	-13.5808	their | them
0.255319	-13.5809	here | use
0.30303	-13.581	the | statement.
0.106383	-13.5811	this | annotation
0.0482014	-13.5812	the | (see
0.555556	-13.5813	ciphers | cipher
0.461538	-13.5814	frame | mt/ref
0.4	-13.5814	300 | from
0.202703	-13.5816	magnitude | the
0.232143	-13.5817	basic | from
0.363636	-13.5818	extracted | those
0.214286	-13.5818	beam | for
0.0238095	-13.5819	the | sentence | the
0.461538	-13.5819	dm | dm
0.461538	-13.5819	specific | further
0.625	-13.582	such | groups
0.411765	-13.5821	order | maximize
0.266667	-13.5822	to | reordering.
0.666667	-13.5822	defined | generates
0.380952	-13.5822	consists | given
0.380952	-13.5822	gender | which
0.0873786	-13.5822	is | form
0.0783939	-13.5822	based on | is
0.031269	-13.5823	first | the | of
0.555556	-13.5824	term | within
0.0787402	-13.5824	corresponding | and
0.15873	-13.5825	500 | of
0.225806	-13.5826	the | multiplication
0.240741	-13.5826	our | suggest
0.5	-13.5826	using | 1999).
0.666667	-13.5829	featured | task,
0.666667	-13.5829	(barzilay | 2003).
0.363636	-13.5829	each | id
0.333333	-13.5829	raw | are
0.30303	-13.5829	successfully | by
0.0780952	-13.5829	the | lm
0.555556	-13.583	rewriting | graph
0.0732113	-13.583	only | on
0.152174	-13.583	evaluation | results
0.168142	-13.583	all | corpus.
0.3125	-13.5831	keyphrases | and
0.346154	-13.5831	can | strings
0.153285	-13.5832	in | loss
0.211268	-13.5832	component | from
0.111554	-13.5832	3 | 6
0.169811	-13.5833	demonstrated | a
0.176471	-13.5833	core | is
0.118421	-13.5833	induction | the
0.4	-13.5834	would | produce
0.4	-13.5834	employing | as
0.0533333	-13.5835	first, | the
0.190476	-13.5835	from | generation
0.666667	-13.5836	was | 2d
0.26087	-13.5837	coreference | mentions
0.8	-13.5838	issue, | first
0.101307	-13.5839	follows | the
0.190476	-13.5839	the | mention.
0.0276586	-13.5839	which | in
0.151079	-13.584	that | (i.e.,
0.25	-13.584	improve | parsing
0.625	-13.584	promising | models
0.625	-13.584	ud | better
0.08933	-13.584	in | ,
0.245283	-13.584	heuristics | in
0.380952	-13.5842	heavily | with
0.380952	-13.5842	they | obtain
0.411765	-13.5842	words: | a
0.333333	-13.5842	we | forced
0.666667	-13.5843	which | co-occur.
0.115385	-13.5843	a | entire
0.15	-13.5844	here | on
0.186813	-13.5844	cannot | be
0.346154	-13.5845	domains | which
0.625	-13.5846	reduces | problem
0.8	-13.5846	decoder | chiang,
0.555556	-13.5847	models | abstract
0.461538	-13.5847	research | identifying
0.461538	-13.5847	out-of-domain | data
0.461538	-13.5847	derivative | is
0.4	-13.5847	one | none
0.043923	-13.5847	the | methods
0.202703	-13.5848	the | mentions,
0.666667	-13.585	wα | wβ
0.0860465	-13.585	other | or
0.294118	-13.585	translation | sms
0.294118	-13.585	of | compounds.
0.294118	-13.585	2.1 | our
0.178218	-13.585	of | vertices
0.236364	-13.5852	these | errors
0.363636	-13.5852	goal | provide
0.211268	-13.5852	out | at
0.100313	-13.5853	support | of
0.555556	-13.5853	smoothing | 2002).
0.333333	-13.5853	chiang | al.
0.8	-13.5854	q: | ∧
0.8	-13.5854	pronunciation | 2009).
0.289474	-13.5855	on | stack
0.100649	-13.5856	group | and
0.146853	-13.5856	other | they
0.666667	-13.5857	set | adjectival
0.666667	-13.5857	note | tuning
0.666667	-13.5857	als | euro
0.0911458	-13.5857	that | be used
0.112903	-13.5857	for | .
0.107407	-13.5858	key | to
0.26087	-13.5858	thus, | all
0.555556	-13.5859	errors | correctly
0.4	-13.5859	generate | topic
0.119816	-13.5859	shown | can
0.625	-13.5859	above | final
0.625	-13.5859	contributions | (3)
0.625	-13.5859	bayesian | dialect
0.255319	-13.5859	analysis | shows
0.186813	-13.5859	we | translate
0.0457816	-13.586	method | the | of
0.184783	-13.5863	approach | one
0.8	-13.5863	psychology | computational
0.666667	-13.5864	mail, | suicide
0.0707547	-13.5864	pair | a
0.188889	-13.5866	that | rich
0.625	-13.5866	algorithm, | first
0.625	-13.5866	word | bigrams,
0.115385	-13.5867	denote the | of
0.15625	-13.5867	measures | are
0.15	-13.5867	of | itself
0.0978593	-13.5867	is | grammar
0.143791	-13.5868	basic | for
0.268293	-13.5869	which | response
0.666667	-13.587	structured | (daum´e
0.666667	-13.587	usually | text.
0.0842572	-13.587	in | task.
0.333333	-13.587	introduced | semantic
0.294118	-13.5871	build | it
0.8	-13.5871	23 | 32
0.186813	-13.5871	semiring | the
0.4	-13.5872	an | effectively
0.275	-13.5873	are | variants
0.4375	-13.5874	models | perplexity
0.176471	-13.5875	classification accuracy | the
0.666667	-13.5877	many | via
0.236364	-13.5877	can | long
0.205479	-13.5877	a | candidates.
0.0648148	-13.5879	definition | the
0.8	-13.5879	final | actual
0.225806	-13.588	a | table.
0.171429	-13.588	on | individual
0.4	-13.588	reading | at
0.0413978	-13.588	the | form
0.145695	-13.588	supervision | the
0.346154	-13.5881	review | work
0.3125	-13.5882	for | argument.
0.215385	-13.5882	domain | this
0.195122	-13.5882	the | detector
0.268293	-13.5883	the | team.
0.0145013	-13.5883	and | for
0.666667	-13.5884	preceding | additional
0.245283	-13.5884	if | result
0.625	-13.5885	offers | performance
0.461538	-13.5886	illustrated | example
0.363636	-13.5886	procedure | described
0.152174	-13.5886	are | knowledge
0.132597	-13.5886	the | categories,
0.0187679	-13.5888	the | set
0.8	-13.5888	advantages | approaches.
0.8	-13.5888	after | collection
0.8	-13.5888	accuracy | low,
0.0782779	-13.5889	whereas | the
0.211268	-13.5889	model | bilingual
0.245283	-13.5889	that | for | tasks.
0.2	-13.589	averaged | and
0.289474	-13.5891	have | classes
0.294118	-13.5893	since | make
0.10453	-13.5894	score | =
0.255319	-13.5894	images | a
0.461538	-13.5895	have | senses,
0.0453956	-13.5895	each | is | of
0.0945559	-13.5896	its | on
0.222222	-13.5896	of | category,
0.202703	-13.5896	have | specific
0.321429	-13.5897	reporting | and
0.16	-13.5897	each | generated
0.0700935	-13.5897	and | types
0.0619684	-13.5897	name | the
0.125	-13.5897	the | successfully
0.625	-13.5898	users | interface
0.161017	-13.5898	that | ones
0.15748	-13.5899	summary | is
0.225806	-13.5899	also | classification
0.321429	-13.59	rest | follows:
0.171429	-13.5901	on | points
0.0858469	-13.5901	the | interaction
0.4375	-13.5902	while | already
0.4375	-13.5902	stack, | and
0.097561	-13.5903	reordering | in
0.0918919	-13.5903	automatic | in
0.169811	-13.5904	used | system.
0.8	-13.5904	film | insulating
0.8	-13.5904	〈xi,yi〉 | {〈xi,yi〉}
0.8	-13.5904	null(null1) | null(null2)
0.8	-13.5904	clasification | act
0.8	-13.5904	#(ct,w) | +µ
0.8	-13.5904	bacteria | ncbi
0.8	-13.5904	arcleft | arcright
0.8	-13.5904	8> | >:
0.8	-13.5904	letters, | notes.
0.8	-13.5904	68.5 | 75.4
0.8	-13.5904	s−1summationdisplay | ρkj(0)]
0.8	-13.5904	ˆc | mink
0.8	-13.5904	product | mattress
0.8	-13.5904	0.075 | β
0.8	-13.5904	characters | latin
0.8	-13.5904	verdient | die
0.8	-13.5904	gw | gwb50
0.8	-13.5904	florist | grocer
0.8	-13.5904	(tetreault | chodorow,
0.8	-13.5904	0.83 | 0.52
0.8	-13.5904	93.2 | uas
0.8	-13.5904	unified | interleaving
0.8	-13.5904	ships | sunk
0.8	-13.5904	0.77 | 0.81
0.8	-13.5904	badminton | indoors
0.8	-13.5904	it’s | just
0.8	-13.5904	(kim | ohta
0.8	-13.5904	mse | msew
0.8	-13.5904	reconstructed | frames
0.8	-13.5904	argc | arga
0.8	-13.5904	f3 | e4
0.8	-13.5904	empirical | discounts
0.8	-13.5904	19 | 46
0.8	-13.5904	synchronous | adjoining
0.8	-13.5904	>= | a*x+b
0.8	-13.5904	quasi-synchronous | smith,
0.8	-13.5904	phenotype | studies.
0.8	-13.5904	unified | (umls)
0.8	-13.5904	“plant” | “factory”
0.8	-13.5904	agrees | disagrees
0.8	-13.5904	0.31 | 0.48
0.8	-13.5904	woz-08 | woz-07
0.8	-13.5904	peter | a.
0.3125	-13.5905	a | literature.
0.555556	-13.5905	pairwise | whether
0.461538	-13.5905	sentence | tokens.
0.461538	-13.5905	has | performance,
0.625	-13.5905	categories | sense
0.333333	-13.5905	in | coming
0.215385	-13.5905	results | overall
0.0967742	-13.5906	latent | of
0.411765	-13.5907	problem | unsupervised
0.202703	-13.5907	and | optimize
0.333333	-13.5908	show | yields
0.333333	-13.5908	formula, | the
0.169811	-13.5909	reasonable | to
0.380952	-13.5909	figure | histogram
0.461538	-13.591	characters, | or
0.4375	-13.5911	order | reliable
0.411765	-13.5911	only | tags.
0.625	-13.5911	sampler | samples
0.666667	-13.5912	1995; | 2006).
0.176471	-13.5912	a | c.
0.126904	-13.5912	results | other
0.129032	-13.5913	do | use
0.205479	-13.5913	weighted | with
0.195122	-13.5913	in | frequencies
0.222222	-13.5914	the | mined
0.461538	-13.5914	additionally, | system
0.151079	-13.5915	the | domain-specific
0.555556	-13.5916	developed | statistical
0.346154	-13.5917	where | potential
0.294118	-13.5917	comparison | performance
0.188235	-13.5919	larger | with
0.236364	-13.5919	classifier | accuracy
0.225806	-13.592	c | g
0.363636	-13.592	extract | web
0.107011	-13.5921	and | measures
0.555556	-13.5922	aac | word
0.117904	-13.5924	and | technology
0.461538	-13.5924	relative | nn
0.4375	-13.5925	method | combining
0.275	-13.5925	with | attributes
0.0700935	-13.5925	statistical | and
0.144737	-13.5925	form | or
0.163793	-13.5926	needed | a
0.197531	-13.5926	• | only
0.289474	-13.5927	set, | only
0.236364	-13.5928	grammar, | to
0.101639	-13.593	with | corresponding
0.140127	-13.5931	r2 | r2
0.255319	-13.5931	these | significant
0.225806	-13.5932	these | instances
0.666667	-13.5933	original | preserved.
0.0842825	-13.5933	common | to
0.211268	-13.5933	tasks, | is
0.555556	-13.5934	importantly, | but
0.25	-13.5934	huge | and
0.0895141	-13.5936	users | to
0.115385	-13.5936	the | beam
0.333333	-13.5936	resources | english
0.321429	-13.5937	model | incorporating
0.625	-13.5937	cheaper | than
0.4375	-13.5938	1 | …
0.363636	-13.5939	approach | alignments
0.30303	-13.5939	according | test
0.666667	-13.594	association | pivot
0.666667	-13.594	snapshot | on
0.121495	-13.594	basic | in
0.555556	-13.594	n-best | asr
0.282051	-13.5942	by | adjective
0.208333	-13.5942	however, | found
0.236364	-13.5943	ngram | of
0.197531	-13.5944	we | giza++
0.141935	-13.5944	based | similarity
0.116379	-13.5945	of | acts
0.0890585	-13.5946	thus, | and
0.666667	-13.5947	3-gram | method
0.0921409	-13.5947	here | in
0.105634	-13.5948	recent | in
0.116883	-13.5948	capture | in
0.208333	-13.5948	are | along
0.208333	-13.5948	from | lexicon
0.190476	-13.5948	they | text
0.2	-13.595	approach | context
0.101974	-13.5951	relevant | in
0.15873	-13.5951	label | are
0.188235	-13.5952	results | high
0.145833	-13.5952	is | as:
0.232143	-13.5952	most | form
0.190476	-13.5952	cost | that
0.275	-13.5956	articles | have
0.26087	-13.5956	additionally | and
0.625	-13.5957	3 | 3.1 | models
0.333333	-13.5957	rhyme | of
0.333333	-13.5957	with | 2003),
0.208333	-13.5957	to | standard.
0.5	-13.5957	news | read
0.5	-13.5957	study | effectiveness
0.222222	-13.5958	order | need
0.461538	-13.5958	or | coherence
0.625	-13.5963	response | reference
0.0432721	-13.5963	to | its
0.321429	-13.5964	they | manually
0.145833	-13.5964	the | scaling
0.333333	-13.5964	handle | by
0.107807	-13.5965	weight | a
0.050695	-13.5966	a | phrase
0.186813	-13.5966	so | use
0.30303	-13.5968	relationships | are
0.282051	-13.5969	# | %
0.159664	-13.5969	noted | in
0.555556	-13.5969	good | coverage
0.192771	-13.597	2005), | a
0.625	-13.597	(mangu | al.,
0.625	-13.597	collection | resulting
0.3125	-13.5971	two | called
0.275	-13.5971	7.3 | the
0.411765	-13.5971	annotation | considered
0.0526316	-13.5971	on | set
0.0804124	-13.5972	terms | a
0.195122	-13.5972	and | free
0.333333	-13.5974	over | across
0.666667	-13.5975	among | difficult
0.666667	-13.5975	achieve | coverage
0.666667	-13.5975	other | ignored.
0.165217	-13.5977	that | expression
0.109434	-13.5977	partial | of
0.5	-13.5977	compute | equation
0.555556	-13.598	metrics | capture
0.0996785	-13.598	identical | the
0.113924	-13.598	direct | and
0.190476	-13.598	a | attachment
0.109434	-13.598	the | removed
0.244898	-13.5981	it | them.
0.666667	-13.5982	those | above.
0.666667	-13.5982	datasets | range
0.666667	-13.5982	into | hierarchy.
0.346154	-13.5982	0.7 | 0.5
0.225806	-13.5982	to | database.
0.2	-13.5982	suggested | a
0.625	-13.5983	iterations | final
0.106618	-13.5983	examples | are
0.11336	-13.5984	we | what
0.411765	-13.5984	thought | be
0.0978593	-13.5984	the | achieves
0.161017	-13.5984	classifier | training
0.180851	-13.5985	to | formal
0.168224	-13.5986	train | data
0.555556	-13.5986	being | (e.g.,
0.555556	-13.5986	uses | probabilistic
0.461538	-13.5987	consisted | two
0.30303	-13.5987	this | considers
0.15748	-13.5989	that | longer
0.225806	-13.5989	richer | of
0.192771	-13.599	“the | a
0.625	-13.599	compositionality | such
0.25	-13.5991	those | only
0.555556	-13.5992	misc | url
0.240741	-13.5992	kappa | for
0.282051	-13.5993	large-scale | for
0.380952	-13.5993	using | alignment.
0.380952	-13.5993	selecting | words
0.333333	-13.5995	training | instances.
0.666667	-13.5996	pragmatic | (e.g.
0.0434007	-13.5996	5 | of
0.30303	-13.5996	leverages | the
0.111554	-13.5996	token | in
0.380952	-13.5997	def= | =
0.461538	-13.5997	object | type
0.0942857	-13.5997	of | identified
0.177083	-13.5997	types | have
0.294118	-13.6	material | research
0.111111	-13.6	which | over
0.411765	-13.6001	mathematical | that
0.666667	-13.6003	cues | scope
0.178947	-13.6004	use | large
0.555556	-13.6004	and | interpolate
0.12766	-13.6004	finally, | with
0.109375	-13.6006	set | feature
0.142857	-13.6006	suggests that | to
0.09375	-13.6006	of | i.e.,
0.346154	-13.6008	note | rules
0.225806	-13.6008	as | not.
0.321429	-13.6008	prevents | of
0.12766	-13.6009	in | tree.
0.625	-13.6009	path | word.
0.5	-13.6009	but | implicit
0.168224	-13.601	to | structural
0.275	-13.6011	variety | different
0.159664	-13.6011	hypothesis | with
0.0840909	-13.6012	of | comparable
0.3125	-13.6013	of all, | to
0.0638853	-13.6013	annotation | in
0.294118	-13.6013	not | completely
0.294118	-13.6013	results | our | outperforms
0.0586932	-13.6013	the | cluster
0.0861244	-13.6015	on | can be
0.555556	-13.6015	field | its
0.197531	-13.6017	failure | the
0.666667	-13.6017	classifier | translationese
0.225806	-13.6017	in | field | of
0.110236	-13.6017	descriptions | the
0.116883	-13.6018	decision | a
0.244898	-13.6018	classification | sentiment
0.139241	-13.602	selection | on
0.4375	-13.6021	set | sampled
0.4375	-13.6021	sets | then
0.555556	-13.6021	million | 2010).
0.321429	-13.6022	in | scale.
0.411765	-13.6023	as | zero.
0.177083	-13.6024	in | sections
0.380952	-13.6025	of | cats
0.232143	-13.6025	once | have
0.3125	-13.6026	only | information.
0.15625	-13.6026	very | not
0.155039	-13.6027	presented | table
0.116883	-13.6027	compared to | and
0.555556	-13.6027	choose | +
0.363636	-13.6027	lie | and
0.411765	-13.6027	findings | reflect
0.411765	-13.6027	y(q)j | (
0.4375	-13.603	way | knowledge
0.4375	-13.603	binding | protein
0.232143	-13.603	distributional | semantic
0.5	-13.603	i | k.
0.666667	-13.6031	polarity | choi
0.666667	-13.6031	apply | predicate
0.236364	-13.6033	sentences | human
0.555556	-13.6033	work | template
0.177083	-13.6033	with | scores.
0.107011	-13.6035	(b) | and
0.15748	-13.6036	that | morphological
0.625	-13.6036	method | sparse
0.244898	-13.6036	we | reach
0.111111	-13.6036	indicate | and
0.282051	-13.6037	a | item.
0.346154	-13.6037	the | time).
0.380952	-13.6037	words | compound
0.182796	-13.6037	used | previous
0.0942857	-13.6037	l | of
0.666667	-13.6038	topics | papers
0.666667	-13.6038	context | n−1
0.666667	-13.6038	how | relates
0.666667	-13.6038	he | him
0.363636	-13.6038	2003; | 2004),
0.4375	-13.6039	correct, | as
0.3125	-13.604	entails | and
0.461538	-13.604	initialize | do
0.212121	-13.604	represent | between
0.333333	-13.6041	of | visible
0.148936	-13.6041	affect | and
0.363636	-13.6042	2002; | 2002;
0.625	-13.6042	(uzuner | al.,
0.0603248	-13.6042	define | of
0.102389	-13.6044	a | texts
0.294118	-13.6044	fact | some
0.30303	-13.6044	performs | when
0.666667	-13.6045	contrast | continuation
0.666667	-13.6045	acl | cited
0.461538	-13.6045	drop | performance
0.22807	-13.6046	and | strict
0.26087	-13.6047	corpus | wikipedia
0.346154	-13.6048	different | additional
0.240741	-13.6048	a | supporting
0.625	-13.6049	reliable | corpus.
0.0687023	-13.605	the | forms
0.155039	-13.605	(if | the
0.135294	-13.605	types | or
0.162393	-13.605	all | score
0.09375	-13.6051	simply | of
0.666667	-13.6052	via | italy
0.666667	-13.6052	discourse | conversation
0.380952	-13.6053	reordering | syntactic
0.380952	-13.6053	of | intervening
0.117117	-13.6054	with | bleu
0.0355699	-13.6055	and | used
0.168224	-13.6056	gender | in
0.625	-13.6056	espresso | seed
0.0405904	-13.6056	in | where
0.232143	-13.6057	of | j,
0.26087	-13.6058	system’s | is
0.666667	-13.6059	should | evidence
0.346154	-13.6059	the | there.
0.0822511	-13.6059	a | sentiment
0.461538	-13.606	ranking | previous
0.104895	-13.606	the | analysis.
0.380952	-13.6061	equation | our
0.141935	-13.6061	in | description
0.1	-13.6062	the | seems
0.555556	-13.6062	find | smaller
0.625	-13.6062	instead | now
0.625	-13.6062	simple: | is
0.135294	-13.6062	for | presented
0.0550725	-13.6062	an | is | in
0.236364	-13.6063	chose | that
0.148936	-13.6063	it | make
0.0721477	-13.6063	to | search
0.173077	-13.6064	using | task
0.380952	-13.6065	made | information
0.141935	-13.6065	best | using
0.255319	-13.6068	segmentation | using
0.3125	-13.6069	demonstrate | approach
0.321429	-13.6069	produces | than
0.146853	-13.6071	to | property
0.666667	-13.6073	introduce | concept
0.666667	-13.6073	detectors | trained
0.25	-13.6073	two | measures
0.363636	-13.6073	prior | both
0.126316	-13.6074	by | topic
0.0301187	-13.6074	are | that
0.0513036	-13.6075	function | the | of
0.053654	-13.6075	automatically | the
0.321429	-13.6076	quantify | in
0.0779727	-13.6076	to | what
0.122549	-13.6076	the | suggested
0.0890585	-13.6077	each | sentence
0.0881612	-13.6077	for | –
0.0716667	-13.6077	of | general
0.00640165	-13.6078	the | in
0.25	-13.6078	components: | the
0.130435	-13.6079	that | users
0.113924	-13.6079	research | on
0.1125	-13.608	with | higher
0.171429	-13.6081	models | these
0.625	-13.6082	conditional | features,
0.165217	-13.6083	that | contextual
0.0911528	-13.6083	of | translation.
0.190476	-13.6083	inflected | the
0.0607867	-13.6084	function | and
0.169811	-13.6086	mwes | in
0.666667	-13.6087	has | broader
0.666667	-13.6087	both | significantly.
0.666667	-13.6087	estimation | smoothing
0.666667	-13.6087	which | violated
0.146853	-13.6088	process | from
0.118182	-13.6088	results | when
0.5	-13.6088	4, | better
0.5	-13.6088	tags | annotator
0.0195647	-13.6089	from | and
0.0861244	-13.6089	texts | and
0.461538	-13.6089	feedback | help
0.144828	-13.609	in | characters
0.100649	-13.6091	the | of | approach.
0.128342	-13.6094	in | edges
0.461538	-13.6094	update | current
0.461538	-13.6094	approach | cost
0.232143	-13.6094	agent | is
0.232143	-13.6094	names | be
0.625	-13.6095	a |  a
0.625	-13.6095	baselines | random
0.625	-13.6095	various | models.
0.244898	-13.6095	defined | such
0.192771	-13.6096	an | (e.g.
0.133721	-13.6096	for | resources
0.363636	-13.6096	support | information
0.0820734	-13.6096	we also | for
0.045098	-13.6097	then | and
0.555556	-13.6098	2008; | chodorow,
0.050571	-13.6099	many | and
0.4375	-13.61	shows | change
0.4375	-13.61	available | them
0.0960961	-13.61	4.1 | and
0.0808511	-13.61	and | content
0.135294	-13.6101	we | scheme
0.261905	-13.6102	however, | likely
0.195122	-13.6102	by | overall
0.136646	-13.6103	are | labels
0.0881612	-13.6103	resources | and
0.208333	-13.6104	several | used
0.0792683	-13.6104	to | we use
0.380952	-13.6105	problem | solution
0.411765	-13.6106	only | 50
0.411765	-13.6106	will | efforts
0.117647	-13.6107	since | an
0.0837104	-13.6107	see | in
0.625	-13.6109	setup | pairs
0.625	-13.6109	metric | computed
0.0791075	-13.6109	method | with
0.268293	-13.6112	formulas | the
0.0664781	-13.6113	we | experiments
0.0196248	-13.6113	the | are | to
0.15625	-13.6114	can be | same
0.282051	-13.6115	manual | not
0.190476	-13.6115	accuracy | more
0.064603	-13.6118	adding | the
0.0963855	-13.612	argument | to
0.158333	-13.6121	in | clear
0.119816	-13.6122	of | responses
0.666667	-13.6123	different | satisfaction,
0.411765	-13.6123	all | substrings
0.163793	-13.6124	toolkit | and
0.461538	-13.6124	or | pronouns
0.0844749	-13.6124	on | text
0.0635538	-13.6125	the | against
0.346154	-13.6125	of | session.
0.215385	-13.6125	extensions | of
0.161017	-13.6125	for | people
0.3125	-13.6126	such | gender
0.132184	-13.6127	parameter | is
0.321429	-13.6127	first, | sentences
0.285714	-13.6128	reasons | this
0.143836	-13.6128	the | is done
0.212121	-13.6128	layer | and
0.411765	-13.6128	approaches | space
0.115385	-13.6129	the | relations,
0.333333	-13.6129	seed | target
0.138365	-13.613	and | character
0.107807	-13.613	of | world
0.157025	-13.6131	distributions | for
0.157025	-13.6131	the | expectations
0.232143	-13.6131	tool | with
0.0483019	-13.6132	method | to
0.188235	-13.6133	of | integration
0.555556	-13.6133	resources | data.
0.0744681	-13.6134	in | analysis
0.261905	-13.6135	given | could
0.625	-13.6135	ic | methods
0.0376417	-13.6135	the | in section
0.0773694	-13.6136	set | set of
0.333333	-13.6136	models | observed
0.5	-13.6136	@ | .
0.5	-13.6136	identical | pairs
0.666667	-13.6137	“avg” | the
0.4375	-13.6137	model | billion
0.145833	-13.6137	to | languages,
0.411765	-13.6137	time-consuming | to
0.321429	-13.6138	into | languages.
0.321429	-13.6138	gigaword | data
0.363636	-13.6139	their | obtained
0.0987261	-13.6139	additionally, | the
0.205479	-13.6139	direct | with
0.129032	-13.614	metric | is
0.119266	-13.614	relatedness | the
0.4375	-13.6141	results | particularly
0.625	-13.6142	bulgarian | slovene
0.625	-13.6142	in proceedings of | 2007.
0.625	-13.6142	nullnull[ | ]
0.625	-13.6142	¯f) | ¯f
0.625	-13.6142	sulins, | graeser,
0.625	-13.6142	[1/ | [2/
0.166667	-13.6144	only those | the
0.165138	-13.6144	a | open
0.255319	-13.6144	our | complete
0.255319	-13.6144	are | labels.
0.155039	-13.6145	compound | a
0.555556	-13.6145	when | senses,
0.25	-13.6145	purity | the
0.411765	-13.6145	low | higher
0.285714	-13.6146	was | designed
0.4375	-13.6146	names | annotated
0.380952	-13.6146	while | enough
0.103093	-13.6146	to capture | the
0.347826	-13.6147	on | 5).
0.188235	-13.6147	to | sparse
0.129032	-13.6148	straightforward | the
0.0470085	-13.6148	2 | in
0.380952	-13.615	extra | with
0.380952	-13.615	parser | labeled
0.666667	-13.6151	even | ones
0.666667	-13.6151	nucle | data
0.346154	-13.6151	comparable | documents
0.222222	-13.6151	these | being
0.275	-13.6151	models | even
0.555556	-13.6151	2005) | 2006).
0.129032	-13.6151	group | a
0.5	-13.6152	recent | 2008),
0.461538	-13.6153	conversion | trees
0.108527	-13.6156	we | times
0.109804	-13.6156	training | translation
0.161017	-13.6156	expected | for
0.666667	-13.6158	times, | found
0.666667	-13.6158	where | changed
0.261905	-13.6158	to | clustering.
0.461538	-13.6158	word-aligned | pairs
0.22807	-13.6158	where | does
0.411765	-13.6159	0.3 | 0.45
0.202703	-13.6159	1 | function
0.103093	-13.6159	and | inference
0.244898	-13.616	in | gaps
0.183908	-13.6161	features | parsing
0.380952	-13.6162	all | 0.
0.0403776	-13.6162	could | the
0.411765	-13.6163	by | newspaper
0.30303	-13.6164	a | alphabet
0.138365	-13.6165	we | discriminative
0.666667	-13.6165	modeling | predicting
0.22807	-13.6165	in | approaches,
0.132948	-13.6168	the | tokens.
0.0745455	-13.6168	approach | by
0.5	-13.6168	preferences | between
0.346154	-13.6169	numbers | were
0.555556	-13.6169	interpolated | model.
0.555556	-13.6169	languages | therefore
0.22807	-13.6173	future work | from
0.149254	-13.6173	the | ngram
0.123153	-13.6174	figure 4 | the
0.166667	-13.6176	achieves | on
0.21875	-13.6176	remaining | is
0.125	-13.6176	we also | from
0.0820399	-13.6176	a | out
0.173469	-13.6177	as | ones
0.461538	-13.6178	overall, | accuracy
0.333333	-13.6179	2007; | 2007),
0.166667	-13.618	approaches | have
0.132184	-13.6182	maximum | for
0.333333	-13.6183	words | categories
0.30303	-13.6183	they | certain
0.104317	-13.6183	constructed | of
0.282051	-13.6184	shows | percentage
0.275	-13.6184	messages | be
0.1	-13.6184	the | method,
0.0996785	-13.6184	correct | a
0.5	-13.6184	like | require
0.205479	-13.6185	correct | all
0.161017	-13.6185	are | position
0.0691824	-13.6186	algorithms | the
0.152672	-13.6187	word | corpus.
0.666667	-13.6187	node | derivation
0.555556	-13.6187	moreover, | scores
0.162393	-13.6187	and | university,
0.294118	-13.6188	labeled | other
0.124378	-13.619	for | tokens
0.2	-13.619	this | leads
0.5	-13.619	preliminary | when
0.099359	-13.6191	manual | and
0.555556	-13.6193	select | optimal
0.461538	-13.6193	observed | scores
0.666667	-13.6194	then, | discourse
0.208955	-13.6194	words | after
0.0892388	-13.6194	of | forms
0.114894	-13.6194	are | rules
0.173469	-13.6195	to | tags.
0.285714	-13.6195	(except | the
0.175258	-13.6195	using | scores
0.244898	-13.6195	from | 100
0.5	-13.6195	improvements | additional
0.294118	-13.6197	smt | (koehn
0.294118	-13.6197	words | meanings
0.363636	-13.6197	arg2 | arg2
0.125654	-13.6197	know | of
0.411765	-13.6198	5 | 150
0.388889	-13.6199	for | children’s
0.666667	-13.6201	fauc(data|a)
0.666667	-13.6201	largest | followed
0.275	-13.6201	their | whether
0.4375	-13.6202	method | improved
0.347826	-13.6203	convergence | with
0.232143	-13.6203	if | find
0.202703	-13.6203	to | questions.
0.363636	-13.6205	impact | parsing
0.321429	-13.6206	n | n-gram
0.30303	-13.6206	therefore, | set
0.347826	-13.6207	require | all
0.380952	-13.6207	assumption | its
0.411765	-13.6207	scale | 0
0.282051	-13.6208	pos-tag | the
0.0558376	-13.6208	the | relative
0.020729	-13.621	we | on
0.180851	-13.621	11 | in
0.5	-13.6211	from | ontology,
0.25	-13.6212	dataset | different
0.145833	-13.6213	asr | and
0.120192	-13.6214	the | style
0.666667	-13.6215	temporal | machines
0.380952	-13.6215	since | sentences.
0.294118	-13.6216	tagged | from
0.145833	-13.6216	and | extended
0.166667	-13.6217	[ | ]
0.168224	-13.6218	with | subject
0.136646	-13.6218	the | encoding
0.141892	-13.6218	to | intended
0.192771	-13.6219	a | predictive
0.380952	-13.6219	treebank | annotated
0.411765	-13.622	global | language
0.411765	-13.622	80% | top
0.134969	-13.6221	method | each
0.294118	-13.6222	repeatedly | to
0.165138	-13.6222	metric | by
0.5	-13.6222	lead | significant
0.232143	-13.6223	are | classified
0.255319	-13.6223	as | −
0.388889	-13.6224	• | i.e.
0.12766	-13.6224	relations | by
0.11157	-13.6225	otherwise | the
0.411765	-13.6225	0.85 | 0.87
0.149254	-13.6225	i | k
0.333333	-13.6226	has | written
0.275	-13.6228	statistic | of
0.133721	-13.6228	when | have
0.21875	-13.6228	sometimes | as
0.268293	-13.6229	configurations | in
0.197368	-13.6229	for | purpose,
0.236364	-13.6229	knight, | et
0.103571	-13.6229	(the | and
0.666667	-13.623	finally, | 4).
0.0633245	-13.623	in | model.
0.11157	-13.623	models | or
0.134503	-13.6232	performance | this
0.347826	-13.6233	summationtextn | of
0.0767754	-13.6233	the | 20
0.178947	-13.6234	words | lexical
0.321429	-13.6234	the | aspect.
0.555556	-13.6234	domain | make
0.294118	-13.6235	most | during
0.380952	-13.6236	model | oracle
0.175258	-13.6236	alignment | between
0.157025	-13.6237	from | aligned
0.666667	-13.6237	outperform | without
0.666667	-13.6237	query | qm
0.666667	-13.6237	over | months
0.159664	-13.6237	the | library
0.24	-13.6238	to handle | in
0.261905	-13.6239	dependent | as
0.4375	-13.624	observed | data,
0.224138	-13.624	a | sensitive
0.161017	-13.624	for | short
0.388889	-13.6241	ones, | that
0.132948	-13.6241	they | but
0.103093	-13.6241	3.2 | to
0.0855107	-13.6243	3 | 0
0.123153	-13.6244	labeling | of
0.555556	-13.6246	fragments | same
0.555556	-13.6246	inference | local
0.333333	-13.6247	slightly | it
0.363636	-13.6248	2 | 3,
0.25	-13.6249	work | nsf
0.14094	-13.6249	contextual | in
0.5	-13.6249	forms | morphological
0.168224	-13.625	to | comments
0.666667	-13.6251	combination | character-based
0.380952	-13.6252	the | scratch.
0.380952	-13.6252	language | community
0.1375	-13.6252	sequence | that
0.181818	-13.6253	comes | and
0.244898	-13.6255	key | model
0.183908	-13.6255	nnp | np
0.125654	-13.6255	that, | is
0.30303	-13.6255	= | ≥
0.188235	-13.6256	by | categories
0.261905	-13.6256	they | general
0.388889	-13.6258	they | definition
0.321429	-13.6258	have | etc.
0.244898	-13.6258	for | tweet
0.190476	-13.6258	to | posts
0.192771	-13.6259	2011) | 2011)
0.363636	-13.6259	feature | measures
0.0716667	-13.626	there is | to
0.0515734	-13.626	this | more
0.045859	-13.6261	any | of
0.171429	-13.6261	is | technique
0.215385	-13.6261	2008) | on
0.173469	-13.6262	for | understanding
0.060355	-13.6262	all | by
0.232143	-13.6263	systems | better
0.24	-13.6263	data | systems.
0.21875	-13.6263	threshold, | the
0.126316	-13.6264	graph, | the
0.294118	-13.6264	which | expressed
0.0724138	-13.6265	in | parsing
0.0696203	-13.6266	describe | and
0.24	-13.6266	each | predicate
0.0904255	-13.6266	3 | are
0.461538	-13.6267	had | book
0.21875	-13.6268	in | users.
0.066092	-13.6269	to | standard
0.555556	-13.627	semantic | cross-lingual
0.347826	-13.6271	with | phrases,
0.0984127	-13.6271	and | particular
0.10728	-13.6271	in | generation
0.186047	-13.6272	in | occurrence
0.461538	-13.6272	two | topics.
0.461538	-13.6272	function | assigns
0.666667	-13.6273	build | complete
0.666667	-13.6273	obviously | all
0.0833333	-13.6273	the | direct
0.411765	-13.6274	of | street,
0.30303	-13.6274	described, | the
0.236364	-13.6275	aligned | source
0.0670641	-13.6275	previous | in
0.14094	-13.6275	s | p
0.21875	-13.6275	bilingual | corpora
0.103943	-13.6276	3.2 | and
0.117117	-13.6278	performance | model
0.147059	-13.6279	successful | the
0.215385	-13.628	our | yields
0.151515	-13.6281	parses | and
0.380952	-13.6281	few | all
0.224138	-13.6281	by | decoder
0.121359	-13.6281	the | her
0.0307503	-13.6283	the | text
0.388889	-13.6283	reverse | not
0.091922	-13.6283	of | results.
0.22807	-13.6285	are | linked
0.0740072	-13.6286	development | of
0.202703	-13.6286	search space | of
0.0851064	-13.6287	that | significantly
0.162162	-13.6287	k, | the
0.666667	-13.6287	dialogue | indicating
0.666667	-13.6287	gaps | target
0.461538	-13.6287	possible | y
0.212121	-13.6287	third | that
0.411765	-13.6287	support vector | learning
0.333333	-13.6287	in | ungrammatical
0.347826	-13.6289	combination | confusion
0.2	-13.6289	as | global
0.205882	-13.6292	(7) | a
0.4375	-13.6292	experiments | models,
0.157025	-13.6293	4 | different
0.162162	-13.6293	in | al
0.100334	-13.6293	we | instead
0.0700326	-13.6294	verb | a
0.21875	-13.6294	for | children
0.666667	-13.6295	chance | very
0.236364	-13.6295	= | k)
0.363636	-13.6295	two | groups
0.275	-13.6296	we add | that
0.411765	-13.6296	automatically | user
0.232143	-13.6298	features | document
0.116592	-13.6299	for | individual
0.0808511	-13.6299	our | other
0.333333	-13.6301	(2010) and | et
0.21875	-13.6301	method | baseline
0.161017	-13.6301	each | labeled
0.461538	-13.6302	features | citation
0.0675676	-13.6302	common | of
0.220339	-13.6303	we | working
0.22807	-13.6303	a | phase
0.190476	-13.6303	from | mt
0.215385	-13.6303	to | newly
0.0362381	-13.6303	the | was | and
0.5	-13.6303	network | node
0.178947	-13.6304	seems | for
0.0144241	-13.6304	the | of | as
0.0962733	-13.6305	this method | the
0.0568421	-13.6305	in | learning
0.0868486	-13.6305	the | technique
0.411765	-13.6305	communication | such
0.21875	-13.6308	behavior | is
0.188235	-13.6309	the | questions,
0.666667	-13.6309	lexical, | cues
0.666667	-13.6309	some | σ
0.129213	-13.6309	identification | a
0.183908	-13.6309	state | each
0.129944	-13.631	construction | in
0.380952	-13.631	an | database.
0.224138	-13.631	techniques | or
0.0787879	-13.631	more | or
0.046729	-13.6311	the | instances
0.285714	-13.6312	nlp | has
0.555556	-13.6312	using | continuous
0.461538	-13.6312	nn | subj
0.30303	-13.6314	problems | data
0.215385	-13.6315	since | possible
0.666667	-13.6316	infrequent | large
0.104693	-13.6316	on | sets
0.461538	-13.6317	same | replaced
0.555556	-13.6318	evaluated | random
0.0182748	-13.6319	the | of | from
0.333333	-13.6319	they | complex
0.244898	-13.6319	to | efficiency
0.388889	-13.6321	the | rows.
0.294118	-13.6321	evaluate | compare
0.119048	-13.6321	introduced | to
0.0281911	-13.6321	results | the | of
0.21875	-13.6322	without | it
0.148148	-13.6322	features | extracted
0.333333	-13.6323	core | our
0.2	-13.6323	set, | set
0.188235	-13.6324	is | memory
0.0691824	-13.6324	2008; | et
0.4375	-13.6325	if | continue
0.380952	-13.6326	100 | 25
0.363636	-13.6326	of | schemes,
0.255814	-13.6326	much | when
0.30303	-13.6327	by | induced
0.169811	-13.6328	it | domain
0.108108	-13.633	the | outside
0.0522052	-13.6331	words, | the
0.380952	-13.6331	manually | into
0.0374134	-13.6331	not | for
0.310345	-13.6332	figure 2 | example
0.117117	-13.6332	as | gold
0.275	-13.6332	training | weights
0.112971	-13.6333	the | arcs
0.168224	-13.6334	at | i
0.333333	-13.6334	uses | syntactic
0.106464	-13.6334	children | the
0.127778	-13.6336	there | also
0.411765	-13.6336	in | lexicons,
0.118483	-13.6337	random | is
0.461538	-13.6337	which | contribute
0.347826	-13.6338	a | frequencies.
0.388889	-13.6338	and | reviewed
0.188235	-13.6338	most | sentences
0.132948	-13.6338	learning | model
0.310345	-13.6339	some | structure
0.4375	-13.6339	and | maintenance
0.4375	-13.6339	learning | ccg
0.129213	-13.6339	aim | and
0.2	-13.6339	drawback | the
0.294118	-13.634	large | very
0.117925	-13.634	check | check
0.388889	-13.6342	before | before
0.388889	-13.6342	outperforms | accuracy
0.555556	-13.6343	binary | (i.e.,
0.117117	-13.6344	j | =
0.186047	-13.6345	and | media
0.151515	-13.6345	task | extraction
0.0480183	-13.6345	by | or
0.411765	-13.6345	when | combine
0.244898	-13.6345	query | word
0.347826	-13.6346	all | queries
0.14094	-13.6346	to | especially
0.119617	-13.6348	performing | and
0.310345	-13.6349	contextual | word
0.555556	-13.6349	pos tags | dependency
0.555556	-13.6349	application | annotation
0.555556	-13.6349	holds, | that
0.145985	-13.6349	presence | a
0.175258	-13.635	for | strategy
0.388889	-13.6351	wsd | words
0.0859951	-13.6351	9 | the
0.205882	-13.6352	work | techniques
0.192308	-13.6352	ter | bleu
0.666667	-13.6352	curran | their
0.14	-13.6352	scaling | the
0.30303	-13.6353	examples | sentence
0.127778	-13.6354	- | with
0.10566	-13.6354	finally, | for
0.0872818	-13.6355	is | higher
0.17	-13.6356	= | z
0.202703	-13.6356	manually | or
0.25	-13.6358	identify | sentences
0.411765	-13.6358	(2008), | which
0.115556	-13.6359	5.1 | and
0.255814	-13.6359	sentences | many
0.333333	-13.6359	an | text,
0.333333	-13.6359	number | various
0.555556	-13.6361	nonlinear | supporting
0.212121	-13.6361	of | vs
0.363636	-13.6362	example | cause
0.294118	-13.6363	instances | both
0.0764706	-13.6363	evaluation | on
0.168224	-13.6365	to | preceding
0.155738	-13.6365	use | similar
0.173469	-13.6366	data | so
0.666667	-13.6367	processing | limited
0.666667	-13.6367	so | introducing
0.666667	-13.6367	transformations | those
0.411765	-13.6367	acted | the
0.388889	-13.6368	all | errors.
0.5	-13.6368	because | relations.
0.310345	-13.6369	given | set,
0.148148	-13.637	information, | to
0.555556	-13.6373	(8) | probability
0.555556	-13.6373	same | ...
0.255814	-13.6373	(1) | 0
0.5	-13.6373	who | them
0.5	-13.6373	we | calibrated
0.133333	-13.6375	a particular | in
0.277778	-13.6375	we | 400
0.321429	-13.6377	algorithm | n
0.0977918	-13.6377	even | are
0.139073	-13.6381	be | within
0.119048	-13.6381	a | wordnet
0.277778	-13.6381	important | about
0.666667	-13.6382	split | segments
0.666667	-13.6382	might | noisy
0.104317	-13.6383	on | analysis
0.411765	-13.6385	parser | some
0.411765	-13.6385	q | k
0.461538	-13.6387	literal | it
0.461538	-13.6387	for | mention-entity
0.461538	-13.6387	mert | weight
0.0689655	-13.6388	for | while
0.666667	-13.6389	role | mappings
0.108108	-13.6389	comparable | to
0.261905	-13.639	contribution | paper
0.5	-13.639	simple | matching
0.108527	-13.6391	significant | for
0.0993377	-13.6391	t | =
0.194805	-13.6392	in | precision.
0.30303	-13.6393	preliminary | our
0.24	-13.6394	from | 3,
0.145985	-13.6395	inference | that
0.666667	-13.6396	systems | ssc
0.555556	-13.6397	types | contextual
0.130682	-13.6397	for | maximum
0.24	-13.6397	is | obvious
0.235294	-13.6397	and | 2010a).
0.277778	-13.6397	in | array
0.321429	-13.6398	can | errors.
0.127778	-13.6399	main | is | in
0.129944	-13.64	for | parameter
0.0943953	-13.64	for | extracted
0.22807	-13.64	process | more
0.0357143	-13.64	the | tree
0.112033	-13.6401	better | with
0.122449	-13.6402	candidates | a
0.212121	-13.6402	results | classifier
0.347826	-13.6403	nlp | their
0.666667	-13.6403	wam | pairs
0.134969	-13.6403	test | were
0.154472	-13.6404	10 | 1
0.117371	-13.6405	the | transliteration
0.142857	-13.6408	means | as
0.162162	-13.6409	the | rightmost
0.255814	-13.6409	rewrite | and
0.0833333	-13.6409	the | u
0.112033	-13.641	are | used in
0.275	-13.641	abstract | important
0.149254	-13.641	this | take
0.333333	-13.641	the | differ.
0.666667	-13.6411	sequence | covering
0.666667	-13.6411	proceeds | follows.
0.183908	-13.6411	center | and
0.095092	-13.6412	in | various
0.277778	-13.6412	studies | between
0.5	-13.6412	performance | traditional
0.461538	-13.6413	manual | three
0.2	-13.6413	word | morphological
0.411765	-13.6417	parser | any
0.5	-13.6417	is | fragment.
0.5	-13.6417	times | size
0.25	-13.6418	to | threshold.
0.0831409	-13.6418	u | the
0.103571	-13.6419	we | techniques
0.157025	-13.642	approaches | by
0.215385	-13.642	from | base
0.347826	-13.6422	text | words.
0.0409519	-13.6423	type | the
0.461538	-13.6423	records | data
0.5	-13.6423	coherence | topics
0.275	-13.6425	given | example
0.347826	-13.6426	generation | target
0.321429	-13.6426	y, | =
0.411765	-13.6426	however, | little
0.30303	-13.6426	user | any
0.0475831	-13.6427	of | syntactic
0.0384995	-13.6427	the following | the
0.212121	-13.6428	product | for
0.208955	-13.6428	is | nps
0.5	-13.6428	form | written
0.133333	-13.6429	as | second
0.30303	-13.6429	remainder | section
0.119048	-13.6429	gender | the
0.0769231	-13.6429	the | describe
0.2	-13.643	to | concepts.
0.139073	-13.6431	is | strong
0.411765	-13.6431	dataset | first
0.0818584	-13.6433	lists | the
0.10566	-13.6433	of | basic
0.2	-13.6433	choosing | a
0.275	-13.6434	that | relations between
0.555556	-13.6434	tweets | topic.
0.0757576	-13.6434	4.2 | the | of
0.188235	-13.6435	aspects | are
0.224138	-13.6435	properties | from
0.411765	-13.6435	thus, | no
0.2	-13.6435	confident | the
0.139073	-13.6436	that | produced
0.130682	-13.6436	or | may
0.310345	-13.6437	annotators | given
0.294118	-13.6437	there | who
0.255814	-13.6438	before, | and
0.215385	-13.6439	proceedings | acl
0.294118	-13.644	our | summarized
0.100334	-13.644	a | made
0.168224	-13.6441	on | see
0.461538	-13.6443	ranking | compare
0.14094	-13.6444	for | results.
0.122449	-13.6445	second, | in
0.555556	-13.6446	consider | path
0.202899	-13.6446	an | whose
0.194805	-13.6447	keywords | and
0.347826	-13.6449	system | conll
0.388889	-13.6449	identify | corresponding
0.388889	-13.6449	the | profit
0.310345	-13.6451	tagger | data
0.0511265	-13.6452	of | event
0.11157	-13.6452	an | target
0.388889	-13.6453	for | ψ
0.285714	-13.6453	time | when
0.163636	-13.6454	with | <
0.666667	-13.6455	stems | (e.g.,
0.666667	-13.6455	approach | treebank.
0.666667	-13.6455	timegraph | temporal
0.666667	-13.6455	trie | less
0.666667	-13.6455	if | xi
0.224138	-13.6455	and | def=
0.148148	-13.6455	presence | in
0.220339	-13.6457	when | input
0.24	-13.6457	ng | a
0.555556	-13.6458	general | category
0.073741	-13.6458	abstract | to
0.165138	-13.6459	reduced | to
0.108108	-13.6459	however, | an
0.173469	-13.6461	generate | by
0.096875	-13.6461	training data | to
0.138158	-13.6462	our experiments | that
0.666667	-13.6462	spatial | features
0.666667	-13.6462	4.2 | distributions
0.666667	-13.6462	bayesian | sparse
0.261905	-13.6462	which | i.e.,
0.192308	-13.6463	the | embedding
0.555556	-13.6464	datasets | experiments.
0.114537	-13.6465	is | string
0.186047	-13.6465	one of | two
0.165138	-13.6465	was | some
0.175258	-13.6465	work | project
0.0977199	-13.6466	important | for
0.363636	-13.6466	f-scores | for
0.347826	-13.6468	do | afrl,
0.192308	-13.6468	order | improve
0.24	-13.6468	data, | using
0.666667	-13.6469	different | posts.
0.333333	-13.6469	paraphrase | task
0.333333	-13.6469	with | structure,
0.159292	-13.6469	4.1 | as
0.388889	-13.647	frameworks | et
0.134146	-13.6471	distinction | the
0.179775	-13.6471	a | vertex
0.0545635	-13.6472	is | given
0.071066	-13.6472	to | multiple
0.5	-13.6472	guide | model
0.24	-13.6473	approach | standard
0.363636	-13.6474	features | introduced
0.122449	-13.6475	word, | and
0.268293	-13.6475	a | recursively
0.166667	-13.6477	is | typically
0.090411	-13.6477	time | for
0.112554	-13.6477	the | pair,
0.0584192	-13.6478	the | maximum
0.208955	-13.6478	of | base.
0.0620155	-13.6479	to | single
0.121212	-13.6479	the | γ
0.144928	-13.6479	because | data
0.119048	-13.6479	for | language.
0.14094	-13.6481	observation | and
0.129944	-13.6482	3.1 | for
0.192308	-13.6482	represent | which
0.0469897	-13.6482	given | in
0.17	-13.6483	in | yield
0.194805	-13.6484	t | v
0.154472	-13.6484	a | preposition
0.294118	-13.6485	in | directions,
0.0833333	-13.6485	this | given
0.277778	-13.6485	has | lot
0.208955	-13.6485	as | hidden
0.160714	-13.6486	since | other
0.2	-13.6486	for | case,
0.388889	-13.6487	thus, | standard
0.095092	-13.6488	we | reference
0.0713073	-13.6489	do | for
0.310345	-13.6489	of | china.
0.461538	-13.6489	summary | was
0.14094	-13.6489	sizes | the
0.363636	-13.649	have | sign
0.333333	-13.6491	is | candidate.
0.0690209	-13.6492	of | present
0.321429	-13.6493	maybe | the
0.0878553	-13.6494	generally | the
0.5	-13.6494	detection | prosodic
0.145985	-13.6495	abstract | translation
0.0714286	-13.6496	correct | of
0.388889	-13.6496	free | which
0.022438	-13.6496	that | are
0.0853659	-13.6498	are | most
0.0308132	-13.6498	only | of
0.666667	-13.6499	defines | corresponding
0.666667	-13.6499	computing | posterior
0.101399	-13.6499	we use | al.,
0.149254	-13.6499	in | almost
0.152	-13.65	be | future
0.141892	-13.65	2007) | a
0.5	-13.65	algorithms | proposed
0.555556	-13.6501	(3) | z
0.277778	-13.6501	model, | its
0.24	-13.6502	it | expected
0.188235	-13.6503	model | applied
0.0415704	-13.6503	is | where
0.0990099	-13.6503	• | by
0.162162	-13.6504	this | developed
0.224138	-13.6504	the | combinations.
0.138158	-13.6505	all | no
0.0550459	-13.6505	phrase | in
0.077551	-13.6505	to | current
0.134146	-13.6505	= | (1) | a
0.153226	-13.6505	some | they
0.347826	-13.6506	lexical | text.
0.666667	-13.6506	2009) | further
0.666667	-13.6506	jian | and
0.363636	-13.6506	both | produce
0.333333	-13.6506	al | from
0.119617	-13.6507	5 | 7
0.555556	-13.6507	the | hmms.
0.555556	-13.6507	even | wrong
0.277778	-13.6507	evaluated | standard
0.179775	-13.6508	subset | that
0.388889	-13.6509	a | algorithmic
0.0629183	-13.6511	we | best
0.186047	-13.6512	scores | set
0.212121	-13.6512	data | improve
0.175258	-13.6512	parsing | model
0.411765	-13.6512	annotates | of
0.150376	-13.6512	a | policy
0.179775	-13.6512	spelling | in
0.0899183	-13.6513	above | and
0.666667	-13.6514	national | association
0.0793991	-13.6514	and | patterns
0.461538	-13.6514	of | covered.
0.2	-13.6514	labeled | data.
0.24	-13.6515	we | keyphrase
0.152	-13.6517	is | choose
0.177778	-13.6517	is | log
0.347826	-13.6518	that | knowing
0.103571	-13.6518	works | and
0.177778	-13.6519	our | patterns
0.461538	-13.652	class | several
0.285714	-13.6522	this | involves
0.0668693	-13.6523	this | other
0.10566	-13.6523	of | annotators
0.294118	-13.6524	200 | accuracy
0.202899	-13.6524	normalize | the
0.0219305	-13.6526	and | are | the
0.411765	-13.6526	hence | model
0.235294	-13.6526	a | techniques.
0.166667	-13.6527	various | that
0.0993377	-13.6527	to | resulting
0.216667	-13.6528	a | precision,
0.216667	-13.6528	callison-burch, | and
0.277778	-13.6528	to | strings.
0.388889	-13.653	split | training,
0.24	-13.6531	baseline | recall
0.162162	-13.6533	such | some
0.197368	-13.6533	used | search
0.24	-13.6534	(2008) | by
0.0800866	-13.6534	of | complex
0.333333	-13.6535	highly | than
0.666667	-13.6536	readers | articles
0.160714	-13.6536	baseline, | and
0.179775	-13.6536	that | necessary
0.060794	-13.6538	the | produced
0.555556	-13.6538	department | waltham,
0.112554	-13.6538	errors | by
0.10687	-13.6538	a | of | instances
0.388889	-13.6539	conventional | with
0.0313221	-13.6539	the | probability
0.100346	-13.6539	is | times
0.5	-13.6539	parameters | gradient
0.103203	-13.654	of | discussed
0.268293	-13.654	monolingual | corpora
0.321429	-13.654	corpus, | only
0.411765	-13.654	our | assumes
0.152	-13.6544	sentence | then
0.294118	-13.6544	consists | text
0.555556	-13.6544	he | different
0.555556	-13.6544	assumption | sentences.
0.5	-13.6544	what | often
0.347826	-13.6545	the | realizer
0.461538	-13.6545	into | approaches.
0.212121	-13.6545	as | components
0.102113	-13.6546	the | stack
0.0358997	-13.6546	best | the
0.165138	-13.6547	their | were
1	-13.655	segmentation | (creutz
1	-13.655	p(t|θ+) p(t|θ−)
1	-13.655	erc | 249520
1	-13.655	yes-answer->yes(yes) | db:1
1	-13.655	slumber | eyelids
1	-13.655	eventlda | accuracy
1	-13.655	50.20 | 55.72
1	-13.655	|0.506
1	-13.655	logp(q|d),
1	-13.655	affect-enriched | clasification
1	-13.655	v. | impr.
1	-13.655	mona | mdiab@ccls.columbia.edu
1	-13.655	callison-burch1 | 2human
1	-13.655	(fractions, | percentages).
1	-13.655	john | lowe.
1	-13.655	dahlmeier1 | integrative
1	-13.655	stationary | stationery
1	-13.655	‘kick | bucket’
1	-13.655	inquirer | 11,788
1	-13.655	accuracies | tradeoff
1	-13.655	nyu’s | ace
1	-13.655	ntcir-8 | meeting:
1	-13.655	nivre | martins
1	-13.655	5010 | rld</
1	-13.655	naming | transporting
1	-13.655	nih/nlm | lm07050.
1	-13.655	“tshyr” | “refers”
1	-13.655	mark | steedman@inf.ed.ac.uk
1	-13.655	torture | acapulco
1	-13.655	log|retweetsk|+ | k|+lavg
1	-13.655	jacy | erg
1	-13.655	canon | fujifilm
1	-13.655	p(wm,n |
1	-13.655	1− | n(n2−1)
1	-13.655	city* | night
1	-13.655	smruthi | rohini
1	-13.655	p | strikeout
1	-13.655	victoria | melbourne
1	-13.655	300 | 3000
1	-13.655	montr´eal, | h3c
1	-13.655	“easier” | easier”
1	-13.655	al-ghad | al-riyadh
1	-13.655	enlt | enet
1	-13.655	maja | (lt),
1	-13.655	xmsa | 10
1	-13.655	budget | feedback.
1	-13.655	<te | gh9
1	-13.655	inra | jouy-en-josas,
1	-13.655	expected-log-gaininterleaved-uncertainty-0.2 | interleaved-uncertainty-0.8expected-reconstruction-error
1	-13.655	absoluteness | relativeness:
1	-13.655	p(t|m).
1	-13.655	color | c1:n,
1	-13.655	dorr, | schwartz.
1	-13.655	‘and | morocco
1	-13.655	protesting | bad
1	-13.655	techniques | undoubtedly
1	-13.655	birth | venus
1	-13.655	commitments | preferences.
1	-13.655	carmen | rada
1	-13.655	series) | 〈hasprop〉(f
1	-13.655	a::d | of::i
1	-13.655	"le | 1-n"/
1	-13.655	[conceived | bsoa
1	-13.655	zones | restricted.
1	-13.655	shit | odee
1	-13.655	fusion | krahmer,
1	-13.655	ngf | nn
1	-13.655	dallas | tx,
1	-13.655	chowdhury | lavelli
1	-13.655	 | d2i
1	-13.655	los | herald
1	-13.655	s48 | s98
1	-13.655	alun | arktis
1	-13.655	r4 | swsd
1	-13.655	falmer, | bn1
1	-13.655	n-top | (n=10
1	-13.655	pcls | ncbc
1	-13.655	•s | •np
1	-13.655	2-00 | ma=
1	-13.655	stoyanov, | gilbert,
1	-13.655	anderswo | projekt
1	-13.655	john | ball
1	-13.655	dary | (b
1	-13.655	.33 | .86
1	-13.655	rm | </te
1	-13.655	haj | alðkyat
1	-13.655	defense | (ndseg)
1	-13.655	nucleus | satellite,
1	-13.655	gamallo | pichel,
1	-13.655	lda−500m−0.1−0.01 | lda−300m−0.1−0.01 | lda−200m−0.1−0.1
1	-13.655	),|pr(maxarg | tcee
1	-13.655	"061 | ="0.5
1	-13.655	nice, | positive,
1	-13.655	vbg | rbs
1	-13.655	“positive” | “negative”
1	-13.655	joanis | agg
1	-13.655	mmil | media
1	-13.655	#, | ±,
1	-13.655	featured | (haitian
1	-13.655	combinations | speel
1	-13.655	(lebanon | 2007;
1	-13.655	raytheon | 02138,
1	-13.655	pittsburgh | 15232
1	-13.655	consonants | consonants
1	-13.655	parenleftbig | l(vectori1)productdisplay
1	-13.655	requirements | task.
1	-13.655	“zum | example”
1	-13.655	dawid | skene
1	-13.655	h1,1 | h2,1
1	-13.655	inertia | baseline)
1	-13.655	bush | lawsuit
1	-13.655	orabona, | i-70125,
1	-13.655	loss | l(z,j,c)
1	-13.655	genes | proteins,
1	-13.655	access | history.
1	-13.655	lscriptmin: | lscriptmax:
1	-13.655	t | l; | t
1	-13.655	(muc | ceaf(e))/3
1	-13.655	20k | w/o
1	-13.655	braille | book
1	-13.655	synonyms, | hyponyms
1	-13.655	bonjour | salut
1	-13.655	searn | current
1	-13.655	human-generated | machine-generated
1	-13.655	p | <wf
1	-13.655	λ1, | λ3
1	-13.655	gestures | router’s
1	-13.655	pisa | informatica
1	-13.655	anderson, | zukunft
1	-13.655	disgust, | emotions
1	-13.655	amt | completion
1	-13.655	d3i | d4i
1	-13.655	marco | marco.baroni@unitn.it
1	-13.655	jin-dong | ohta,
1	-13.655	pr3 | pr1
1	-13.655	languages | tsl.
1	-13.655	n:animal | v:eats
1	-13.655	direction | en-fr
1	-13.655	“cat, | mammal”
1	-13.655	stellenwert | dabei
1	-13.655	judged | unidirectional
1	-13.655	learnthigh | effect-size
1	-13.655	u11 | u21
1	-13.655	supertreelet | t1.
1	-13.655	city* | good*
1	-13.655	jurafsky | linguistics
1	-13.655	“compatible | cm”
1	-13.655	direct-mapping | reversed-mapping
1	-13.655	costa-jussà | banchs
1	-13.655	coast | says
1	-13.655	0.000 | 0.010
1	-13.655	economy, | programme.
1	-13.655	t¨ass¨a | korkean
1	-13.655	size | 7.3
1	-13.655	null∗t | a∈ω,w
1	-13.655	λx2 | lic.
1	-13.655	brown | (francis
1	-13.655	nara | prefecture,
1	-13.655	summarize | viewpoints
1	-13.655	protestierten | gegen
1	-13.655	terrorism | abroad
1	-13.655	expanding | joint
1	-13.655	silke | whitt,
1	-13.655	essay | (attali
1	-13.655	fronting | fronting.
1	-13.655	log | interaction.
1	-13.655	alternative | referred
1	-13.655	(blaheta | 2007a).
1	-13.655	(top | (middle
1	-13.655	review | disclosed
1	-13.655	blogspot | ace
1	-13.655	solved | solver.
1	-13.655	ni] | -に
1	-13.655	guodong | qiaoming
1	-13.655	conll-2011 | “modeling
1	-13.655	0.793 | 0.823
1	-13.655	u: | home
1	-13.655	15 | accur
1	-13.655	(collection|series) | |
1	-13.655	marcello | evan
1	-13.655	38 | 36
1	-13.655	edge(nullnullnull) | node(null)
1	-13.655	comes | forms:
1	-13.655	(red | (blue
1	-13.655	illinois | urbana
1	-13.655	ku, | sun,
1	-13.655	breck, | claire
1	-13.655	word-choice | desired
1	-13.655	component-whole | entity-destination
1	-13.655	interplay | repository
1	-13.655	villeroy | boch.
1	-13.655	angeles, | 90045
1	-13.655	pr(c1|s,t)
1	-13.655	σ | s1s2
1	-13.655	(feder/fse) | 2010”
1	-13.655	giorno | mese
1	-13.655	93.09 | 93.06
1	-13.655	coarsening | label
1	-13.655	h… | ‘writing
1	-13.655	null4 | null6
1	-13.655	qsd | (i>j)
1	-13.655	easy–to–read | difficult–
1	-13.655	transducens | d’alacant,
1	-13.655	 | d1i
1	-13.655	pθ(f|e)
1	-13.655	thomson | cambridge,
1	-13.655	categorizated | ventanilla
1	-13.655	general-purpose | protocols.
1	-13.655	0.002 | 0.003
1	-13.655	“n | n”
1	-13.655	basis | (rbf)
1	-13.655	urdu | mukund
1	-13.655	stringnet | knowledgebase
1	-13.655	lrscore | distance.
1	-13.655	omar | dept.
1	-13.655	handsome | streets
1	-13.655	references andreevskaia | and
1	-13.655	thiago | salgueiro
1	-13.655	lemmatization | (minnen
1	-13.655	(present | absent)
1	-13.655	eugenio, | serafin,
1	-13.655	peter | christopher
1	-13.655	em-nb | em-nb
1	-13.655	68.1 | 77.8
1	-13.655	cnr | moruzzi,
1	-13.655	”starbucks” | ”coffee”
1	-13.655	image | hand-written
1	-13.655	surgery | 08/2007
1	-13.655	codeword | codeword
1	-13.655	wan, | xiao.
1	-13.655	adjacency | wij
1	-13.655	(collection∨ | @v(villeroy
1	-13.655	epitopes | bmpr-ia
1	-13.655	theme(coreentity|event),
1	-13.655	congressional | transcripts.
1	-13.655	ubiquitous | germany
1	-13.655	ldc2003e14, | ldc2005t06.
1	-13.655	47% | 24%
1	-13.655	star | celdt
1	-13.655	georgia | 30602
1	-13.655	pedagogy | psychoanalysis
1	-13.655	mt08-wb | dev10-wb
1	-13.655	elicit | illicit
1	-13.655	<tlink | signalid="s1"
1	-13.655	cooked | ate
1	-13.655	nba | jordan
1	-13.655	h(i+1) | d(ax,ay)
1	-13.655	step: | p(r|x)
1	-13.655	argmaxa∈ai | q∗ij(s,a),
1	-13.655	dotted | subtree.
1	-13.655	word-pairs | brackets
1	-13.655	crf-std | crf-lex
1	-13.655	pp-attachment | coordination
1	-13.655	4050 | 8090
1	-13.655	computation | c0,...,cm,
1	-13.655	drug | enzymes,
1	-13.655	0.001k | cutoff)
1	-13.655	0.815 | 0.860
1	-13.655	g1 | (v1,e1)
1	-13.655	q; | o(q);
1	-13.655	toolkit | employed.
1	-13.655	zheda | hangzhou
1	-13.655	vowels, | vowel
1	-13.655	 | 
1	-13.655	prefer | indoors
1	-13.655	rebecca | susan
1	-13.655	rao | ravichandran,
1	-13.655	convert | graph.
1	-13.655	nodef | p@70,
1	-13.655	harvesting | (hovy
1	-13.655	2.00% | 8.00%
1	-13.655	frontof(figure:washer, | facing(figure:washer,
1	-13.655	(3 | (6
1	-13.655	basic-sr | beam-sr
1	-13.655	lasdiff | —
1	-13.655	fa8750-09-c-0181, | award
1	-13.655	[0, | .]}.
1	-13.655	supertagger | relaxing
1	-13.655	libin | jinxi
1	-13.655	 coa+ng |  pork
1	-13.655	(williams | balakrishnan,
1	-13.655	jason | florham
1	-13.655	0.6 | documents-#labeled
1	-13.655	f−score | xicon
1	-13.655	blinked | lashes,
1	-13.655	(guthrie | hepple,
1	-13.655	left | pathological
1	-13.655	 | 
1	-13.655	0.5083 | 0.4667
1	-13.655	vincent | mercer.
1	-13.655	vnullnull1(v,y) | vnullnull2(v,y)
1	-13.655	berkeleyparser | bohnetparser
1	-13.655	boun | f-sco
1	-13.655	technologiestiftung | co-financed
1	-13.655	os= | >und
1	-13.655	cmc | (pestian
1	-13.655	farkas | szarvas
1	-13.655	sent-par | phrase-par
1	-13.655	kristina | redmond,
1	-13.655	e., | p.,
1	-13.655	65.9 | 58.5
1	-13.655	87.65 | 94.01
1	-13.655	acoustic | unvoiced
1	-13.655	turn-taking | making.
1	-13.655	fubl | ubl
1	-13.655	supposed | treelet
1	-13.655	segments | phoneme.
1	-13.655	21.1 | 72.1
1	-13.655	(hsu | glass,
1	-13.655	carreras, | michael
1	-13.655	hedge | numerical
1	-13.655	iwslt | 42.62
1	-13.655	%, | ¬.
1	-13.655	80k | (l)proposed
1	-13.655	mcp | 1.15,
1	-13.655	ct | kuniondisplay
1	-13.655	rosenfeld | feldman,
1	-13.655	instructor | planner
1	-13.655	2030 | 6070
1	-13.655	tempeval-2010 | association
1	-13.655	began | february
1	-13.655	def= | h1..t
1	-13.655	blogs | forums,
1	-13.655	chart | (ci),
1	-13.655	v; | v
1	-13.655	prep | postp
1	-13.655	john(x1):20 | compose(e1,x1,x2):0
1	-13.655	meth | fut
1	-13.655	s32s85s76s73s83s83s69 | (desr).
1	-13.655	— | 30%
1	-13.655	s51s48s48s48 | s56s54
1	-13.655	.94 | .81
1	-13.655	dn(f) | pn(e|f).
1	-13.655	die | bremsen
1	-13.655	nps, | quantified
1	-13.655	choi-nicolov’s | nivre’s
1	-13.655	visited | (s,a)
1	-13.655	lagrangian | lp
1	-13.655	(spitkovsky | 2010a,
1	-13.655	cd-pb | cd-dhpb
1	-13.655	(malioutov | (eisenstein
1	-13.655	action, | action,
1	-13.655	,exp | 〉,int
1	-13.655	wsem | wsem.
1	-13.655	2-grams | 6-grams
1	-13.655	chi | vietnam
1	-13.655	sprime | [b1
1	-13.655	vbd[was] | in:[by]
1	-13.655	volkswagen | lichtenberg-professorship
1	-13.655	xa | nadiyah
1	-13.655	rld" | rm>
1	-13.655	has | 10/2007.
1	-13.655	u.s. | hereon.
1	-13.655	l | sts
1	-13.655	she denies | chills,
1	-13.655	white | 11.7
1	-13.655	hr0011-06-c-0023 | (rgc)
1	-13.655	tr1r | ···u2r−1a
1	-13.655	amherst, | 01003
1	-13.655	enro | enel
1	-13.655	fondazione | kessler
1	-13.655	|x )=
1	-13.655	scheme="argument | consequences"
1	-13.655	software | http://www.csie.ntu.edu.tw/~cjlin/libsvm
1	-13.655	u18(x,y,di,α) | u26b(x,y,di,α)
1	-13.655	argument1 | argument2
1	-13.655	leon | thibaux
1	-13.655	p(k|t,r | 1)p(k|t,r
1	-13.655	toutanova | moore
1	-13.655	perusoikeuskirja | vahvistaa
1	-13.655	macquarie | (bernard,
1	-13.655	adjunct | adjunct
1	-13.655	(filatova | takamura
1	-13.655	day) | 〈creator〉(v
1	-13.655	p<.01), | efect
1	-13.655	else | λd(vectori1)+d(vectori2)
1	-13.655	1998; | inter
1	-13.655	adik-beradik | beradik
1	-13.655	construct | 2,162
1	-13.655	iterated | (kirby,
1	-13.655	meting | (sigdial),
1	-13.655	¨oo, | o¨onull
1	-13.655	“semantic | overlapping”
1	-13.655	(coster | kauchak,
1	-13.655	(first | dots)
1	-13.655	excellent, | fortunate,
1	-13.655	wei | kam-fai
1	-13.655	removed=terms | category=terms
1	-13.655	strapparava | mihalcea
1	-13.655	phrases | granularities
1	-13.655	85.3 | 87.1
1	-13.655	bkg | res
1	-13.655	p(d|c).
1	-13.655	41.9 | 47.4
1	-13.655	haffari | sarkar
1	-13.655	(girju | pennacchiotti,
1	-13.655	p(btarrowhookleft→a | 1|at
1	-13.655	(natarajan | saleem
1	-13.655	路 | 下
1	-13.655	2to | exposition,the
1	-13.655	universityof | park,maryland
1	-13.655	<arg0, | <arg1,
1	-13.655	p(e|f) | p(f|e)
1	-13.655	heng | 11367,
1	-13.655	tagyoung | daniel
1	-13.655	[1, | .,
1	-13.655	57.1 | touchdownpartialcount
1	-13.655	percent, | 95
1	-13.655	fevers, | or
1	-13.655	hard | algorithms,
1	-13.655	msrpc | cqae
1	-13.655	14.44 | 3.47
1	-13.655	airola | al.
1	-13.655	trust, | anticipation.
1	-13.655	former, | latter,
1	-13.655	nr | nff
1	-13.655	confirmed | ->artist,
1	-13.655	automatic | discovery.
1	-13.655	commission | programme)
1	-13.655	e1: | e2:
1	-13.655	bridewel, | hanbury,
1	-13.655	(wsj | 0).
1	-13.655	out-of-vocabulary | (oovs).
1	-13.655	agreeing | 325
1	-13.655	hunter | colorado
1	-13.655	desert | dessert
1	-13.655	s114s32 | s115
1	-13.655	argued | 1980).
1	-13.655	okumura, | riedhammer
1	-13.655	socer | twenty
1	-13.655	tend | noisy.
1	-13.655	(tanaka | umemura,
1	-13.655	senti|aligni−1,j−k)
1	-13.655	testing | candidatesentences
1	-13.655	san | 95120
1	-13.655	> | tid=''t1''
1	-13.655	pre | postconditions
1	-13.655	5800 | n95
1	-13.655	lnlr | β
1	-13.655	federal | theseus
1	-13.655	split | (k,m,lscript)
1	-13.655	sec | 00
1	-13.655	doc-avg | doc-comb,
1	-13.655	(billot | 1989;
1	-13.655	lll | (n´edellec,
1	-13.655	lda−1000m−0.1−0.1 | lda−100m−0.1−0.01
1	-13.655	demographics | highly-discriminant
1	-13.655	2017 | "063
1	-13.655	lisa’s | number.”
1	-13.655	texas* | houston*
1	-13.655	ja | komission
1	-13.655	mec/micinn | itrans2
1	-13.655	mvm−3m−0.1−0.01 | lda−100m−0.1−0.1
1	-13.655	respect | control
1	-13.655	month, | formed
1	-13.655	gmail | track
1	-13.655	(eskin, | kvˇetoˇn
1	-13.655	s006 | tutor
1	-13.655	dry | shortness
1	-13.655	textual | (rte2)
1	-13.655	“crf | crf-mf”
1	-13.655	ntcir | moat,
1	-13.655	[pas07]: | paşca.
1	-13.655	homophones, | synonyms.
1	-13.655	mvm−50m−0.1−0.01 | mvm−20m−0.1−0.01
1	-13.655	certificates | optimality,
1	-13.655	noun, | determiner,
1	-13.655	belz | kow
1	-13.655	mpnc | mmnr−1
1	-13.655	cough | breath
1	-13.655	nicola | bojar,
1	-13.655	mediastinal | right
1	-13.655	(fda) | maximize
1	-13.655	swim | shortness
1	-13.655	identity | driver’s
1	-13.655	patwardhan | riloff.
1	-13.655	denis | anaphoricity
1	-13.655	(o | ru)"
1	-13.655	porte-parole | reconnu
1	-13.655	svr-rec | svr-base,
1	-13.655	(definitionindictionary | (orth
1	-13.655	(dinesh | al.,
1	-13.655	206; | roisman,
1	-13.655	universitat | catalunya
1	-13.655	electrode | first
1	-13.655	“x | 맛있다
1	-13.655	constraint | sw/.
1	-13.655	prevent | decay.
1	-13.655	"und | wf>
1	-13.655	name: | industry:
1	-13.655	onlineb | upv-combo
1	-13.655	[[u2]]c | [[u3]]c
1	-13.655	sandia | laboratories
1	-13.655	(knowledge | robots)
1	-13.655	xiong, | shouxun
1	-13.655	joakim | joakim.nivre@lingfil.uu.se
1	-13.655	wonder | acclaimed
1	-13.655	(merlo | plas,
1	-13.655	s) | g),
1	-13.655	srilm-h | 26.6g
1	-13.655	+prior | +cov
1	-13.655	bond, | flickinger.
1	-13.655	inc. | finmeccanica
1	-13.655	ad-hoc | ir
1	-13.655	(goldberg | tsarfaty,
1	-13.655	intramural | nih,
1	-13.655	pb(w3|w1w2) | pb(w3|w2)
1	-13.655	al., 2009; | et.
1	-13.655	u.s. | iis-0713402
1	-13.655	topic | (rosen-zvi
1	-13.655	40k | (l)prop.
1	-13.655	historians | living
1	-13.655	(8 | memory).
1	-13.655	recursive | mercer,
1	-13.655	88.55 | 92.39
1	-13.655	adhoc | sweet
1	-13.655	crf | noprop-dep
1	-13.655	navigation | desktop
1	-13.655	61.1 | 62.8
1	-13.655	pr (w|g)
1	-13.655	(1a) | (1b)
1	-13.655	weka | update.
1	-13.655	technische | darmstadt,
1	-13.655	proper | numbers.
1	-13.655	t111 | tpi(r)2r
1	-13.655	zornitsa | 90292-6695
1	-13.655	pre-translation | permutation
1	-13.655	human | ugc.
1	-13.655	non-local | sampling.
1	-13.655	server | engine.
1	-13.655	“divide | conquer”
1	-13.655	“largest” | (nullnullnull@null).
1	-13.655	(blitzer | 2006)
1	-13.655	(3a) | (3b),
1	-13.655	field::n | width::n
1	-13.655	hall, | jens;
1	-13.655	goddesses | parametrizable
1	-13.655	sα | sβ
1	-13.655	bleu-c | 0.1344
1	-13.655	tistica | dataexi
1	-13.655	habash-soudibuckwalter | (habash
1	-13.655	coper, | buchanan.
1	-13.655	children | autism.
1	-13.655	research | almaden
1	-13.655	s1: | guard
1	-13.655	paraphrases | l1-english
1	-13.655	plsa | d5
1	-13.655	directory | (odp)
1	-13.655	+np: | +vbd:
1	-13.655	software | nanjing
1	-13.655	lda−500m−0.1−0.1 | lda−50m−0.1−0.01
1	-13.655	fieldgoalpartialcount | 40.9
1	-13.655	rd., | 310027,
1	-13.655	63.22 | 3.31
1	-13.655	flowers | stay
1	-13.655	she was | afinitor
1	-13.655	schid="0" | </prop>
1	-13.655	tason | kuninkaan
1	-13.655	(reisinger | mooney,
1	-13.655	verified | rtk
1	-13.655	wrong | 1,310
1	-13.655	% | mvm−30m−0.1−0.01
1	-13.655	antarctica | shelf.
1	-13.655	sub(zuv,x) | del(zuv)
1	-13.655	texas* | 1865-1901
1	-13.655	mvm−5m−0.1−0.01 | lda−1000m−0.1−0.01
1	-13.655	nmflib | nmfcon
1	-13.655	m4 | m7
1	-13.655	università | dipartimento
1	-13.655	interfaces | (maxent,
1	-13.655	2% | 11%
1	-13.655	zprime p(o|zprime)p(zprime|c)
1	-13.655	ayes | ano
1	-13.655	promoterdependence | transcriptionfrom
1	-13.655	l1 | l2,
1	-13.655	(caines | buttery,
1	-13.655	(c3) | (c4)
1	-13.655	grosz, | k.
1	-13.655	feldman | hana,
1	-13.655	(craig, | mcdaniel
1	-13.655	chilpancingo | hour
1	-13.655	logical | metonymies
1	-13.655	s56s52 | s32s100s80s85s80s65
1	-13.655	nos. | council
1	-13.655	87.36 | 93.99
1	-13.655	(collection|series)).
1	-13.655	clac-nb: | and
1	-13.655	p(r7 | r3null
1	-13.655	〈hasprop〉(f | v&b))))
1	-13.655	(vyas | 2009).
1	-13.655	gcc | compiler.
1	-13.655	“light | “support
1	-13.655	mt08-nw | dev10-nw
1	-13.655	φ(x,y) | zero
1	-13.655	prd2t(t|w) | prt2d(t|w)
1	-13.655	michael | m.auli@sms.ed.ac.uk
1	-13.655	(magnini | cavaglia,
1	-13.655	ecfunded | cosyne
1	-13.655	uniprot | subtiwiki
1	-13.655	(10-fold | validation)
1	-13.655	lda−50m−0.1−0.1 | dpmm−0.1−0.1
1	-13.655	dec | dec
1	-13.655	(d´ıaz | dub´e
1	-13.655	subrequest | [b0,c0]
1	-13.655	ruz, | maute,
1	-13.655	storage | bytes/n-gram
1	-13.655	re-rank | finite-length
1	-13.655	actiontarget | transcriptionby
1	-13.655	long-term | 1850
1	-13.655	families | tongues
1	-13.655	self | protocol
1	-13.655	vlachos | biostatistics
1	-13.655	c., | h.,
1	-13.655	cpu | 4.0.1
1	-13.655	下 | 下雨
1	-13.655	term-level | entailment
1	-13.655	administration | boeing
1	-13.655	10-1020-1530-2040-2550-30400-50500-60600-70700-80800-90 | #labeled
1	-13.655	field-s | [[n
1	-13.655	hki | tkj
1	-13.655	jerónimo | (c5)
1	-13.655	scfg | left-hand-side
1	-13.655	argm-dis | argm-mnr
1	-13.655	charles | newark,
1	-13.655	patent | 1.20%
1	-13.655	select:1 | winword.exe
1	-13.655	(bartlet | 207;
1	-13.655	spanish: | lse:
1	-13.655	ingenio | (csd2007-00018),
1	-13.655	g2 | keywords
1	-13.655	# | ensl
1	-13.655	λ2 | vectori1,vectori2,l(vectori1)=l(vectori2)
1	-13.655	p(a|data,α).
1	-13.655	topic | zd,n
1	-13.655	exclusive | relations.
1	-13.655	accompanied | kind.
1	-13.655	···u1r−1a | tpi(1)21
1	-13.655	p(e|q).
1	-13.655	|v|,
1	-13.655	brooke | richard
1	-13.655	bomb | victims
1	-13.655	sciences | france
1	-13.655	klein | (dmv),
1	-13.655	hungarian, | slovene,
1	-13.655	1− 2nd+|y|
1	-13.655	nlu | others.
1	-13.655	bietet | wissenschaft
1	-13.655	illustration | generalized
1	-13.655	effect | disease
0.5	-13.655	(primeprime) | top
0.5	-13.655	hussein | terrorist
0.5	-13.655	1e+07 | asonucle
0.5	-13.655	gender | n=3
0.5	-13.655	bilm | avg.
0.5	-13.655	s32 | s32s76s83
0.285714	-13.6551	as | nodes.
0.183908	-13.6551	a | modifier
0.0647887	-13.6551	and | three
0.347826	-13.6553	these | point
0.255814	-13.6553	may | less
0.186047	-13.6554	for | lack
0.321429	-13.6554	using | stream
0.268293	-13.6555	(a | with
0.555556	-13.6556	able | see
0.075	-13.6558	available | and
0.165138	-13.6558	builds | the
0.0932945	-13.6558	setup | the
0.126374	-13.6558	are | necessarily
0.0896739	-13.6559	of | wordnet
0.75	-13.6559	volume | volume
0.75	-13.6559	worth* | fort*
0.75	-13.6559	(dt) | distributional
0.75	-13.6559	vocabulary | stereotypes
0.75	-13.6559	large-scale | data-driven
0.75	-13.6559	it’s | it’s
0.75	-13.6559	french | (bonneau-maynard
0.75	-13.6559	illustrates | process.
0.75	-13.6559	machine | 2000a;
0.75	-13.6559	scopes | biomedical
0.75	-13.6559	twenty | twenty
0.75	-13.6559	moses | translating
0.75	-13.6559	map | integers,
0.75	-13.6559	mountain | 94043,
0.75	-13.6559	grct | lct
0.75	-13.6559	sassano | 2000;
0.75	-13.6559	j∩ci|
0.75	-13.6559	person | voice
0.75	-13.6559	63 | 14
0.75	-13.6559	query | filling
0.75	-13.6559	|v|.
0.75	-13.6559	(2009b) | details.
0.75	-13.6559	instances | ke
0.75	-13.6559	labels | location,
0.75	-13.6559	restricted | head-driven
0.75	-13.6559	statements | statement
0.75	-13.6559	cr | ranks
0.75	-13.6559	world | proofs.
0.75	-13.6559	took | 20,
0.75	-13.6559	tree-rewriting | contain
0.75	-13.6559	calculate | agreement.
0.75	-13.6559	minutes, | years
0.75	-13.6559	km04 | saj10a
0.75	-13.6559	66.1 | 62.9
0.175258	-13.6559	templates | in
0.116279	-13.6561	does | it
0.461538	-13.6561	up | form
0.363636	-13.6562	methodology | different
0.103704	-13.6563	prior | in
0.171717	-13.6564	the | dimensionality
0.0632571	-13.6564	10 | of
0.285714	-13.6566	used | parameter
0.175258	-13.6566	therefore, we | a
0.181818	-13.6566	(6) | to
0.75	-13.6567	d. | f.
0.75	-13.6567	learning | (ramshaw
0.75	-13.6567	d. | ng.
0.75	-13.6567	findings | 2007;
0.75	-13.6567	thread | conversation
0.75	-13.6567	amjad | eecs
0.75	-13.6567	equations | (15),
0.75	-13.6567	standard deviations | replicated
0.75	-13.6567	aspect | aspect.
0.75	-13.6567	hpsg | enju
0.75	-13.6567	60.1 | contrastive
0.75	-13.6567	meteor | wer
0.75	-13.6567	nodes, | node.
0.75	-13.6567	pruning | efficiency
0.75	-13.6567	〈genowner〉 | via
0.75	-13.6567	114 | 16
0.75	-13.6567	shortest | connecting
0.75	-13.6567	reweighting | graph
0.75	-13.6567	1992; | 1995;
0.75	-13.6567	hmm | prlg.
0.75	-13.6567	dem | sich
0.75	-13.6567	government | ict
0.75	-13.6567	won | 2010
0.75	-13.6567	toolkit | enju parser
0.75	-13.6567	total | 0.0001
0.136364	-13.6567	thank | anonymous
0.347826	-13.6568	sentences | logical
0.310345	-13.6568	how | particular
0.101045	-13.6568	this | evaluation
0.136364	-13.6568	tagging | a
0.555556	-13.6569	class | functions
0.555556	-13.6569	1999; | 2001;
0.150794	-13.6569	more | about
0.461538	-13.6571	recognized | was
0.255814	-13.6571	or | polarity
0.150794	-13.6571	set | new
0.347826	-13.6572	remove | which
0.261905	-13.6572	to | reader
0.175258	-13.6574	and | adjective
0.428571	-13.6574	relation | under
0.75	-13.6575	anchor | title
0.75	-13.6575	tracking | relaxation
0.75	-13.6575	ecdf | partially
0.75	-13.6575	weighted | maximized.
0.75	-13.6575	example | fifty
0.75	-13.6575	[p | rolls
0.75	-13.6575	pa | patient
0.75	-13.6575	hierarchies | learning.
0.75	-13.6575	incorporate | hashing.
0.75	-13.6575	correctly | lex-items
0.75	-13.6575	recorded | hit
0.75	-13.6575	reach | reach
0.75	-13.6575	effort | u.s.
0.555556	-13.6575	word | 1992).
0.118483	-13.6577	retrieval | and
0.5	-13.6577	θ | ·
0.25	-13.6578	by | (i.e.
0.216667	-13.6579	kim | al.,
0.428571	-13.6579	previous | showed that
0.255814	-13.6579	these | relatively
0.055102	-13.658	during | the
0.0279757	-13.658	all | and
0.461538	-13.6581	using | (liu
0.294118	-13.6583	the | meets
0.75	-13.6584	2.1 | percentage
0.75	-13.6584	psychosentiwordnet | about
0.75	-13.6584	define | ranges,
0.75	-13.6584	person, | organization.
0.75	-13.6584	- | couples
0.75	-13.6584	problem | exacerbated
0.75	-13.6584	further | needed.
0.75	-13.6584	(lavergne | 2010).
0.75	-13.6584	rule | unrestricted
0.75	-13.6584	requires | knowledge.
0.75	-13.6584	ams | vocabulary
0.130178	-13.6585	2.2 | in
0.388889	-13.6586	by | down
0.347826	-13.6587	linguistic | knowledge
0.666667	-13.6588	reaction | two
0.310345	-13.6588	this | constraints.
0.268293	-13.6588	simple | has
0.216667	-13.6588	use | clustering
0.0818182	-13.6589	such | which
0.126374	-13.6589	0 | 4 | 5
0.461538	-13.6591	algorithms | shown
0.75	-13.6592	represent | concept,
0.75	-13.6592	adverbs, | verbs
0.75	-13.6592	assessment | proficiency
0.75	-13.6592	different | (?;
0.75	-13.6592	time* | be
0.75	-13.6592	2001; | dolan
0.75	-13.6592	only | eaten.
0.75	-13.6592	bigram | unigrams.
0.75	-13.6592	original | version,
0.75	-13.6592	absence | explicit
0.75	-13.6592	suffix | morphemes
0.75	-13.6592	serena | classifier
0.75	-13.6592	confluence | random
0.0914286	-13.6592	must | a
0.171717	-13.6592	that | other.
0.154472	-13.6592	that | techniques
0.428571	-13.6593	with | to,
0.0527804	-13.6593	the | patterns
0.0874036	-13.6594	for | like
0.2	-13.6594	terms | semantic
0.5	-13.6594	test | effectiveness
0.5	-13.6594	mapping | :
0.666667	-13.6595	acknowledges | (afrl)
0.119048	-13.6595	a | documents.
0.17	-13.6598	reordering model | the
0.1875	-13.6598	the | co-reference
0.310345	-13.6599	where | goal
0.411765	-13.6599	2.1 | types
0.294118	-13.66	units | or
0.75	-13.66	provides | robustness
0.75	-13.66	unlike | evaluation,
0.75	-13.66	hope | encourage
0.75	-13.66	starting | sj
0.75	-13.66	role | islamic
0.555556	-13.66	in | userusage
0.109756	-13.66	standard | with
0.0582192	-13.66	systems | to
0.1875	-13.6601	4.4 | a
0.155738	-13.6601	using | evaluation
0.216667	-13.6601	that | reduced
0.666667	-13.6602	extraction | subjective
0.461538	-13.6602	approach | reranking
0.062417	-13.6602	the | token
0.175824	-13.6603	has | positive
0.0188906	-13.6605	our | of
0.0996564	-13.6605	do | with
0.5	-13.6605	models | (relms
0.220339	-13.6606	is | beginning
0.044137	-13.6607	show | of
0.145985	-13.6607	third | a
0.120603	-13.6608	topic | are
0.388889	-13.6608	tables | and 3
0.285714	-13.6608	score | obtained
0.75	-13.6609	funded | india.
0.75	-13.6609	additional | graphic.
0.75	-13.6609	liu | texas
0.75	-13.6609	convergence | fig.
0.148148	-13.6609	example: | the
0.666667	-13.661	level | hidden
0.0911681	-13.6611	based on | we
0.0804348	-13.6611	is | positive
0.14094	-13.6611	to | valid
0.104869	-13.6611	we | above
0.5	-13.6611	a | subcorpus
0.5	-13.6611	only-if | 0
0.17	-13.6613	we | core
0.17	-13.6613	8 | 2
0.411765	-13.6613	traditional | extraction
0.095679	-13.6614	can | information
0.150794	-13.6614	remainder | the
0.175824	-13.6615	2. | word
0.212121	-13.6616	a | off
0.75	-13.6617	metric | adequacy.
0.75	-13.6617	integrating | syntax
0.75	-13.6617	modeling | relevance
0.75	-13.6617	january | 2010
0.75	-13.6617	man* | related
0.75	-13.6617	id | lemma
0.75	-13.6617	reflect | real
0.22807	-13.6617	agreement | not
0.363636	-13.6619	intend | our
0.157895	-13.6619	entity | an
0.24	-13.662	information | form
0.333333	-13.6621	collected | two
0.0284231	-13.6622	is | it
0.255814	-13.6622	: | i
0.108871	-13.6623	those | or
0.25	-13.6623	or | what
0.0478026	-13.6624	3 | in
0.202899	-13.6624	fuzzy | to
0.205882	-13.6625	the | mechanisms
0.666667	-13.6625	rules | 2005)
0.216667	-13.6625	use | models.
0.0709343	-13.6625	automatically | of
0.75	-13.6626	application | computer
0.75	-13.6626	developed | logic
0.75	-13.6626	auli | edinburgh
0.75	-13.6626	wikulu | web
0.128492	-13.6627	unlike | to
0.428571	-13.6627	wordnet | synset
0.411765	-13.6627	detect | their
0.142857	-13.6628	there | such as
0.5	-13.6628	0.03 | convergence
0.321429	-13.6629	rule | :
0.363636	-13.6631	also, | can
0.666667	-13.6632	distinction | non-english
0.063172	-13.6632	the | of | tokens
0.461538	-13.6633	phenomena | linguistic
0.189873	-13.6633	case | two
0.75	-13.6634	generate | ciphertext
0.75	-13.6634	6 conclusion | parsing.
0.75	-13.6634	rcv1 | points
0.75	-13.6634	가 | 보다
0.75	-13.6634	comparing | dual
0.135484	-13.6635	information | their
0.125	-13.6636	approaches to | of
0.555556	-13.6637	additional | 2010).
0.130952	-13.6637	for | manually
0.347826	-13.6638	system, | over
0.1875	-13.6638	collection | by
0.197368	-13.6638	a | recall.
0.0638889	-13.6638	but | are
0.0609137	-13.6639	as | but
0.189873	-13.6639	good | not
0.363636	-13.6639	having | different
0.5	-13.6639	ranking | weights
0.5	-13.6639	induction | disambiguation
0.321429	-13.664	annotation, | that
0.428571	-13.6641	generic | semantic
0.411765	-13.6641	detailed | analysis
0.75	-13.6642	generated | image.
0.75	-13.6642	opinions, | news
0.75	-13.6642	belief | adaptation
0.555556	-13.6643	state | emission
0.153226	-13.6645	have | so
0.666667	-13.6647	relatedness | senses
0.25	-13.6648	and | (2006) | for
0.085213	-13.6648	that | rules
0.255814	-13.6648	with | annotation.
0.411765	-13.665	have | great
0.0751445	-13.665	in | found
0.75	-13.6651	system | equations.
0.75	-13.6651	noise | after
0.75	-13.6651	01iw08003 | german
0.75	-13.6651	string-to-dependency | language
0.120603	-13.6652	learning | which
0.294118	-13.6652	into | 2009).
0.171717	-13.6652	a | translation,
0.235294	-13.6652	used | compare
0.216667	-13.6654	you | it
0.333333	-13.6654	automatically | new
0.12	-13.6655	use | but
0.666667	-13.6655	4.3 | testing
0.0909091	-13.6655	workshop | pages
0.294118	-13.6656	using | parser.
0.555556	-13.6656	time, | time.
0.363636	-13.6656	nugget | the
0.194805	-13.6657	are | ranked
0.148438	-13.6657	5.4 | the
0.189873	-13.6658	are | too
0.134146	-13.6658	and | roles
0.113537	-13.6658	mostly | the
0.285714	-13.6659	induction | semantic
0.75	-13.6659	combination | integration
0.75	-13.6659	very | 0.8),
0.75	-13.6659	block | procedure
0.75	-13.6659	designed | detailed
0.75	-13.6659	decoding | 1-best
0.75	-13.6659	occurring | once
0.75	-13.6659	across | categories,
0.75	-13.6659	bilingual | 2001;
0.75	-13.6659	approximately | points.
0.160714	-13.6659	2 | task
0.181818	-13.6659	first | text
0.3	-13.6661	word | e.g.
0.0504348	-13.6661	a | some
0.321429	-13.6662	source | additional
0.461538	-13.6664	should | always
0.2	-13.6664	primeprime | primeprime
0.388889	-13.6665	we | g2p
0.388889	-13.6665	words | phrase.
0.388889	-13.6665	correct | cases
0.0447458	-13.6667	of | data.
0.163636	-13.6667	text | information
0.0841346	-13.6667	(2) | a
0.5	-13.6667	techniques | algorithms
0.75	-13.6668	detection | urdu
0.75	-13.6668	sum | parts.
0.75	-13.6668	in | suc.
0.75	-13.6668	glow | local
0.75	-13.6668	• | mismatch:
0.363636	-13.6668	evaluation, | set
0.0816327	-13.6668	on | sentence
0.277778	-13.6668	appear | more
0.321429	-13.6669	obtain | semantic
0.189873	-13.6669	containing | from
0.0764588	-13.6669	we | distribution
0.0446852	-13.6669	class | the
0.428571	-13.667	# | 13
0.0930931	-13.6671	of | 1,
0.230769	-13.6671	if | above
0.363636	-13.6672	represents | state
0.194805	-13.6673	pairwise | in
0.0939394	-13.6673	of | system,
0.244898	-13.6673	to | usefulness
0.12	-13.6674	game | the
0.0616175	-13.6674	to | specific
0.0652174	-13.6674	the | items
0.194805	-13.6675	involve | a
0.3	-13.6675	the | ciphertext
0.205882	-13.6676	setting | with
0.75	-13.6676	annotator | answer
0.75	-13.6676	estimated | iterative
0.75	-13.6676	2007), | authors
0.75	-13.6676	4 | [5..6]
0.75	-13.6676	pam | topic
0.75	-13.6676	addition | multiplication.
0.75	-13.6676	term | filtering
0.75	-13.6676	latent | advantage
0.0757282	-13.6676	a | example
0.666667	-13.6677	difficult | require
0.666667	-13.6677	strategy | real
0.333333	-13.6677	conducted | each
0.150794	-13.6677	necessary | to
0.224138	-13.6679	language | corpora
0.145985	-13.6679	with | role
0.555556	-13.6681	systems | world
0.461538	-13.6684	computing | scores
0.213115	-13.6684	the | facilitates
0.75	-13.6685	lexicons | induce
0.75	-13.6685	german, | french
0.75	-13.6685	stsg, | time.
0.75	-13.6685	order | calculated.
0.75	-13.6685	modules | pipeline
0.75	-13.6685	shown | tab.
0.75	-13.6685	effort | labeling
0.75	-13.6685	authors | gender
0.75	-13.6685	& | b.
0.255814	-13.6685	for | locations
0.160714	-13.6685	assigned | as
0.2	-13.6686	straightforward | a
0.321429	-13.6687	information | questions
0.555556	-13.6687	out | every
0.285714	-13.6688	this paper | proposed
0.261905	-13.6689	set | documents.
0.127072	-13.6689	more | their
0.0983607	-13.6689	class | for
0.388889	-13.6691	range | linguistic
0.388889	-13.6691	simply | these
0.255814	-13.6691	apply | other
0.17	-13.6692	on | they are
0.411765	-13.6692	we also | various
0.310345	-13.6693	the | paragraphs.
0.310345	-13.6693	from | in-domain
0.75	-13.6693	movie | (mr)
0.75	-13.6693	modeling | orders
0.75	-13.6693	made | 2011,
0.0996564	-13.6693	main | that
0.285714	-13.6694	(2007) | an
0.461538	-13.6695	component | whether
0.461538	-13.6695	precision, | it
0.163636	-13.6695	we | external
0.137255	-13.6696	of | generally
0.144928	-13.6696	a | ontology
0.0727273	-13.6696	models, | the
0.130178	-13.6697	is | potential
0.171717	-13.6698	annotation | each
0.244898	-13.6698	with | unigram
0.268293	-13.6699	2001), | of
0.023355	-13.6699	when | the
0.277778	-13.6699	lists | or
0.388889	-13.67	5 | selected
0.0818182	-13.6701	the | spanish
0.162162	-13.6701	that | degree
0.75	-13.6701	pittsburgh | pittsburgh,
0.75	-13.6701	provides | visualizing
0.75	-13.6701	sunnyvale, | 94089
0.75	-13.6701	string | equal
0.75	-13.6701	during | iteration,
0.75	-13.6701	expression | describes
0.411765	-13.6701	given | make
0.134615	-13.6702	algorithm | training
0.163636	-13.6704	result | an
0.461538	-13.6705	approach | fully
0.235294	-13.6705	comparison, | to
0.123077	-13.6705	calculate | and
0.411765	-13.6706	certain | while
0.666667	-13.6707	t | ct
0.261905	-13.6707	pronunciation | and
0.110169	-13.6707	of | national
0.194805	-13.671	necessarily | a
0.1875	-13.671	bigram | is
0.75	-13.671	compound | vocabulary
0.75	-13.671	restaurant | (crp)
0.75	-13.671	ratings | comments
0.75	-13.671	contents | provide
0.75	-13.671	text | preprocessed
0.75	-13.671	colours | shown
0.75	-13.671	& | 204;
0.115207	-13.671	second | by
0.213115	-13.671	shift | to
0.461538	-13.6711	of | headlines.
0.179775	-13.6711	given | phrase
0.138158	-13.6712	we | change
0.173913	-13.6712	have | following
0.5	-13.6712	able | discriminate
0.261905	-13.6713	of | second-order
0.285714	-13.6714	johnson | a
0.333333	-13.6714	problem | problems
0.333333	-13.6714	second | contains
0.666667	-13.6715	makes | inference
0.3	-13.6715	which | longer
0.216667	-13.6715	should | their
0.163636	-13.6715	for | pair,
0.277778	-13.6715	for | events,
0.75	-13.6718	china | funds
0.75	-13.6718	meaning | source.
0.75	-13.6718	techniques, | coverage
0.75	-13.6718	’@ | ˇ@
0.75	-13.6718	modifier | relative
0.230769	-13.6718	pair | alignment
0.27027	-13.6719	associates | the
0.123656	-13.672	includes | as
0.411765	-13.672	name | person
0.130952	-13.672	step | that
0.102941	-13.6721	the | versus
0.363636	-13.6721	those | correct
0.363636	-13.6721	text | techniques
0.388889	-13.6722	of | neighbours.
0.666667	-13.6722	ball | each
0.294118	-13.6722	revealed | in
0.220339	-13.6722	the | root node
0.428571	-13.6723	examples | good
0.5	-13.6723	sentences | expressed
0.156522	-13.6724	on | per
0.411765	-13.6724	document | author
0.3	-13.6725	feature | n-gram
0.555556	-13.6725	classified | having
0.183908	-13.6725	advanced | and
0.181818	-13.6725	that | quality.
0.166667	-13.6726	with | senses
0.153226	-13.6726	doing | the
0.75	-13.6727	symbols | stack
0.75	-13.6727	detecting | usage
0.75	-13.6727	older | tend
0.75	-13.6727	governor | within
0.277778	-13.6727	comparison | method
0.152	-13.6728	we | representing
0.294118	-13.6728	depth | similarity
0.165049	-13.6729	methods | information
0.156522	-13.673	is | 2003).
0.388889	-13.6731	is | sample.
0.555556	-13.6731	extended | features.
0.168317	-13.6733	nlp | such
0.2	-13.6734	works | on
0.388889	-13.6735	distribution | uniform
0.0687398	-13.6736	a | alignment
0.154472	-13.6736	for example, | more
0.555556	-13.6737	dependency | mod
0.461538	-13.6737	adjective | vectors
0.461538	-13.6737	find | sequence
0.111111	-13.6737	comparison | to
0.0720721	-13.6738	when | on
0.144928	-13.6738	dialog | the
0.347826	-13.674	classifier | labels
0.333333	-13.6741	you | like
0.555556	-13.6743	competence | a
0.75	-13.6744	robustness | similarity
0.75	-13.6744	specialized | directly
0.75	-13.6744	open | (vilar
0.75	-13.6744	again, | outperforms
0.75	-13.6744	(jordan | al.,
0.75	-13.6744	accuracy | lowering
0.75	-13.6744	stimuli | category
0.75	-13.6744	assignment | product
0.75	-13.6744	arabic | datasets.
0.368421	-13.6744	step | update
0.261905	-13.6745	cases | more
0.3	-13.6746	which | preceding
0.0719424	-13.6747	parsing | to
0.411765	-13.6747	0.5 | 2.0
0.159292	-13.6747	verb | not
0.388889	-13.6748	highly | semantic
0.285714	-13.6749	apply | method
0.255814	-13.6749	next | will
0.0835322	-13.675	as | some
0.0609137	-13.6751	v | the
0.3	-13.6752	(1 | =
0.75	-13.6752	first | (2006).
0.75	-13.6752	one | character-based
0.75	-13.6752	start | above
0.461538	-13.6752	systems | utterance
0.5	-13.6752	models | (vogel
0.192308	-13.6753	of | normalization
0.388889	-13.6753	meaning | representation
0.666667	-13.6753	genia | abstracts
0.255814	-13.6754	the | city,
0.230769	-13.6754	and | pp.
0.368421	-13.6756	θ | [φ
0.121212	-13.6757	social | a
0.461538	-13.6758	chart-based | the
0.13253	-13.6758	the | judges
0.285714	-13.6759	how | those
0.208955	-13.6759	2 | extraction
0.666667	-13.676	quantitative | linguistic
0.147287	-13.676	3 | corpus
0.126374	-13.676	are able | the
0.388889	-13.6761	provide | context
0.388889	-13.6761	however, | explicitly
0.75	-13.6761	root | (where
0.75	-13.6761	over | 50%.
0.75	-13.6761	scheme, | structure
0.75	-13.6761	rows | table,
0.75	-13.6761	search | (kumar
0.75	-13.6761	statistic | observed
0.75	-13.6761	web | isolated
0.216667	-13.6761	ben | of
0.197183	-13.6761	to | mwes
0.143885	-13.6761	cases | not
0.0609911	-13.6762	also | by
0.137255	-13.6763	the | phonetic
0.162162	-13.6764	roles | in
0.13253	-13.6764	accuracy | using
0.27027	-13.6765	for | tweets.
0.216667	-13.6766	thus | =
0.17	-13.6767	right-hand | the
0.333333	-13.6767	of | acld
0.461538	-13.6768	project | university
0.224138	-13.6768	∼ | for
0.224138	-13.6768	can | among
0.0334487	-13.6768	three | the
0.149606	-13.6768	system | their
0.0903955	-13.6769	the | z
0.75	-13.6769	before | performed.
0.75	-13.6769	had | contributed
0.75	-13.6769	lwlm | when
0.75	-13.6769	exact | expensive
0.555556	-13.6769	(x | (y
0.216667	-13.6769	note that | =
0.5	-13.6769	database | web
0.127907	-13.677	improved | to
0.310345	-13.677	generation | dialogue
0.347826	-13.6771	questions | same
0.109244	-13.6771	the | on.
0.0597561	-13.6771	related | in
0.24	-13.6771	to | interval
0.116822	-13.6772	and | candidates
0.202899	-13.6772	for | assessment
0.461538	-13.6773	aic | for
0.310345	-13.6774	transliteration | language
0.160714	-13.6774	performance | feature
0.5	-13.6774	as | suffix.
0.666667	-13.6775	major | automatic
0.209677	-13.6775	query | document
0.185185	-13.6775	reasons | in
0.173913	-13.6775	we | namely
0.411765	-13.6776	can be | pattern
0.12	-13.6777	reference | for
0.17	-13.6778	locations | the
0.163462	-13.6778	on | pairs.
0.75	-13.6778	form | asks
0.75	-13.6778	micro-averaged | 1
0.75	-13.6778	lmbot | →ψ
0.75	-13.6778	hmm-lda | selection
0.461538	-13.6778	examples | (1)
0.0709343	-13.6778	of | shared
0.244444	-13.6781	one | language.
0.555556	-13.6781	when | became
0.126374	-13.6781	and | semantics
0.157895	-13.6782	lexical | an
0.205882	-13.6783	the | z.
0.461538	-13.6784	very | results.
0.0810811	-13.6784	the | author | the
0.175824	-13.6785	4 | experiments
0.321429	-13.6785	and | stimulus
0.321429	-13.6785	using | achieves
0.411765	-13.6785	introduced | algorithm
0.75	-13.6786	treebanks | common
0.75	-13.6786	and | mckeown.
0.75	-13.6786	so | net
0.75	-13.6786	transliterations | transliteration
0.75	-13.6786	etc. | obtain
0.75	-13.6786	information-oriented | tasks
0.2	-13.6786	this | scheme
0.061039	-13.6787	structure | and
0.124324	-13.6788	2008), | al.,
0.0576037	-13.6788	important | of
0.555556	-13.6788	conclusion | section
0.666667	-13.679	head | head,
0.0584795	-13.679	the | directly
0.226415	-13.679	(2003) | in
0.368421	-13.679	word | deletion
0.428571	-13.6791	assigned | highest
0.101818	-13.6792	used | all
0.0440294	-13.6792	as | an
0.0747126	-13.6792	for example, | for
0.171717	-13.6794	4.1 | from
0.75	-13.6795	feng | lapata
0.75	-13.6795	have | fingerprints
0.255814	-13.6795	corpus | evaluate
0.177778	-13.6795	in | structures.
0.0577367	-13.6797	after | of
0.0868421	-13.6797	our | has
0.220339	-13.6797	features | important
0.5	-13.6797	implemented | standard
0.666667	-13.6798	resulting | output.
0.461538	-13.6799	models, | where
0.085213	-13.6799	reported | and
0.141844	-13.6799	and | response
0.27027	-13.6799	most | small
0.17	-13.6801	the | recursive
0.25	-13.6801	to | components.
0.235294	-13.6801	over | three
0.75	-13.6804	coverage | precision,
0.75	-13.6804	run | ghz
0.75	-13.6804	string | calculated
0.75	-13.6804	testing | his
0.666667	-13.6806	length | σ
0.216667	-13.6806	one | time.
0.142857	-13.6807	we | dependencies
0.115207	-13.6807	conditional | and
0.172043	-13.6808	discriminative | et
0.0808989	-13.6809	a | through
0.5	-13.6809	critical | which
0.132911	-13.681	four | with
0.0974026	-13.6811	under | a
0.75	-13.6812	size | 16,
0.75	-13.6812	recognition, | process
0.75	-13.6812	type, | trigger
0.75	-13.6812	aac | potential
0.75	-13.6812	system | transliterations.
0.75	-13.6812	terminal | basic
0.75	-13.6812	intermediate | advanced
0.165049	-13.6812	however, | because
0.235294	-13.6812	on | evaluation.
0.555556	-13.6813	common | pair
0.216667	-13.6813	and | biased
0.255814	-13.6813	an | making
0.137255	-13.6813	textual | a
0.108	-13.6813	of | parts
0.388889	-13.6814	a | option.
0.261905	-13.6814	class | its
0.5	-13.6814	methods | introduced
0.173913	-13.6816	but | most
0.3	-13.6817	degrades | the
0.321429	-13.6818	modeling | over
0.555556	-13.6819	among | comparable
0.666667	-13.6821	generation | tasks.
0.75	-13.6821	(although | significant
0.75	-13.6821	complex | hindi
0.75	-13.6821	range | age
0.226415	-13.6821	prominent | and
0.117073	-13.6822	follow | of
0.321429	-13.6822	since | space
0.102941	-13.6822	our | best
0.411765	-13.6822	to | record.
0.244444	-13.6823	which | i.e.
0.216667	-13.6823	redundant | and
0.168317	-13.6823	that | correlated
0.368421	-13.6823	evaluation | ability
0.230769	-13.6823	with | until
0.205882	-13.6825	our | experiment
0.202899	-13.6826	contains | on
0.411765	-13.6827	to | financial
0.255814	-13.6827	verb, | in
0.666667	-13.6828	richness | language
0.0868421	-13.6829	we | multiple
0.75	-13.6829	taken | helps
0.75	-13.6829	part | geo.
0.75	-13.6829	techniques | co-training
0.75	-13.6829	incorrect | q1.
0.75	-13.6829	as | sanity
0.363636	-13.6829	future | include
0.209677	-13.683	that | wrong
0.125	-13.683	approach | word
0.121693	-13.6831	approach | language
0.5	-13.6831	give | provide
0.224138	-13.6832	this | indicate
0.13986	-13.6832	we | concepts
0.143885	-13.6832	each | n
0.321429	-13.6833	words | probabilities
0.321429	-13.6833	using | outperforms
0.3	-13.6834	part-of-speech | features
0.153226	-13.6834	are | length
0.75	-13.6838	implementation, | hash
0.75	-13.6838	different | adopt
0.75	-13.6838	corpus | dialogs
0.555556	-13.6838	survey | work
0.0467572	-13.6838	parser | the
0.140845	-13.6838	not | due
0.347826	-13.6839	independence | to
0.25	-13.684	responses | from
0.185185	-13.684	sample | with
0.388889	-13.6841	tuning | used
0.116279	-13.6842	known | in
0.108787	-13.6842	entailment | and
0.138889	-13.6842	experiment | on
0.224138	-13.6842	nr | nr
0.25	-13.6843	as | valid
0.666667	-13.6844	content | repeated
0.147287	-13.6844	a | actual
0.147287	-13.6844	we | review
0.0604534	-13.6844	rules | in
0.75	-13.6846	knowledge | diferent
0.75	-13.6846	character | 0.
0.155172	-13.6846	and | actions
0.144928	-13.6846	have | performance
0.127907	-13.6847	studies | al.,
0.461538	-13.6847	2. | label
0.277778	-13.6847	the | stays
0.5	-13.6849	0.08 | 0.10
0.363636	-13.685	these | alternative
0.310345	-13.6851	table | shows | examples
0.3	-13.6852	bootstrapping | from
0.461538	-13.6852	obviously, | not
0.285714	-13.6853	or | forms
0.0529412	-13.6853	7 | the
0.150794	-13.6853	to | best | our
0.363636	-13.6854	with | metrics,
0.75	-13.6855	capture | toward
0.428571	-13.6855	however, | substantial
0.244444	-13.6856	apply | training
0.0353601	-13.6857	where | a
0.555556	-13.6857	despite | method
0.189873	-13.6857	our | summarization
0.159292	-13.6857	to | issues
0.0993151	-13.6858	4 | 6
0.428571	-13.686	funded | commission
0.428571	-13.686	preposition | but
0.135484	-13.686	we | our | two
0.144928	-13.6862	the | modeled
0.347826	-13.6863	the | ast.
0.388889	-13.6863	scores, | not
0.220339	-13.6863	the | modality
0.216667	-13.6863	dependencies | are
0.75	-13.6864	children | robot
0.75	-13.6864	corresponds | npk,
0.75	-13.6864	inducer | gold
0.555556	-13.6864	graphical | plate
0.173913	-13.6864	was | national
0.261905	-13.6865	europarl | news
0.172043	-13.6865	an | measure
0.368421	-13.6865	will | towards
0.137255	-13.6866	is not | or
0.285714	-13.6867	results | initial
0.666667	-13.6867	conversational | user
0.27027	-13.6867	a | indices
0.461538	-13.6868	parsing | span
0.159292	-13.6868	connected | a
0.321429	-13.687	random | were
0.555556	-13.687	e. | c.
0.0919881	-13.687	against | and
0.24	-13.687	word | thus
0.0553781	-13.6872	however, | that
0.75	-13.6872	constraints | part-whole
0.25	-13.6872	them | other
0.0333717	-13.6872	as a | of
0.168317	-13.6873	this | certain
0.666667	-13.6874	provided | required
0.666667	-13.6874	extraction | primarily
0.096463	-13.6874	there | such
0.220339	-13.6875	to | removing
0.285714	-13.6876	provides | number
0.0909091	-13.6876	the | unit
0.125	-13.6877	but | such
0.333333	-13.6878	improvements | models
0.368421	-13.6878	ranking | aspect
0.461538	-13.6879	how | entities
0.411765	-13.6879	text | patterns
0.310345	-13.688	translations | parallel
0.388889	-13.6881	but | yet
0.75	-13.6881	records | every
0.75	-13.6881	remain | annotation
0.0891365	-13.6881	to | complex
0.333333	-13.6881	assigned | label
0.0552017	-13.6882	considered | the
0.368421	-13.6882	all | streams
0.1	-13.6885	4 | from
0.0906516	-13.6885	and | common
0.0551432	-13.6886	if | =
0.133758	-13.6886	interpolation | the
0.75	-13.6889	estimates | expected
0.75	-13.6889	nonterminals | strings
0.388889	-13.689	the | semiring.
0.666667	-13.689	0, | words
0.213115	-13.689	our | models,
0.140845	-13.6892	all | information | the
0.133758	-13.6892	trained | corpus
0.277778	-13.6892	from | 1)
0.333333	-13.6893	human | automatically
0.290323	-13.6895	services | to
0.461538	-13.6895	of | 4480
0.0946372	-13.6895	a | subject
0.255814	-13.6895	evaluated | system
0.141844	-13.6895	you | and
0.123656	-13.6896	reordering | for
0.192308	-13.6896	colour | in
0.3	-13.6896	a | heavily
0.197183	-13.6896	observations | a
0.127907	-13.6897	scheme | in
0.213115	-13.6897	the | words:
0.205882	-13.6898	from | various
0.261905	-13.6898	first | towards
0.75	-13.6898	title, | article
0.461538	-13.69	larger | size
0.0909091	-13.69	word | one
0.5	-13.69	parser | genia
0.5	-13.69	thus | generate
0.333333	-13.6901	finds | by
0.27027	-13.6901	measures | different
0.127168	-13.6901	2011 | on | processing,
0.277778	-13.6902	output | sentence
0.24	-13.6903	turns | turns
0.388889	-13.6904	emails | men
0.117647	-13.6904	error | by
0.0707965	-13.6904	3 | 1
0.25	-13.6906	systems | task.
0.192308	-13.6907	distribution | document
0.321429	-13.6907	many | being
0.75	-13.6907	patterns | groups
0.230769	-13.6907	by | hypotheses
0.555556	-13.6908	that | viewpoint.
0.555556	-13.6908	includes | features.
0.173913	-13.6908	amount | by
0.209677	-13.6909	previous | one
0.428571	-13.6909	where | markov
0.0909091	-13.691	algorithm | as
0.168317	-13.691	this | frequency
0.134615	-13.691	providing | and
0.0835381	-13.6911	a | model | from
0.226415	-13.6911	assumption | can
0.175824	-13.6912	expressed | or
0.130178	-13.6914	this | uses
0.75	-13.6915	prepositions | syntactic
0.555556	-13.6915	many | wrong
0.230769	-13.6915	by | spanish
0.333333	-13.6916	good | if
0.117073	-13.6917	this paper | an
0.0413174	-13.6917	the | words in
0.152	-13.6917	task | more
0.194444	-13.6917	these | important
0.118227	-13.6918	and | polarity
0.129412	-13.6919	feature | 1
0.141844	-13.6919	noun | an
0.666667	-13.692	issues | synchronous
0.0986395	-13.692	features, | to
0.115207	-13.692	with | three
0.555556	-13.6921	step | α
0.197183	-13.6921	they also | the
0.0592593	-13.6922	of | speech
0.0810185	-13.6923	linear | and
0.113122	-13.6923	empirical | and
0.5	-13.6923	using | heuristic.
0.75	-13.6924	na¨ıve | implementation
0.75	-13.6924	sigdat | methods
0.0774059	-13.6924	the | h
0.25	-13.6924	first | do
0.333333	-13.6924	present, | is
0.461538	-13.6926	by | headed
0.333333	-13.6927	synsets | wordnet
0.165049	-13.6927	of | variables.
0.277778	-13.6927	and | environment.
0.0676329	-13.6928	in | probability
0.666667	-13.6928	evaluation | respective
0.428571	-13.6929	performed | following
0.5	-13.6929	in | layers.
0.5	-13.6929	method | document.
0.0833333	-13.693	there | which
0.0620783	-13.6931	knowledge | and
0.152	-13.6931	each | is | using
0.75	-13.6933	prefers | if
0.0648968	-13.6934	it is | as
0.428571	-13.6934	option | not
0.159292	-13.6934	is | character
0.333333	-13.6935	isolate | of
0.411765	-13.6936	syntactic | learned
0.150794	-13.6936	cost | is
0.134615	-13.6938	the | addressed
0.388889	-13.6939	classification | svm
0.290323	-13.6939	correct | its
0.27027	-13.6939	an | turn
0.230769	-13.6939	words | alignment
0.182927	-13.694	strongly | to
0.75	-13.6941	tagged | polar
0.75	-13.6941	limiting | possible
0.75	-13.6941	virtual | task.
0.75	-13.6941	svms | perform
0.75	-13.6941	making | harder
0.75	-13.6941	ssc | already
0.555556	-13.6941	features | morpho-syntactic
0.5	-13.6941	changes | reordering
0.197183	-13.6942	grammatical | are
0.061039	-13.6942	the | sentence,
0.347826	-13.6943	while | performed
0.666667	-13.6943	wikipedia | text,
0.0716846	-13.6943	we | scores
0.170213	-13.6943	the | character.
0.159292	-13.6943	statistical | translation.
0.428571	-13.6944	similarity | fragments
0.0414913	-13.6945	but | a
0.141844	-13.6945	contained | and
0.5	-13.6946	approaches | extracting
0.5	-13.6946	text | detect
0.163462	-13.6947	semantic | text
0.0914454	-13.6947	simple | for
0.0859375	-13.6947	of | m
0.0694915	-13.6948	2 | with
0.310345	-13.6948	exceptions | and
0.121693	-13.695	shows that | to
0.75	-13.695	papers | title
0.411765	-13.695	1© | 2©
0.277778	-13.695	our | text,
0.666667	-13.6951	basic | set.
0.5	-13.6952	verify | using
0.555556	-13.6953	considering | words.
0.110169	-13.6954	the | default
0.290323	-13.6955	predict | using
0.159292	-13.6955	examples | are | and
0.0684039	-13.6955	in | dialogue
0.0813953	-13.6955	the | actions
0.0802752	-13.6957	and | associated
0.230769	-13.6958	the | epoch
0.666667	-13.6959	were | long.
0.75	-13.6959	facilitate | collection
0.75	-13.6959	extraction), | education
0.75	-13.6959	constituents | whole
0.75	-13.6959	distribution | 2001;
0.75	-13.6959	annotation scheme | level
0.75	-13.6959	iso | act
0.75	-13.6959	engkoo | its
0.0336901	-13.6961	the | order
0.32	-13.6962	de, | en
0.194444	-13.6964	on the other hand, | are
0.5	-13.6964	event | fillers.
0.0350577	-13.6965	and | all
0.0963455	-13.6965	that | like
0.388889	-13.6966	learned | models
0.310345	-13.6966	our | reduced
0.555556	-13.6966	separate | during
0.75	-13.6967	extensions | additional
0.220339	-13.6967	proposition | a
0.202899	-13.6968	it | phrases
0.175824	-13.6969	our | performance.
0.461538	-13.6969	only | effects
0.5	-13.6969	graph | collection
0.5	-13.6969	results | predicates
0.126437	-13.6971	all | target
0.220339	-13.6972	ambiguous | as
0.235294	-13.6972	so | no
0.32	-13.6973	on | effort
0.156522	-13.6975	parts | a
0.165049	-13.6975	extensive | the
0.0893855	-13.6976	we | future
0.285714	-13.6979	japanese | word
0.0973154	-13.6979	for | system.
0.118812	-13.698	formula | the
0.290323	-13.6982	approach | web
0.666667	-13.6982	85 | 65
0.222222	-13.6982	an | procedure
0.0665635	-13.6982	in | many
0.140845	-13.6984	the | members
0.75	-13.6985	knowledge | lab
0.75	-13.6985	deal | interest
0.75	-13.6985	tokenizer | 2003)
0.155172	-13.6985	is | goal
0.230769	-13.6985	overlap | for
0.230769	-13.6985	has | low
0.161905	-13.6985	to | fewer
0.347826	-13.6987	multiple | models
0.0370549	-13.6987	it | to | a
0.0714286	-13.6987	and | knowledge
0.180723	-13.6987	some | do
0.25	-13.6987	stack | the | stack.
0.5	-13.6987	al | has
0.209677	-13.6989	who | an
0.0955414	-13.699	automatic | for
0.347826	-13.6991	for example, | label
0.0756646	-13.6991	we compare | the
0.09699	-13.6991	r | s
0.555556	-13.6992	confusion | &
0.75	-13.6993	dd | results
0.75	-13.6993	chan | 2009),
0.25	-13.6993	like | many
0.151261	-13.6994	r | with
0.310345	-13.6995	alignment | some
0.147287	-13.6995	2) | in
0.461538	-13.6996	approach | optimization
0.263158	-13.6996	topic | both
0.168317	-13.6996	model | no
0.0812065	-13.6997	now | of
0.666667	-13.6997	algorithm | descent
0.368421	-13.6997	z | +
0.428571	-13.6998	clustering | articles
0.5	-13.6998	during | random
0.189873	-13.6999	can | errors
0.333333	-13.7001	annotation | single
0.1875	-13.7002	however | is
0.222222	-13.7002	presented | paper
0.75	-13.7002	stefano, | and
0.428571	-13.7003	refine | in
0.428571	-13.7003	different | tasks,
0.116505	-13.7003	to | problem,
0.141844	-13.7004	and | characters
0.222222	-13.7005	we | book
0.555556	-13.7005	data, | prior
0.0883978	-13.7006	of | before
0.347826	-13.7007	white, | and
0.388889	-13.7007	a | fragment.
0.163462	-13.7007	method | all
0.112613	-13.7007	of | links
0.147287	-13.7008	we | evidence
0.194444	-13.7009	readability | a
0.0651515	-13.7009	features | by
0.103448	-13.701	a | object
0.0769231	-13.701	the | language,
0.75	-13.7011	conclude, | with
0.75	-13.7011	institute | 12015
0.75	-13.7011	have | types:
0.75	-13.7011	structure | hiden
0.388889	-13.7012	arise | are
0.191781	-13.7012	judge | of
0.310345	-13.7013	,β | c
0.138889	-13.7013	abstract | present
0.0715564	-13.7014	length | and
0.0940439	-13.7014	also | model
0.101083	-13.7014	outperforms | and
0.146154	-13.7015	but | do
0.0925926	-13.7015	in | work.
0.109244	-13.7016	general | as
0.119792	-13.7016	phrase | or
0.23913	-13.7016	and | y.
0.5	-13.7016	number | ngrams
0.114833	-13.7017	two | models
0.461538	-13.7017	finally, | 7
0.461538	-13.7017	bigrams | using
0.0686767	-13.7017	occur | the
0.368421	-13.7019	features | weighted
0.333333	-13.702	themselves | as
0.213115	-13.702	many | semantic
0.388889	-13.7021	target | contain
0.0753564	-13.7021	sense | in
0.163462	-13.7026	to | patterns.
0.146154	-13.7026	that | c
0.368421	-13.7027	two | scoring
0.75	-13.7028	was | cssr
0.230769	-13.7028	this | it.
0.230769	-13.7028	in | inference.
0.114833	-13.7029	that | combination
0.0923077	-13.7029	of | f
0.310345	-13.7031	manual | annotation
0.368421	-13.7031	of | agent’s
0.132075	-13.7033	used | text
0.173913	-13.7033	each | word.
0.226415	-13.7035	being | more
0.222222	-13.7036	baseline | method
0.149606	-13.7036	sampling | in
0.75	-13.7037	different | 2009a;
0.75	-13.7037	only | annot.
0.191781	-13.7037	of | capturing
0.191781	-13.7037	abstract | has
0.106996	-13.7038	to | (the
0.428571	-13.7038	script | with
0.428571	-13.7038	message | visual
0.194444	-13.7038	use | relation
0.213115	-13.7038	two | annotated
0.138889	-13.7039	rules, | of
0.0966667	-13.7039	4.1 | a
0.461538	-13.7039	occurs | different
0.461538	-13.7039	learning | margin
0.105469	-13.7039	one | our
0.5	-13.7039	is | preferable
0.138889	-13.7041	and | share
0.189873	-13.7041	message | is
0.127168	-13.7041	our | extraction
0.0556769	-13.7042	of | them
0.206349	-13.7043	on | review
0.0476575	-13.7043	each | by
0.347826	-13.7044	from | precision.
0.666667	-13.7044	defines | class
0.666667	-13.7044	procedure | sense
0.555556	-13.7044	examples | annotators
0.0976431	-13.7044	increases | the
0.571429	-13.7045	(except | involving
0.571429	-13.7045	she | he
0.0854922	-13.7046	bigram | the
0.244444	-13.7048	find | their
0.428571	-13.7048	since | produce
0.0340862	-13.705	in | data
0.461538	-13.705	often | sentence.
0.185185	-13.705	2011 | task
0.125	-13.7051	as | system.
0.5	-13.7051	surface | features.
0.75	-13.7054	construct | repository
0.75	-13.7054	policy | turn
0.75	-13.7054	maltparser | features,
0.263158	-13.7054	information | local
0.145038	-13.7055	which | words.
0.108787	-13.7055	around | the
0.114833	-13.7057	when | used
0.0782609	-13.7057	this | method
0.571429	-13.7058	literature | approaches
0.571429	-13.7058	(∃u,v) | uvnegationslash=
0.3	-13.706	presented | different
0.0492228	-13.706	second | of
0.152542	-13.7061	similarity | semantic
0.23913	-13.7061	are | subjective
0.0909091	-13.7062	get | of
0.3	-13.7063	can | improving
0.75	-13.7063	whom | when
0.75	-13.7063	summarize | topics
0.75	-13.7063	|ci|
0.75	-13.7063	epi | due
0.75	-13.7063	characters | keep
0.333333	-13.7063	biomedical | 2011,
0.222222	-13.7065	that | phrases.
0.461538	-13.7066	graphs | popular
0.173913	-13.7066	both | english
0.277778	-13.7066	(zhou | al.,
0.27027	-13.7067	people | more
0.244444	-13.7068	whether | target
0.5	-13.7068	were | judged
0.166667	-13.7069	four | from
0.0540541	-13.7069	1, | the
0.263158	-13.7069	subset | training
0.156522	-13.707	however, | will
0.143939	-13.707	those | do
0.0511152	-13.707	order | a
0.104247	-13.707	point | a
0.310345	-13.7071	natural | tasks.
0.571429	-13.7071	removing | elements
0.571429	-13.7071	closure | cells
0.75	-13.7072	summarizing | document
0.75	-13.7072	antecedent | within
0.75	-13.7072	feature | null,
0.75	-13.7072	bayes | linear
0.235294	-13.7073	corpus, | an
0.117647	-13.7074	however, | data
0.368421	-13.7074	research | done
0.5	-13.7074	models | brown
0.5	-13.7074	performance | ie
0.347826	-13.7076	in | haghighi
0.555556	-13.7076	hierarchical | 2007)
0.185185	-13.7076	supported | national | and
0.172043	-13.7076	beginning | and
0.115385	-13.7077	required | to
0.137931	-13.7077	for | corpora.
0.310345	-13.7078	used | assign
0.571429	-13.7078	model | acc.
0.0701754	-13.7079	a | related
0.0614973	-13.7079	models | on
0.110132	-13.7079	the | forward
0.368421	-13.7079	scores | range
0.2	-13.7079	of | ccg
0.75	-13.7081	but | steadily
0.75	-13.7081	construction | (as
0.75	-13.7081	still | again
0.75	-13.7081	(or | documents.
0.178571	-13.7082	each | head
0.0453548	-13.7082	the | classifier
0.114833	-13.7083	1 introduction | as
0.0436535	-13.7084	to | target
0.571429	-13.7085	probabilistic | topics.
0.571429	-13.7085	 is |  a
0.0841837	-13.7087	actually | the
0.163462	-13.7087	complex | that
0.203125	-13.7087	used | combination
0.230769	-13.7087	similarity | distance
0.428571	-13.7088	them | sentences.
0.255814	-13.7088	expressions | their
0.555556	-13.7089	0.40 | 0.60
0.75	-13.709	after | fill
0.140845	-13.709	with | candidate
0.571429	-13.7091	probabilities | pscfgs.
0.571429	-13.7091	ex | (check)
0.103053	-13.7092	we | local
0.5	-13.7092	such | control
0.12234	-13.7092	reference | is
0.347826	-13.7093	phenomenon | as
0.388889	-13.7093	outside | from
0.213115	-13.7093	x1 | a
0.165049	-13.7096	of | vocabulary.
0.0923077	-13.7096	decoding | and
0.2	-13.7096	of | areas
0.290323	-13.7097	is | vi
0.194444	-13.7097	construction | that
0.428571	-13.7098	being | part
0.571429	-13.7098	attachment | (las)
0.155172	-13.7099	of | expressions.
0.244444	-13.71	a | purposes.
0.194444	-13.71	methodology | for
0.0890805	-13.71	in | relative
0.32	-13.71	localization | a
0.347826	-13.7101	3rd | in
0.226415	-13.7101	fact | same
0.555556	-13.7102	small | performance.
0.197183	-13.7102	for | phrases.
0.27027	-13.7102	4 | dataset
0.130435	-13.7103	relation, | the
0.461538	-13.7104	other | strategies
0.571429	-13.7105	university | toronto
0.571429	-13.7105	author | author’s
0.571429	-13.7105	accuracy(%) | accuracy
0.75	-13.7107	generating | exercises
0.140845	-13.7107	which | standard
0.0356016	-13.7108	i | the
0.0550162	-13.7108	the | next
0.3	-13.7109	by | links
0.0652504	-13.7109	lexicon | the
0.0614973	-13.711	the | of | particular
0.0675241	-13.711	we | use | in
0.12963	-13.711	relatedness | and
0.0661417	-13.7111	consists | of
0.155172	-13.7112	to | examples.
0.0284974	-13.7112	the same | the
0.347826	-13.7113	into | separate
0.25	-13.7113	while | might
0.428571	-13.7113	more | composed
0.428571	-13.7113	type | →
0.368421	-13.7113	preposition | between
0.191781	-13.7115	clinical | and
0.0652504	-13.7115	the | algorithms
0.27027	-13.7115	mention | entity | entity
0.277778	-13.7115	unlike | use
0.3	-13.7116	all | 100
0.75	-13.7116	text | wiki
0.197183	-13.7116	as | parameter
0.347826	-13.7117	semantics | between
0.156522	-13.7118	observation | in
0.206349	-13.7118	is | faster
0.333333	-13.7118	kendall’s | a
0.203125	-13.7118	rule | if
0.277778	-13.7118	in | adverbs
0.103448	-13.7119	a | selection
0.2	-13.7119	a | estimates
0.244444	-13.712	set | (e.g.
0.226415	-13.712	strong | on
0.148438	-13.7121	that | combined
0.5	-13.7121	j=1 | m
0.666667	-13.7122	user | detail
0.263158	-13.7122	phrase | trees
0.263158	-13.7122	words | distance
0.428571	-13.7123	achieved | systems
0.347826	-13.7125	systems | performance.
0.0285879	-13.7125	is | an
0.75	-13.7125	there | total.
0.165049	-13.7125	discuss | section
0.571429	-13.7125	grammars | statistical
0.0685619	-13.7125	in | specific
0.116505	-13.7125	such | language
0.168421	-13.7126	categories, | and
0.0821256	-13.7126	also | or
0.0940439	-13.7127	element | the
0.255814	-13.7127	1 | frequency
0.125	-13.713	as | (e.g.
0.0737864	-13.7131	phrases | and
0.119171	-13.7131	exact | and
0.0867209	-13.7132	for | without
0.571429	-13.7132	collocation | compositionality
0.0960265	-13.7132	in | (1)
0.180723	-13.7133	aligned | that
0.0829268	-13.7133	that | terms
0.10917	-13.7133	our work | of
0.75	-13.7134	lexicon | entries.
0.75	-13.7134	because | summaries,
0.25	-13.7134	quotes | the
0.222222	-13.7135	not | comparable
0.461538	-13.7136	consist | one
0.0676568	-13.7136	low | the
0.13986	-13.7136	only | most
0.0344393	-13.7137	it | a | of
0.222222	-13.7138	the | overfitting
0.571429	-13.7138	entries | considered
0.388889	-13.7139	adaptations | the
0.0809524	-13.714	the | indicating
0.117647	-13.7141	web | in
0.147287	-13.7141	in | 5, | and
0.555556	-13.7141	out | 15
0.104247	-13.7141	as | positive
0.140845	-13.7142	in | elements
0.178571	-13.7143	that | equivalent
0.178571	-13.7143	also | rules
0.75	-13.7143	distributions | lda
0.388889	-13.7144	reported | both
0.388889	-13.7144	of | prior.
0.180723	-13.7144	that | developed
0.180723	-13.7144	to | preferences
0.428571	-13.7144	structures | spoken
0.666667	-13.7146	cb | clause
0.0628571	-13.7146	(the | the
0.0963455	-13.7146	line | and
0.157407	-13.7146	and | emotion
0.15	-13.7146	while | may
0.106122	-13.7147	is | sentences.
0.0634006	-13.7147	analysis | a
0.0461073	-13.7148	in | both
0.290323	-13.7149	training | tuning
0.428571	-13.7149	parser | 2009).
0.0911854	-13.715	topics | and
0.0478508	-13.7151	given | is
0.75	-13.7151	open | lexicon
0.75	-13.7151	combinations | models.
0.75	-13.7151	verb, | adjective
0.178571	-13.7152	efforts | al.,
0.0726577	-13.7152	the | get
0.571429	-13.7152	expression | opinions
0.2	-13.7152	and | variables.
0.173913	-13.7153	penalty | to
0.333333	-13.7153	of | remained
0.25	-13.7154	to | dimensionality
0.108696	-13.7154	is | step
0.126437	-13.7154	precision | that
0.152542	-13.7154	official | of
0.152542	-13.7154	paraphrase | as
0.105691	-13.7154	national | of
0.0497778	-13.7155	combination | the
0.290323	-13.7156	manual | this
0.32	-13.7156	models | interpolation
0.32	-13.7156	4 | various
0.0756646	-13.7156	performs | the
0.5	-13.7157	will | case.
0.0979021	-13.7158	work | language
0.75	-13.716	j) | l
0.75	-13.716	counts | frequencies
0.75	-13.716	nonterminal | unique
0.666667	-13.7161	consistency | events
0.666667	-13.7161	zero | subject
0.148438	-13.7161	all | so
0.333333	-13.7161	sentence | words,
0.310345	-13.7162	and | lexicons.
0.0982456	-13.7162	a | news
0.172043	-13.7163	analysis | using
0.428571	-13.7164	role | within
0.32	-13.7164	algorithm | cluster
0.23913	-13.7165	in | referent
0.290323	-13.7166	for | presentation
0.368421	-13.7166	with | components.
0.168421	-13.7167	to | other,
0.143939	-13.7168	of | restricted
0.0963455	-13.7168	randomly | of
0.23913	-13.7168	consequently, | a
0.666667	-13.7169	high | costs
0.75	-13.7169	probability | transitions
0.461538	-13.7169	obtained | million
0.120419	-13.717	is | bleu
0.368421	-13.717	also | interest
0.206349	-13.7171	the | pos.
0.32	-13.7171	for example, | documents
0.102662	-13.7172	to | just
0.226415	-13.7173	structure | using
0.226415	-13.7173	and | s1
0.216667	-13.7173	nonterminals | a
0.333333	-13.7173	method | frequency
0.1875	-13.7174	testing, | the
0.368421	-13.7174	using | correction
0.230769	-13.7174	thus, | model
0.23913	-13.7174	pathway | in
0.166667	-13.7175	are | capture
0.130435	-13.7175	labeled | from
0.127168	-13.7175	could | from
0.388889	-13.7176	aligned | pair
0.0821256	-13.7176	the | core
0.666667	-13.7177	human | salient
0.75	-13.7178	approached | problem
0.14876	-13.7178	principle | the
0.571429	-13.7179	(abeill´e | al.,
0.23913	-13.7179	however, | able
0.388889	-13.718	sparsity, | to
0.290323	-13.718	meteor | on
0.461538	-13.718	model | gmm
0.5	-13.718	kernel | i.e.
0.209677	-13.7181	we | simpler
0.555556	-13.7181	also, | like
0.333333	-13.7181	– | possible
0.347826	-13.7183	so | related
0.175824	-13.7183	svm | on
0.461538	-13.7185	funding | research
0.244444	-13.7186	also | patterns
0.0788288	-13.7186	to | data,
0.0669856	-13.7186	rate | the
0.571429	-13.7186	extracted | textrunner
0.5	-13.7186	tokenized | tokenized
0.5	-13.7186	confidence | hypothesis
0.347826	-13.7187	against | human
0.75	-13.7187	conversation | therefore
0.75	-13.7187	optimization | log-likelihood
0.75	-13.7187	no. | china
0.209677	-13.7188	over | document
0.226415	-13.7189	let | ∈
0.119171	-13.7189	gender | and
0.388889	-13.719	responses | english
0.277778	-13.719	our | closely
0.32	-13.719	summationdisplay | parenrightbig
0.0992908	-13.7191	these | features
0.194444	-13.7191	that | linked
0.213115	-13.7191	technique | by
0.23913	-13.7191	features | derived
0.146154	-13.7192	real | to
0.5	-13.7192	first | designed
0.3	-13.7194	what | will
0.428571	-13.7194	get | improvement
0.32	-13.7194	type | role
0.368421	-13.7196	(2008) | approach
0.191781	-13.7197	baselines | a
0.0215589	-13.7197	the | was | the
0.0419274	-13.7198	topic | the
0.388889	-13.7199	that | missed
0.310345	-13.7199	determining | or
0.310345	-13.7199	a | parameter.
0.28125	-13.7199	for | form.
0.571429	-13.7199	in | simplifications.
0.226415	-13.72	arguments | or
0.136986	-13.7203	it | context
0.0655226	-13.7204	of | .
0.290323	-13.7204	and | labeling.
0.75	-13.7205	f-measure | classes.
0.75	-13.7205	from | http:
0.0996441	-13.7206	this | data.
0.263158	-13.7206	explore | which
0.571429	-13.7206	workers | provide
0.461538	-13.7207	3.2 | process
0.461538	-13.7207	sentences, | lexical
0.666667	-13.7209	allows | document,
0.666667	-13.7209	methods | acronym
0.176471	-13.7209	three | words
0.0457614	-13.721	to be | and
0.5	-13.721	supported | commission
0.290323	-13.7211	research | systems
0.143939	-13.7211	overall | with
0.14876	-13.7211	from | articles
0.347826	-13.7212	alignment | translation.
0.090379	-13.7212	the | of | structures
0.388889	-13.7213	by | markers
0.209677	-13.7213	while | systems
0.571429	-13.7213	informatics, | abstract
0.75	-13.7214	information | management,
0.75	-13.7214	ido | bar-ilan
0.75	-13.7214	values | summed
0.75	-13.7214	lead | poor
0.368421	-13.7214	like | ,
0.0246725	-13.7215	we | by | the
0.244444	-13.7215	filtering | by
0.3	-13.7215	on | graph.
0.0302731	-13.7215	score | the
0.111111	-13.7216	of | ability
0.23913	-13.7219	any | of | afrl,
0.333333	-13.722	some | during
0.571429	-13.722	response | nlu
0.75	-13.7222	π24, | high
0.75	-13.7222	82 | 12
0.75	-13.7222	both | margin.
0.310345	-13.7224	similarities | as
0.0566474	-13.7224	to | similar
0.0484694	-13.7224	the | role
0.25641	-13.7225	jdpa | the
0.571429	-13.7226	study | pre-test
0.388889	-13.7227	was | 2011).
0.182927	-13.7227	recently, | to
0.555556	-13.7227	errors | significantly
0.218182	-13.7227	mean | on
0.105691	-13.7227	but | have
0.122905	-13.7227	the | chance
0.0260594	-13.7228	we | that | in
0.157407	-13.7228	candidates | by
0.461538	-13.7229	clusterings | by
0.368421	-13.7231	test | unseen
0.209677	-13.7233	scores | but
0.101887	-13.7233	1 | all
0.571429	-13.7233	sentences | esl
0.571429	-13.7233	unlikely | same
0.2	-13.7233	for | incremental
0.5	-13.7234	study | possible
0.109649	-13.7235	the | dev
0.388889	-13.7236	an | (which
0.182927	-13.7236	a | terms.
0.0920245	-13.7236	the | asr
0.127168	-13.7236	in | construction
0.25641	-13.7237	we | decipherment
0.0989399	-13.7237	as | best
0.0669935	-13.7239	rank | the
0.0738523	-13.7239	given | and | in
0.054721	-13.7239	with | using
0.388889	-13.724	such | derivation
0.666667	-13.724	include | indicator
0.75	-13.724	rank | entries.
0.555556	-13.724	wikipedia | authority
0.180723	-13.724	to | fine-grained
0.571429	-13.724	mention | antecedent
0.571429	-13.724	hence | across
0.133333	-13.7242	annotation | an
0.194444	-13.7242	list | used
0.333333	-13.7244	in | experiments, | only
0.35	-13.7244	,q | ,q
0.347826	-13.7245	after | obtained
0.0911854	-13.7245	each | into
0.226415	-13.7246	, | :
0.461538	-13.7246	way | corpora
0.263158	-13.7246	currently | by
0.134228	-13.7246	feature | between | and
0.555556	-13.7247	predicted | values
0.133333	-13.7248	count | a
0.75	-13.7249	following | hold:
0.75	-13.7249	engineer | alternative
0.75	-13.7249	algorithm | 1989)
0.172043	-13.7249	to | types,
0.104418	-13.725	of | computing
0.428571	-13.725	first, | user
0.136054	-13.725	morphology | the
0.137931	-13.725	by | positive
0.3	-13.7251	representation, | is
0.244444	-13.7253	were | up
0.368421	-13.7253	queries, | are
0.388889	-13.7254	f1 | 70.1
0.4	-13.7254	“play | forward
0.160377	-13.7254	and | empty
0.0423729	-13.7254	learning | and
0.0996441	-13.7255	for | 2009).
0.168421	-13.7255	systems | used
0.143939	-13.7256	in | attribute
0.209677	-13.7256	we | viterbi
0.461538	-13.7257	error | instead
0.0992908	-13.7257	that | class
0.35	-13.7257	4.2 | analysis
0.75	-13.7258	up | down.
0.75	-13.7258	books, | electronics
0.388889	-13.7259	connectives | two
0.388889	-13.7259	problem | algorithms
0.28125	-13.7259	the | good,
0.555556	-13.726	3.1 | actions
0.0938511	-13.7263	split | and
0.5	-13.7263	developed | part
0.127273	-13.7264	techniques | al.,
0.0960265	-13.7264	go | the
0.23913	-13.7264	learning | supervised
0.310345	-13.7265	work | monolingual
0.100372	-13.7265	is | generate
0.104	-13.7266	shows | results
0.153846	-13.7266	empty | and
0.75	-13.7267	inference | factored
0.75	-13.7267	hungarian | spanish
0.388889	-13.7268	scores | pairs.
0.388889	-13.7268	punctuation | words
0.101124	-13.7268	will | an
0.15	-13.7268	while | system
0.110132	-13.7269	unit | the
0.222222	-13.7269	features | automatically
0.25	-13.727	current | have
0.0957096	-13.727	semantic | which
0.0895954	-13.7273	text | that
0.35	-13.7273	words | mean
0.0381002	-13.7273	in | their
0.23913	-13.7273	process | over
0.172043	-13.7274	m | )
0.571429	-13.7274	therefore, | character
0.111628	-13.7275	is | m
0.35	-13.7277	test | comparing
0.0925926	-13.7277	the | the | side.
0.141791	-13.7279	translation | both
0.106996	-13.728	t | s
0.178571	-13.7281	for | compound
0.146154	-13.7281	average | words
0.25	-13.7282	show | achieve
0.0567787	-13.7282	which | the | which
0.0285462	-13.7283	a | of | from
0.0707804	-13.7284	known | the
0.333333	-13.7284	an exhaustive | of
0.127273	-13.7284	consistent | and
0.368421	-13.7284	no | size
0.123596	-13.7284	independent | and
0.75	-13.7285	most | measurement
0.25641	-13.7286	premise | to
0.0846561	-13.7289	summarization | and
0.141791	-13.7289	1999; | et
0.101124	-13.7289	2 | two
0.32	-13.7289	trees | but
0.145038	-13.729	3. | is
0.23913	-13.729	the | subtree.
0.206349	-13.7291	scheme | with
0.25	-13.7291	model | 2006).
0.139706	-13.7291	research | has
0.0370188	-13.7291	an | that | the
0.0935484	-13.7292	we | k
0.333333	-13.7292	acronym | as
0.555556	-13.7293	compiled | words
0.75	-13.7294	absolute | relative)
0.75	-13.7294	research | phenomena.
0.75	-13.7294	baseline | 1.0
0.388889	-13.7296	system | feedback
0.178571	-13.7296	we use | sentences
0.428571	-13.7296	score | minus
0.28125	-13.7296	that, | only
0.214286	-13.7296	use | estimate
0.222222	-13.7299	types | relations
0.277778	-13.7299	findings | this
0.555556	-13.73	improvement | reduction
0.185185	-13.73	class, | of
0.333333	-13.73	contains | over
0.230769	-13.73	are | process.
0.571429	-13.7301	$ | +
0.4	-13.7302	2009) | models
0.347826	-13.7303	types | arguments
0.347826	-13.7303	set | values.
0.347826	-13.7303	tokyo, | beijing,
0.75	-13.7303	rating | about
0.75	-13.7303	english | adj
0.124294	-13.7303	a | 2003).
0.388889	-13.7305	(section | section
0.3	-13.7305	highest-scored | the
0.101887	-13.7305	results | two
0.0554324	-13.7306	correlation | the
0.214286	-13.7306	translation | chinese
0.0615595	-13.7307	report | the
0.234043	-13.7308	another | between
0.12963	-13.7308	the | accuracy,
0.27027	-13.7308	lack | training
0.0648148	-13.7309	or | not
0.214286	-13.7309	tag | pos
0.133333	-13.731	correlations | and
0.113208	-13.731	the | years
0.0428667	-13.7311	although | the
0.75	-13.7312	linear | 3.1
0.428571	-13.7312	(from | or
0.428571	-13.7312	method | match
0.213115	-13.7312	explicitly | for
0.555556	-13.7313	distinctive | phrases
0.0643713	-13.7313	present | in
0.117347	-13.7313	observation | of
0.0735586	-13.7315	this | both
0.35	-13.7315	words | segmented
0.0756303	-13.7318	6 | for
0.203125	-13.7318	more than | as
0.263158	-13.7319	is | only.
0.2	-13.7319	possible | pairs
0.666667	-13.732	produce | binary
0.555556	-13.732	pittsburgh, | 15260
0.555556	-13.732	h | h,
0.555556	-13.732	command | command
0.555556	-13.732	(is|’s)
0.75	-13.7321	non-english | method
0.146154	-13.7321	was | most
0.428571	-13.7322	(t, | is
0.146341	-13.7322	in | native
0.571429	-13.7322	knight, | 2008).
0.0818859	-13.7322	of | subject
0.23913	-13.7322	proposed | better
0.0668842	-13.7323	important | in
0.152542	-13.7323	in | open
0.2	-13.7323	of | 4th
0.5	-13.7323	chain | its
0.388889	-13.7324	also | ways
0.0513308	-13.7324	also | with
0.154545	-13.7324	reduce | in
0.234043	-13.7325	of | bag-of-words
0.124294	-13.7325	was | into
0.0680272	-13.7325	selected | of
0.206349	-13.7326	and | dual
0.0898204	-13.7326	the | fragments
0.0731225	-13.7327	the | to make
0.186667	-13.7327	has | probability
0.290323	-13.7328	and | creative
0.263158	-13.7328	down | with
0.35	-13.7328	important | question
0.35	-13.7328	response | were
0.388889	-13.7329	crfs | al.,
0.0631424	-13.7329	hand, | the
0.12963	-13.7329	for | features,
0.5	-13.7329	pivot | source
0.5	-13.7329	3000 | 1
0.23913	-13.733	trees, | in
0.428571	-13.7332	meaning | paraphrase
0.35	-13.7332	7 | type
0.32	-13.7332	values | +
0.222222	-13.7333	hypotheses | are
0.042623	-13.7335	we | the | corpus
0.136986	-13.7335	in | resource
0.120419	-13.7335	also | results
0.2	-13.7335	as | evidence
0.122905	-13.7335	of | emotion
0.5	-13.7335	aspects | dependency
0.206349	-13.7336	the | pushes
0.35	-13.7336	sequence | output
0.35	-13.7336	• | problems
0.368421	-13.7337	object | same
0.75	-13.7339	five | runs.
0.25	-13.7339	furthermore, | more
0.139706	-13.734	polarity | a
0.203125	-13.734	tokens | an
0.3	-13.7341	approach | learned
0.4	-13.7341	comparing | comparing
0.197183	-13.7341	context | semantic
0.368421	-13.7341	or | texts.
0.115578	-13.7342	inference | for
0.428571	-13.7343	such | solution
0.3	-13.7344	to | deceptive
0.461538	-13.7345	performance | accuracy.
0.128049	-13.7345	considering | a
0.0947712	-13.7346	analysis | on
0.28125	-13.7347	construct | model
0.0796253	-13.7347	proceedings of | association
0.08	-13.7347	and | defined
0.0513069	-13.7347	head | the
0.388889	-13.7348	model | approximately
0.75	-13.7348	propose | third-order
0.75	-13.7348	most | hand-crafted
0.75	-13.7348	described | examples.
0.147541	-13.7348	the | canonical
0.222222	-13.7349	and | granularity
0.0154164	-13.7349	on | and
0.461538	-13.7351	s32s83 | s72
0.461538	-13.7351	=+ | bel(a,
0.194444	-13.7352	in | historical
0.290323	-13.7353	problem | 2009).
0.35	-13.7353	children | language
0.35	-13.7353	process | order
0.27027	-13.7353	specific | different
0.27027	-13.7353	second | their
0.209677	-13.7354	this | ability
0.119792	-13.7354	e | =
0.100746	-13.7356	remove | the
0.163265	-13.7356	basic | are
0.137931	-13.7356	we | finding
0.571429	-13.7356	msd | from
0.571429	-13.7356	s. | b.
0.186667	-13.7356	constructed | from
0.234043	-13.7359	these | result
0.137931	-13.7359	the | request
0.2	-13.7359	max | summationdisplay
0.100746	-13.736	a | correlation
0.25641	-13.736	extract | phrase
0.194444	-13.7361	example | was
0.164948	-13.7361	pattern | with
0.145038	-13.7363	some | other
0.25641	-13.7363	of | identified.
0.176471	-13.7363	science | usa
0.178571	-13.7364	of | parser,
0.101124	-13.7364	of | expected
0.0416141	-13.7364	results | for
0.4	-13.7365	base | output
0.0844327	-13.7365	from | english
0.75	-13.7366	metrics | ratings
0.75	-13.7366	perform | tagging.
0.074	-13.7366	two | or
0.25	-13.7366	use | binary
0.0707804	-13.7367	for | relation
0.152542	-13.7367	made | with
0.428571	-13.7368	all | sorted
0.218182	-13.7368	surprising | of
0.172043	-13.7369	to | terms.
0.234043	-13.737	model | variable
0.0609756	-13.737	and | university
0.206349	-13.7371	(np | (np
0.368421	-13.7372	rules | correct
0.2	-13.7372	and | paraphrasing
0.0715631	-13.7373	k | of
0.263158	-13.7373	the | mapping.
0.118557	-13.7374	in | taken
0.0935484	-13.7374	however, | are | to
0.0669935	-13.7375	very | in
0.178571	-13.7375	context | its
0.75	-13.7375	alignments | alignment,
0.75	-13.7375	• | reverb
0.0629575	-13.7375	obtained | of
0.0601604	-13.7376	when | are | the
0.290323	-13.7377	boundary | are
0.0926518	-13.7377	languages | a
0.571429	-13.7377	number | times.
0.571429	-13.7377	as | converse
0.197183	-13.7377	from | feedback
0.35	-13.7378	trained | biomedical
0.35	-13.7378	assumption | similar
0.107759	-13.7379	using | based
0.4	-13.7379	an unsupervised | method
0.108696	-13.7379	predicates | of
0.0567376	-13.7379	w | the
0.186667	-13.7379	extraction | semantic
0.203125	-13.7379	3.3 | are
0.388889	-13.738	found | extraction
0.0751566	-13.738	for | following
0.0883191	-13.738	project | the
0.0712946	-13.7381	the | surface
0.145038	-13.7382	order | not
0.0805687	-13.7382	our | performance
0.107296	-13.7383	we will | that
0.5	-13.7383	additionally, | both
0.666667	-13.7384	experiments, | comparable
0.75	-13.7384	slot | person
0.75	-13.7384	distributional | similarity,
0.75	-13.7384	01234 | p
0.263158	-13.7385	over | points
0.14876	-13.7386	many | language
0.234043	-13.7387	metrics | used
0.116162	-13.7387	6 | 0
0.136986	-13.7387	a | semantically
0.119171	-13.7388	features | information
0.388889	-13.739	patterns | similar
0.290323	-13.7391	statistical | most
0.206349	-13.7391	from | weight
0.0923567	-13.7391	with | standard
0.214286	-13.7391	4.2 | model
0.166667	-13.7392	the | keywords
0.333333	-13.7392	noisy | not
0.333333	-13.7392	arabic | using
0.0539112	-13.7392	out | of
0.0627737	-13.7392	arguments | the
0.0655226	-13.7393	it is | to | that
0.75	-13.7393	bengali | using
0.32	-13.7393	the | agent,
0.32	-13.7393	qmp | the
0.244444	-13.7394	that | solve
0.25641	-13.7394	crf | on
0.194444	-13.7394	limited | that
0.146341	-13.7396	the | held
0.5	-13.7396	english | conll
0.28125	-13.7397	however, | annotated
0.571429	-13.7398	parse | (collins
0.077951	-13.7399	semantic | by
0.0952381	-13.74	in | systems.
0.186667	-13.7401	our | consists | and
0.32	-13.7401	only | identified
0.75	-13.7402	jin-dong | center
0.75	-13.7402	t-score | best
0.75	-13.7402	chinese, | english.
0.14876	-13.7402	algorithm | using
0.0505618	-13.7403	a | lexical
0.145161	-13.7403	if | probability
0.35	-13.7404	labels | then
0.122222	-13.7405	target | or
0.571429	-13.7405	consider | aa
0.0660225	-13.7406	is | about
0.056338	-13.7407	this | set
0.119171	-13.7407	for | word.
0.4	-13.7408	perform | inference
0.133333	-13.7408	over | using
0.0343348	-13.7409	is | by | and
0.106383	-13.7409	we | take
0.203125	-13.7409	much | has
0.0677966	-13.7409	first | on
0.174419	-13.741	of | overlapping
0.0507042	-13.741	two | that
0.75	-13.7411	features | feature | results.
0.75	-13.7411	integrated | multimodal
0.75	-13.7411	study | modeled
0.75	-13.7411	mi | introduced
0.290323	-13.7412	lm | models
0.0666667	-13.7413	a | dialogue
0.0952381	-13.7413	token | and
0.206349	-13.7413	for | actions
0.045085	-13.7414	corpus | in
0.136986	-13.7415	have | learning
0.176471	-13.7416	proceedings | pp.
0.333333	-13.7416	a | person.
0.666667	-13.7417	propose | generative model
0.0650155	-13.7418	in | systems
0.0870787	-13.7418	current | a
0.0748441	-13.7418	among | of
0.1	-13.7419	sentences | or
0.428571	-13.742	of | family,
0.333333	-13.742	even | source
0.234043	-13.7421	the | compound.
0.154545	-13.7421	expressions | are
0.140741	-13.7421	from | 1.
0.25641	-13.7422	used | implementation
0.218182	-13.7423	dependents | the
0.170213	-13.7424	a | condition
0.0981818	-13.7424	authors | and
0.226415	-13.7425	with | metrics.
0.571429	-13.7425	questions | jeopardy!
0.571429	-13.7425	source-language | have
0.35	-13.7425	should | contain
0.140741	-13.7425	different | use
0.2	-13.7425	is | belongs
0.388889	-13.7427	of | phenomenon.
0.180723	-13.7427	token | with
0.180723	-13.7427	for | estimation
0.4	-13.7428	messages | other
0.4	-13.7428	extraction | (e.g.
0.0808824	-13.7429	does not | and
0.191781	-13.7429	all | trees
0.75	-13.743	constraints | ∀i
0.263158	-13.743	users | at
0.368421	-13.743	lexical | binary
0.368421	-13.743	extracted | previous
0.203125	-13.7431	says | the
0.25	-13.7432	systems | data.
0.571429	-13.7432	cognate | cognate
0.666667	-13.7433	joint | along
0.666667	-13.7433	classifying | articles
0.19697	-13.7436	collaborative | and
0.189189	-13.7436	for | ordering
0.4	-13.7437	language | score.
0.25641	-13.7438	that | somewhat
0.75	-13.7439	representation, | defined
0.060241	-13.7439	of | instances
0.307692	-13.744	when | mentions
0.666667	-13.7441	ten | them
0.189189	-13.7442	we | beam
0.137681	-13.7442	if | given
0.141791	-13.7443	as | subject
0.145161	-13.7444	number | correct
0.5	-13.7444	included | had
0.161616	-13.7445	our | achieve
0.0394061	-13.7445	the | pair
0.428571	-13.7446	but, | not
0.571429	-13.7446	inherits | from
0.35	-13.7446	first | character
0.0398126	-13.7447	for | model
0.226415	-13.7447	some | might
0.112676	-13.7447	we found | to
0.120879	-13.7448	system | features
0.135135	-13.7448	np | jj
0.0676819	-13.7448	to | dialogue
0.666667	-13.7449	increase | (by
0.0718336	-13.7449	another | to
0.112676	-13.745	special | and
0.157407	-13.7451	pair | an
0.32	-13.7451	score | translation.
0.222222	-13.7452	texts | other
0.3	-13.7453	properties | more
0.0892857	-13.7453	the | unknown
0.571429	-13.7453	annotations | checked
0.571429	-13.7453	action | estimated
0.35	-13.7455	performed | well
0.218182	-13.7455	much | but
0.5	-13.7456	translation | forward
0.666667	-13.7457	accuracy | classifiers.
0.75	-13.7457	p(w|a)
0.178571	-13.746	two | either
0.130719	-13.746	this, | to
0.571429	-13.746	.76 | -
0.571429	-13.746	mt, | has
0.571429	-13.746	model | mt04
0.0743802	-13.7463	to | sentence.
0.164948	-13.7465	∈ | where
0.180723	-13.7466	demonstrate | our
0.368421	-13.7466	rank | given
0.428571	-13.7467	on | categories,
0.27027	-13.7468	thresholds | to
0.108225	-13.7468	system | was
0.333333	-13.7469	to | writing.
0.186667	-13.7469	outperformed | the
0.307692	-13.747	assume | document
0.368421	-13.747	these | below.
0.064915	-13.7471	target | is
0.4	-13.7471	quantify | on
0.0503262	-13.7471	the | c
0.0604027	-13.7473	from | set
0.210526	-13.7473	as | classifier.
0.388889	-13.7475	their | unsupervised
0.3	-13.7475	with | allowed
0.75	-13.7475	process | constructing
0.0914826	-13.7475	for | less
0.189189	-13.7475	case | only
0.0932476	-13.7476	in | random
0.28125	-13.7476	fully | model
0.35	-13.7476	it | ten
0.0651828	-13.7477	topic | a
0.125749	-13.7478	graphs | and
0.121547	-13.7479	candidate | with
0.244444	-13.748	our | efficient
0.214286	-13.748	and | roughly
0.25	-13.7481	baseline | three
0.122222	-13.7481	compared | that
0.571429	-13.7481	metrics: | word
0.307692	-13.7482	applied | extraction
0.184211	-13.7482	problem | our
0.32	-13.7483	hypotheses | word
0.127273	-13.7484	1) | to
0.155963	-13.7484	which | trained
0.176471	-13.7485	which | sentences.
0.122222	-13.7487	the | defines
0.428571	-13.7488	achieve | similar
0.571429	-13.7488	2002; | mckeown,
0.244444	-13.7489	or | information.
0.666667	-13.749	d. | j.
0.666667	-13.749	actions | associated
0.666667	-13.749	comments | into
0.206349	-13.7491	model | on | data.
0.151786	-13.7491	solution | is
0.0535152	-13.7491	feature | for
0.0790698	-13.7491	works | the
0.166667	-13.7492	seems | that
0.307692	-13.7493	achieves | compared
0.206349	-13.7493	this | context.
0.194444	-13.7493	can then | to
0.35	-13.7493	smt | moses
0.172414	-13.7494	system | lexical
0.153153	-13.7494	model | error
0.160377	-13.7494	resource | for
0.184211	-13.7496	top | words
0.4	-13.7496	21 | 20
0.4	-13.7496	from | forest.
0.194444	-13.7496	baseline | both
0.28125	-13.7497	most | instances
0.0245684	-13.7498	it is | the
0.218182	-13.75	neighborhood | the
0.3	-13.7501	given | fact
0.112676	-13.7501	is | instead
0.571429	-13.7502	component | automatically
0.571429	-13.7502	extends | dependency
0.75	-13.7503	find | anaphoric
0.75	-13.7503	wider | knowledge
0.142857	-13.7504	α | to
0.170455	-13.7505	discussion | on
0.666667	-13.7506	approach | formulation
0.191781	-13.7506	trained | sentences
0.35	-13.7506	posterior | model
0.388889	-13.7508	values | parameter
0.226415	-13.7508	standard | words
0.151786	-13.7508	language | speech
0.454545	-13.7509	0.75 | 0.72
0.184211	-13.7509	and | 2001),
0.428571	-13.7509	of | advance.
0.571429	-13.7509	matches any | property
0.4	-13.751	finally, | achieved
0.263158	-13.751	1 | ...
0.155963	-13.7511	random | on
0.307692	-13.7512	was | it.
0.75	-13.7512	their | difficult.
0.25	-13.7512	unlike | it
0.23913	-13.7513	answering | et
0.666667	-13.7514	example | cancer
0.135135	-13.7514	d | c
0.35	-13.7514	remaining | an
0.104418	-13.7515	using | system
0.4	-13.7515	nnp | .
0.4	-13.7515	method | appropriate
0.0707635	-13.7515	create | the
0.2	-13.7515	event | role
0.15	-13.7517	significantly | on
0.5	-13.7517	cluster | contexts
0.290323	-13.7518	branches | the
0.454545	-13.752	thread | post
0.0887574	-13.752	as | syntactic
0.25641	-13.7522	system | performs
0.0707635	-13.7522	implementation | the
0.333333	-13.7522	paid | to
0.214286	-13.7522	space | not
0.0538793	-13.7523	any | in
0.571429	-13.7523	thus, | extracted
0.571429	-13.7523	several | 1994;
0.0875706	-13.7524	which | data
0.137681	-13.7524	risk | the
0.12987	-13.7525	apply | for
0.368421	-13.7525	in | round
0.333333	-13.7526	stateof-the-art | et
0.0818859	-13.7526	in | 5.
0.32	-13.7526	though | was
0.0156845	-13.7526	this | to
0.25	-13.7527	make | text
0.218182	-13.7527	this | motivated
0.184211	-13.753	items | by
0.75	-13.753	under | miprcv
0.75	-13.753	values | inverse
0.15	-13.753	the | margin
0.234043	-13.7532	improvement | significant
0.044235	-13.7532	source | and
0.106383	-13.7533	and | concepts
0.107759	-13.7534	are | three
0.428571	-13.7535	0.84 | 0.60
0.244444	-13.7537	ball | the
0.2	-13.7537	media | a
0.234043	-13.7538	defining | as
0.307692	-13.7538	probability | grammar
0.333333	-13.7538	locations | are
0.368421	-13.7538	methods | important
0.203125	-13.7538	2 | english
0.75	-13.7539	generate | queries.
0.191781	-13.7539	4 | has
0.13245	-13.7539	section 2 | and
0.244444	-13.754	to | generation.
0.4	-13.754	papers | used
0.0992647	-13.7541	is | task.
0.184211	-13.7541	and | measures.
0.118919	-13.7541	work | learning
0.5	-13.7542	ranking | four
0.5	-13.7542	(2011), | that
0.0551181	-13.7543	of | original
0.0616246	-13.7543	system | by
0.218182	-13.7543	an | language.
0.214286	-13.7543	each | count
0.180723	-13.7544	learning | new
0.571429	-13.7544	(charniak, | parser
0.35	-13.7544	allows | multiple
0.104	-13.7545	probabilities | a
0.114428	-13.7545	took | the
0.229167	-13.7546	thus | have
0.0765864	-13.7546	c | for
0.666667	-13.7547	language | impaired
0.141791	-13.7547	as | events
0.134228	-13.7547	framework | that
0.25	-13.7548	goal | such
0.111111	-13.7549	its | which
0.32	-13.7549	ti | is
0.0745342	-13.755	bilingual | and
0.101562	-13.7552	feature | or
0.191781	-13.7553	in | uniform
0.35	-13.7553	in | but,
0.0882353	-13.7553	r | p
0.333333	-13.7554	improvement of | baseline
0.128049	-13.7554	table | show
0.244444	-13.7555	verb | main
0.0683012	-13.7555	our | word
0.180723	-13.7557	to | criteria
0.571429	-13.7558	time | fast
0.571429	-13.7558	swsd | target
0.0571081	-13.756	are | different
0.234043	-13.7561	used | reordering
0.307692	-13.7561	however, | complexity
0.0831169	-13.7561	of | clustering
0.35	-13.7562	when | fixed
0.127273	-13.7562	process | with
0.19697	-13.7563	following | were
0.158879	-13.7563	this | semantics
0.203125	-13.7563	become | for
0.122905	-13.7563	we | state-of-the-art
0.122905	-13.7563	aspect | and
0.307692	-13.7564	sentences | produce
0.454545	-13.7565	construct | rule
0.571429	-13.7565	semantic | objects,
0.75	-13.7567	policies | trained
0.333333	-13.7567	the | alphabet.
0.134228	-13.7568	this | existing
0.118919	-13.7569	projected | the
0.368421	-13.757	part | (pos)
0.454545	-13.7571	condition | no
0.0955631	-13.7571	capture | and
0.0861111	-13.7571	any | be
0.428571	-13.7572	expressed | text
0.5	-13.7572	which | filled
0.5	-13.7572	this | undesirable
0.3	-13.7574	parsing, | can
0.1133	-13.7574	one | different
0.263158	-13.7575	examples | given
0.0796253	-13.7576	the | expressed
0.272727	-13.7577	experimental | effectiveness
0.117347	-13.7577	clause | the
0.0622287	-13.7577	they | as
0.128049	-13.7579	and | 2003),
0.5	-13.7579	lower | expected
0.19697	-13.758	agreement | between
0.234043	-13.7581	the | studies.
0.0945946	-13.7582	iterations | the
0.25	-13.7582	adds | in
0.0877193	-13.7583	however, | by
0.0955631	-13.7584	requires | and
0.244444	-13.7585	message | by
0.25641	-13.7585	when | generated
0.25641	-13.7585	primeprime | prime
0.229167	-13.7585	requires | not
0.115578	-13.7585	we | paraphrase
0.32	-13.7585	because | languages
0.5	-13.7585	existing | sentiment
0.0714286	-13.7586	to | additional
0.28125	-13.7586	contribution | overall
0.571429	-13.7586	composition | sets
0.218182	-13.7586	words | morphological
0.25641	-13.7588	to | long-distance
0.428571	-13.7588	illustrates | our
0.130719	-13.7588	such | only
0.4	-13.7589	phrase | embedded
0.0945946	-13.759	for | grammar
0.333333	-13.7591	35 | 50
0.5	-13.7591	decoding | mbr
0.116751	-13.7592	description | a
0.571429	-13.7593	many | indicate
0.571429	-13.7593	vectors | acceptable
0.4	-13.7594	sprime | that
0.307692	-13.7595	type | similar
0.75	-13.7595	image | vectors
0.102362	-13.7597	the | detailed
0.5	-13.7597	baseline | oracle
0.428571	-13.7598	semantic | selectional
0.109589	-13.7599	was | with | in
0.0630631	-13.7599	simple | to
0.35	-13.76	features, | sentence
0.0794393	-13.7601	in | .
0.23913	-13.7601	are | hard
0.229167	-13.7602	models | random
0.101167	-13.7603	incorrect | the
0.666667	-13.7604	model | aligner.
0.666667	-13.7604	ro | sl
0.4	-13.7604	thesis, | of
0.0684211	-13.7605	often | and
0.0635401	-13.7605	following | is
0.11828	-13.7605	is | few
0.0992647	-13.7607	if | its
0.571429	-13.7607	with | 0.2.
0.571429	-13.7607	working | improve
0.102767	-13.7608	to | aligned
0.666667	-13.7612	required | generation
0.0711538	-13.7612	extract | and
0.120219	-13.7612	for | containing
0.14876	-13.7613	an | like
0.571429	-13.7614	emission | both
0.454545	-13.7615	given | evidence
0.25	-13.7616	the | infrastructure
0.333333	-13.7616	(miller | al.,
0.234043	-13.7618	modified | as
0.4	-13.7619	val | 3
0.154545	-13.762	the | reader
0.454545	-13.7621	find | clusters
0.454545	-13.7621	systems | available.
0.666667	-13.7621	conducted | subjects
0.666667	-13.7621	had | indicating
0.571429	-13.7621	need | efficient
0.571429	-13.7621	seminal | that
0.75	-13.7622	3.1 | (2010).
0.75	-13.7622	propagate | across
0.181818	-13.7623	presented | model
0.0920635	-13.7624	cases | a
0.19697	-13.7624	co-segmentation | the
0.28125	-13.7624	for | predicate.
0.0877193	-13.7624	category | and
0.229167	-13.7625	shows | while
0.139706	-13.7626	search | on
0.0507177	-13.7627	sentences | in
0.050237	-13.7628	model | by
0.229167	-13.7628	data | made
0.28125	-13.7628	average | 1
0.154545	-13.7629	with | verbs
0.139706	-13.7629	where | over
0.218182	-13.7629	tweets | or
0.155963	-13.7629	in | parses
0.428571	-13.763	papers | this
0.428571	-13.763	rule | pattern
0.263158	-13.763	an | e.g.
0.0759219	-13.763	10 | and
0.75	-13.7632	5 | sim
0.75	-13.7632	situation | denoting
0.75	-13.7632	wizard | rl-based
0.072549	-13.7632	of | term
0.103734	-13.7632	this | out
0.168539	-13.7633	in | pipeline
0.146341	-13.7633	discuss | for
0.368421	-13.7633	sparse | by
0.122807	-13.7633	grammar | with
0.106838	-13.7633	5 | are
0.0942761	-13.7634	algorithm | by
0.111628	-13.7635	to | match
0.0559441	-13.7635	a | event
0.137681	-13.7635	while | all
0.35	-13.7635	value | concept
0.184211	-13.7636	meanings | in
0.428571	-13.7636	have | varying
0.229167	-13.7636	split | by
0.307692	-13.7637	changes | word
0.666667	-13.7637	domain | independent
0.091195	-13.7638	result | is
0.454545	-13.7638	by | gift
0.191781	-13.7638	predicts | of
0.19697	-13.7639	data | examples
0.103734	-13.7639	we | classes
0.2	-13.764	initial | using
0.13245	-13.764	of | training.
0.244444	-13.7642	approach | topic
0.222222	-13.7642	this | generally
0.0584498	-13.7642	np | the
0.571429	-13.7642	them | framework.
0.454545	-13.7644	see | right
0.454545	-13.7644	antecedent | candidate
0.454545	-13.7644	(marton | 2008; | al.,
0.35	-13.7644	β, | is
0.174419	-13.7645	which | state
0.333333	-13.7645	potential | have
0.0897833	-13.7646	the | dependent
0.144	-13.7646	property | and
0.109091	-13.7647	for | tag
0.368421	-13.7647	university | tokyo
0.368421	-13.7647	recall | 0.83
0.5	-13.7647	and | t.j.
0.203125	-13.7648	technical | and
0.0879765	-13.7649	dictionary | of
0.263158	-13.7649	first, | number
0.210526	-13.7649	its | given
0.75	-13.765	usually | expressed
0.75	-13.765	mixture | towards
0.75	-13.765	160 | 100
0.75	-13.765	concepts | limited
0.25	-13.765	our | i.e.,
0.571429	-13.765	could | place
0.121547	-13.7652	choose | to
0.0856354	-13.7652	initial | to
0.170455	-13.7652	similar | words
0.4	-13.7653	in addition, | how
0.174419	-13.7653	c | set
0.22449	-13.7654	task, | which
0.19403	-13.7655	this | advantage
0.28125	-13.7656	use | cosine
0.428571	-13.7657	zero, | is
0.172414	-13.7658	an | must
0.290323	-13.7658	20 | were
0.106383	-13.7658	equal | of
0.0340017	-13.7658	features | in
0.122093	-13.7659	are | labeled
0.75	-13.766	was | thoroughly
0.75	-13.766	occurrence | patterns.
0.189189	-13.7661	the | penalizes
0.109589	-13.7661	sense | as
0.0727273	-13.7661	the | participants
0.0730223	-13.7662	semantics | the
0.0960854	-13.7662	is | content
0.0562874	-13.7663	we also | and
0.25	-13.7663	accuracy | recall
0.0592105	-13.7664	with | at
0.0447309	-13.7664	information | a
0.0644654	-13.7665	distribution | a
0.5	-13.7665	different | gene
0.454545	-13.7666	frequently | likely
0.28125	-13.7666	shown | useful
0.307692	-13.7667	because | relatively
0.100775	-13.7667	scores | by
0.145161	-13.7667	word | by | word
0.146341	-13.7667	parallel | corpus
0.75	-13.7669	adjacency | information
0.129032	-13.7669	is | longer
0.666667	-13.767	ex | expectations
0.666667	-13.767	annotation | address
0.19403	-13.767	is | encoded
0.307692	-13.7671	less | per
0.571429	-13.7671	occurrences | occurrence
0.121547	-13.7672	experiments | used
0.121547	-13.7672	the | biased
0.243902	-13.7673	7 conclusions | and
0.0900621	-13.7674	in | supervised
0.111111	-13.7674	set | different
0.0469522	-13.7675	models | a
0.090625	-13.7675	that | important
0.206897	-13.7675	table | reports
0.32	-13.7676	translation | derivations
0.454545	-13.7677	they | complementary
0.033833	-13.7677	of | information
0.666667	-13.7678	word | chosen.
0.0879765	-13.7678	the | denotes
0.428571	-13.7678	realization | that
0.176471	-13.7678	next | with
0.118919	-13.7678	results | the | baseline
0.0734694	-13.7679	the | aspects
0.0914826	-13.768	example, | or
0.1	-13.7682	features | this
0.571429	-13.7685	measure | dialog
0.571429	-13.7685	model | 80.6
0.243902	-13.7686	machine | scores
0.214286	-13.7686	more | grammar
0.75	-13.7688	100 | judged
0.4	-13.7688	entity | sentence.
0.454545	-13.7689	improvements | better
0.181818	-13.7689	algorithm, | is
0.126506	-13.769	of | t.
0.174419	-13.769	such | large
0.23913	-13.769	obtain | word
0.0725806	-13.7691	in | candidate
0.290323	-13.7691	with | (cf.
0.189189	-13.7691	∈ | set
0.368421	-13.7693	less | since
0.191176	-13.7694	rules | into
0.116402	-13.7694	and | problems
0.333333	-13.7695	forms | words
0.101167	-13.7696	we | experiments.
0.75	-13.7697	classification, | etc.
0.75	-13.7697	need | translations,
0.0981818	-13.7698	the | suggests
0.428571	-13.7699	σ | r
0.333333	-13.7699	subject | other
0.0867052	-13.7699	created | and
0.454545	-13.77	within | project
0.210526	-13.7701	and | ace
0.222222	-13.7703	model | matrix
0.428571	-13.7705	much | size
0.35	-13.7705	variational | et
0.0872093	-13.7705	the | become
0.75	-13.7706	tool | gender
0.333333	-13.7707	part-of-speech | information
0.22449	-13.7707	2.2 | are
0.189189	-13.7708	other | results
0.28125	-13.7708	and | yi
0.135714	-13.7709	to | etc.
0.5	-13.7709	each | symmetric
0.163265	-13.771	(2009), | and
0.428571	-13.771	draw | our
0.12426	-13.7711	we | edges
0.0909091	-13.7711	for all | in
0.333333	-13.7711	bayesian | using
0.19403	-13.7711	information | relations
0.28125	-13.7712	the | noisy.
0.229167	-13.7713	automatically | corpus
0.0806846	-13.7713	there | two
0.222222	-13.7714	maximum | all
0.111111	-13.7714	we | alignments
0.203125	-13.7714	which | frequent
0.22449	-13.7715	always | or
0.5	-13.7715	detecting | opinion
0.428571	-13.7716	iterations | gibbs
0.203125	-13.7716	is | named entity
0.307692	-13.7717	annotation | type
0.154545	-13.7718	this | performed
0.210526	-13.772	they | high
0.307692	-13.7721	ge | with
0.0826873	-13.7721	followed | the
0.571429	-13.7721	percentage | pairs.
0.0491356	-13.7722	common | the
0.040293	-13.7726	all | for
0.290323	-13.7727	output | then
0.0489574	-13.7727	a | may
0.666667	-13.7728	conv | uuniondisplay
0.666667	-13.7728	za | æ
0.666667	-13.7728	premodifier | possessive
0.666667	-13.7728	categorization | replaces
0.666667	-13.7728	“look | prp”
0.666667	-13.7728	noise | φ(x,y).
0.666667	-13.7728	1+2 | 3+4+5
0.666667	-13.7728	hay | ✇
0.666667	-13.7728	deaf | hearing
0.666667	-13.7728	instructor | interactions
0.666667	-13.7728	∀j | ∈s⇒(lj
0.666667	-13.7728	revisions | revision
0.666667	-13.7728	pijsummationdisplay | p(β0)
0.666667	-13.7728	(lafferty, | pereira
0.666667	-13.7728	scale; | raters
0.666667	-13.7728	angbracketleftbiggbracketleftbig | bracketrightbigangbracketrightbigg
0.666667	-13.7728	include | predicted
0.666667	-13.7728	song | dance
0.666667	-13.7728	朋友 | 伙伴
0.666667	-13.7728	<theme | stat3
0.666667	-13.7728	d103 | d100
0.666667	-13.7728	lexicon | roget
0.666667	-13.7728	loc | tmp
0.666667	-13.7728	cit´e, | france
0.666667	-13.7728	0.63 | 0.69
0.666667	-13.7728	np-sbj | vp
0.666667	-13.7728	.92 | .14
0.666667	-13.7728	ss | india
0.666667	-13.7728	maosong | tsinghua
0.666667	-13.7728	(gap+) | (male)
0.666667	-13.7728	|df|
0.666667	-13.7728	0.03 | 0.57
0.666667	-13.7728	[2009] | p(θ,y,b|x,x
0.666667	-13.7728	〈w1,wprime1〉 | 〈w2,wprime2〉
0.229167	-13.7728	it | length
0.0600273	-13.7729	a | so
0.35	-13.7731	studies | speech
0.428571	-13.7732	(2010), | it
0.0711538	-13.7733	usually | the
0.146341	-13.7733	state | by
0.2	-13.7733	1 | |
0.100775	-13.7734	lists | of
0.0593792	-13.7734	that | most
0.368421	-13.7734	its | still
0.571429	-13.7735	average | (real)
0.153153	-13.7735	which | relations
0.125749	-13.7736	which | context
0.186667	-13.7739	source | alignment
0.0188704	-13.7739	the | set | the
0.104167	-13.774	linear | is
0.153153	-13.774	u | =
0.5	-13.774	post | user
0.125749	-13.7741	5 | using
0.0789474	-13.7741	systems | al.,
0.263158	-13.7741	set | count
0.0933333	-13.7742	weight | to
0.4	-13.7743	approaches | 2008)
0.35	-13.7744	when | long
0.35	-13.7744	another | direction
0.454545	-13.7745	gap | position
0.174419	-13.7745	(3) | with
0.0808081	-13.7746	term | and
0.28125	-13.7747	defined | following
0.112195	-13.7747	= | :
0.32	-13.7748	can | reasonable
0.234043	-13.7749	from | scoring
0.0680628	-13.7749	a | output
0.128205	-13.775	system | better
0.454545	-13.7751	dependencies | stanford
0.0502415	-13.7752	a | dependency
0.75	-13.7753	discuss | limitations
0.5	-13.7753	context | space.
0.296296	-13.7754	combination | systems.
0.0833333	-13.7755	introduce | of
0.166667	-13.7756	the | supports
0.135714	-13.7756	on | length
0.571429	-13.7756	train | reviews
0.0930233	-13.7757	are | many
0.0613409	-13.7757	1 | 4
0.333333	-13.7757	intelligence | and
0.333333	-13.7757	figure 1: | annotation
0.218182	-13.7757	note that | these
0.203125	-13.7757	has | various
0.161616	-13.7758	on | existing
0.191176	-13.7759	100 | from
0.290323	-13.7759	and | 2005,
0.243902	-13.7759	model | represents
0.0793269	-13.776	features | word
0.133803	-13.776	as | common
0.0721443	-13.7761	out | in
0.35	-13.7762	in | 1997)
0.203125	-13.7762	26 | and
0.234043	-13.7764	while | (e.g.,
0.0623145	-13.7764	expected | the
0.174419	-13.7764	she | a
0.144	-13.7765	a | marked
0.23913	-13.7765	its | tree
0.5	-13.7765	some | knowledge,
0.191176	-13.7768	transitions | and
0.191176	-13.7768	from | had
0.145161	-13.7768	as | head
0.146341	-13.7768	we | variable
0.144	-13.7768	4 | training
0.25641	-13.777	patterns | their
0.210526	-13.777	that, | can
0.117647	-13.7771	the | 13
0.206897	-13.7771	our | extracting
0.75	-13.7772	pattern | correcting
0.0423497	-13.7773	main | the
0.0964286	-13.7774	uk, | 2011.
0.0702087	-13.7774	in | proposed
0.142857	-13.7774	few | for
0.0513078	-13.7774	set | as
0.333333	-13.7774	uniform | =
0.25	-13.7775	data-driven | in
0.35	-13.7775	most | questions
0.35	-13.7775	contains | pair
0.164948	-13.7775	that, | be
0.203125	-13.7775	the | direction,
0.296296	-13.7776	 | 
0.191176	-13.7778	we | sms
0.214286	-13.7779	most | related
0.454545	-13.778	such | readability
0.454545	-13.778	representation | contains
0.0601093	-13.778	a | s | a
0.23913	-13.778	red | to
0.307692	-13.7782	generate | pairs
0.1	-13.7782	of | structure.
0.28125	-13.7782	the | navigli,
0.133803	-13.7783	all | do
0.4	-13.7783	mapping | logical
0.0757576	-13.7783	the | short
0.116402	-13.7784	a | simply
0.35	-13.7784	create | pairs
0.5	-13.7784	not | stemming
0.454545	-13.7785	extension | algorithm
0.428571	-13.7785	a | algorithm | linear
0.571429	-13.7785	from | impasse
0.368421	-13.7785	identifier | a
0.0795181	-13.7786	and | present
0.149123	-13.7786	any | =
0.1133	-13.7787	are | part
0.218182	-13.7787	has | value
0.243902	-13.7789	ongoing | to
0.23913	-13.7789	baseline | without
0.454545	-13.7791	update | probability
0.75	-13.7791	dependency | protein-protein
0.428571	-13.7791	model | globally
0.0996169	-13.7791	hypothesis | in
0.290323	-13.7792	annotators | but
0.170455	-13.7793	include | such
0.28125	-13.7793	completion | a
0.307692	-13.7794	thus | 1
0.036716	-13.7794	is | at
0.243902	-13.7795	the | 3.3).
0.290323	-13.7796	including | but
0.132867	-13.7796	paraphrases | in
0.32	-13.7796	to | weight.
0.307692	-13.7798	(pp | (np
0.177215	-13.7798	for | logical
0.4	-13.7798	performance | affected
0.368421	-13.7799	initial | learning
0.0780142	-13.78	this | performance
0.75	-13.78	development | accurate
0.28125	-13.78	and | position.
0.181818	-13.78	of | dt
0.101562	-13.7801	ran | the
0.428571	-13.7801	use | paraphrasing
0.263158	-13.7801	these | larger
0.35	-13.7801	here | very
0.110577	-13.7801	each | probability
0.206897	-13.7802	when | document
0.454545	-13.7803	all | (this
0.368421	-13.7804	how | get
0.214286	-13.7804	our | basic
0.234043	-13.7805	in | causes
0.296296	-13.7806	ml | in
0.35	-13.7806	of | non-empty
0.0921053	-13.7806	the | absolute
0.172414	-13.7808	from | university
0.0694698	-13.7808	this paper | in
0.145161	-13.7808	the | marks
0.0379747	-13.7808	which | be
0.210526	-13.7809	lower | which
0.296296	-13.781	two | components
0.35	-13.781	 | if
0.181818	-13.781	entailment | al.,
0.218182	-13.7812	of | reasonably
0.0722892	-13.7813	to | general
0.4	-13.7813	constructions | not
0.454545	-13.7814	noisy-and | model
0.296296	-13.7814	grammar | dependency
0.28125	-13.7814	as | derivations
0.141732	-13.7814	weighted | is
0.571429	-13.7814	teacher | whether
0.2	-13.7815	by | method.
0.0656814	-13.7815	in | work,
0.5	-13.7815	is, | find
0.126582	-13.7816	may | but
0.0652529	-13.7816	coreference | the
0.104603	-13.7816	even | not
0.153153	-13.7817	models | both
0.0392962	-13.7817	the | terms | the
0.135714	-13.7818	as | languages
0.0576184	-13.7818	of | parsing
0.75	-13.7819	summarization | trans-temporal
0.75	-13.7819	0.72 | 0.61
0.75	-13.7819	organizers | evaluation
0.263158	-13.782	benefit | by
0.263158	-13.782	source | translation.
0.191176	-13.7821	used | without
0.139535	-13.7821	words | also
0.101626	-13.7822	point | in
0.243902	-13.7823	mst | malt
0.4	-13.7824	1998; | 2002;
0.4	-13.7824	most | e.g.
0.454545	-13.7825	instead | fixed
0.28125	-13.7825	responses | score
0.164948	-13.7825	2004; | et | 2005;
0.158416	-13.7825	the | suffixes
0.210526	-13.7825	for | pairwise
0.036864	-13.7827	this | a | for
0.333333	-13.7828	only | users
0.571429	-13.7828	queries | find
0.147826	-13.7828	this | sentences.
0.5	-13.7828	for | hearst,
0.163265	-13.7829	all | accuracy
0.75	-13.7829	terminates | no
0.168539	-13.7829	word | related
0.25641	-13.7831	method | search
0.272727	-13.7832	approach | linear
0.116402	-13.7834	of | provided by
0.128205	-13.7834	5 | 8
0.28125	-13.7835	most | machine translation
0.12987	-13.7835	the authors | to
0.75	-13.7838	shared | (hajiˇc
0.12	-13.7839	corpora | with
0.272727	-13.7839	senses | their
0.4	-13.7839	sentence | weighted
0.296296	-13.784	consists | about
0.142857	-13.7841	results | 3.
0.22449	-13.7842	make | system
0.136691	-13.7842	models | use
0.19403	-13.7842	also | related
0.156863	-13.7843	process | each
0.571429	-13.7843	aware | fact
0.0915033	-13.7845	given | can
0.243902	-13.7845	separate | are
0.272727	-13.7846	were | evaluated
0.234043	-13.7847	once | it
0.75	-13.7848	survival | effects
0.218182	-13.7848	but | how
0.0599721	-13.785	applied | and
0.428571	-13.785	necessary | order
0.368421	-13.785	has | counts
0.0501931	-13.7851	and | text
0.11165	-13.7851	components | and
0.188406	-13.7852	syntactic | parse
0.039356	-13.7852	each | with
0.110577	-13.7852	= | w
0.137681	-13.7853	current | as
0.25641	-13.7854	sentences | topic
0.142857	-13.7854	an | noun
0.153153	-13.7854	to | ambiguity
0.125	-13.7855	following | which
0.20339	-13.7856	synsets | a
0.0903427	-13.7856	core | the
0.181818	-13.7856	this | complete
0.25641	-13.7857	using | results.
0.0789474	-13.7858	unsupervised | to
0.0778302	-13.7859	first, | to
0.100386	-13.7859	of | identifying
0.243902	-13.786	training | gold
0.210526	-13.786	models | improve
0.5	-13.786	parser | uas
0.5	-13.786	have | disease
0.133803	-13.7861	selection | as
0.0673575	-13.7861	data, | of
0.0728745	-13.7862	the | 5.
0.0448223	-13.7863	has | for
0.0717131	-13.7863	approach | with
0.229167	-13.7863	aligned | which
0.0516529	-13.7863	event | and
0.234043	-13.7864	the | paraphrase.
0.307692	-13.7864	set | 30
0.129032	-13.7864	actually | to
0.571429	-13.7864	aligned | potential
0.571429	-13.7864	sequence, | each
0.106195	-13.7865	but | our
0.454545	-13.7866	were | experiment.
0.170455	-13.7866	working | and
0.75	-13.7867	observations | similar
0.35	-13.7867	unique | set
0.12	-13.7868	the | in-domain
0.0295289	-13.7868	our | that
0.0909091	-13.7868	it | into
0.108108	-13.7868	dialogue | that
0.290323	-13.7869	(2009) | translation
0.229167	-13.7869	the | on-line
0.368421	-13.7869	punctuation | all
0.25	-13.787	bilingual | alignment
0.571429	-13.7872	¯em | in
0.571429	-13.7872	output | 1-best
0.571429	-13.7872	grammars | context-free
0.571429	-13.7872	summationdisplay | msummationdisplay
0.046473	-13.7873	the | general
0.112745	-13.7873	sampling | and
0.0795181	-13.7874	is | length
0.296296	-13.7874	suggesting that | be
0.0933333	-13.7876	the | correction
0.75	-13.7876	from | completely.
0.140625	-13.7877	and | affect
0.150442	-13.7877	decided | to
0.25	-13.7879	words | important
0.571429	-13.7879	compliance | data
0.571429	-13.7879	(judge | al.,
0.571429	-13.7879	part | gigaword
0.20339	-13.788	words, | more
0.133803	-13.7881	citation | the
0.22	-13.7881	probability | model.
0.135714	-13.7883	our | significant
0.168539	-13.7885	show | outperforms
0.037234	-13.7885	however, | to
0.75	-13.7886	2009) | require
0.571429	-13.7886	results | arabic,
0.218182	-13.7886	trend | and
0.272727	-13.7887	patterns | they
0.129032	-13.7887	is | close
0.170455	-13.7888	number | examples
0.4	-13.7889	constraints | hierarchical
0.4	-13.7889	tools | annotation
0.296296	-13.7889	shown | tasks
0.272727	-13.7891	of | lexically
0.5	-13.7891	algorithm | parses
0.0736842	-13.7892	and | much
0.25	-13.7892	plan to | more
0.333333	-13.7892	use | automated
0.368421	-13.7892	text | modeling
0.0294604	-13.7893	in | and | are
0.571429	-13.7893	original | 2007).
0.571429	-13.7893	will | srl
0.243902	-13.7894	each | until
0.0909091	-13.7894	are | likely
0.35	-13.7894	7 | experiments
0.75	-13.7895	first, | dialogues
0.75	-13.7895	scheme | marneffe
0.375	-13.7895	the | 2.4.
0.146552	-13.7896	and | sets,
0.32	-13.7897	various | other
0.5	-13.7898	17 | 12
0.214286	-13.7899	and | members
0.454545	-13.79	and | bins.
0.210526	-13.79	all | data,
0.22449	-13.7902	researchers | on
0.35	-13.7903	summarization | 2010).
0.333333	-13.7904	with | networks
0.19403	-13.7904	main contribution | the
0.0512048	-13.7905	previous | of
0.179487	-13.7908	fuzzy | and
0.12426	-13.7909	constituents | the
0.102459	-13.7912	and | above
0.158416	-13.7912	in | translate
0.264706	-13.7912	include, | the
0.229167	-13.7913	by | darpa
0.163265	-13.7914	as | cannot
0.75	-13.7914	build | characters.
0.75	-13.7914	agreement | weak
0.307692	-13.7915	color | to
0.428571	-13.7915	lot | information
0.571429	-13.7915	are | italics.
0.229167	-13.7916	the | runs.
0.181818	-13.7916	with | already
0.177215	-13.7917	and | sometimes
0.571429	-13.7923	was | evaluations.
0.375	-13.7923	possible | errors
0.75	-13.7924	gale | hr0011-06-c-0022.
0.35	-13.7925	models | clusters
0.368421	-13.7925	doing | be
0.32	-13.7926	vertex | null
0.170455	-13.7927	is | unit
0.375	-13.7928	mean | t-units
0.214286	-13.7929	in | women
0.151786	-13.793	applications | as
0.105263	-13.7931	character | of
0.290323	-13.7932	(5k) | (10k)
0.234043	-13.7933	behavior | by
0.75	-13.7933	adopt | term
0.22	-13.7934	alternatively, | of
0.454545	-13.7935	雨 | 滑
0.4	-13.7935	use | heuristics
0.28125	-13.7935	text | so
0.238095	-13.7936	and | equivalence
0.141732	-13.7936	whether | word
0.5	-13.7936	2.2 | research
0.161616	-13.7938	as | directly
0.131944	-13.7939	aim | a
0.0364103	-13.7939	we | a | from
0.28125	-13.7939	log | ,
0.0305206	-13.794	given | of
0.0724346	-13.7941	agreement | and
0.0327122	-13.7941	on | with
0.0842697	-13.7941	the | with | corresponding
0.229167	-13.7942	such | web
0.5	-13.7942	linguist | and
0.75	-13.7943	highest | chosen.
0.75	-13.7943	user: | suche
0.75	-13.7943	also | domingos,
0.75	-13.7943	sets | ontologies
0.104167	-13.7944	as | result
0.0328346	-13.7944	use | in
0.0525188	-13.7944	algorithm | in
0.136691	-13.7946	in | 2).
0.454545	-13.7947	head | arg2
0.333333	-13.7947	sentences | dataset
0.375	-13.7947	annotation | evaluate
0.126582	-13.7949	results | method
0.25641	-13.7949	is, | it
0.0322061	-13.7949	the | described
0.0884146	-13.7949	k | a
0.307692	-13.795	gives | when
0.146552	-13.795	to | boundaries
0.0944056	-13.7953	a | processing
0.333333	-13.7955	target | additional
0.119318	-13.7956	applies | the
0.172414	-13.7957	to | precision.
0.16	-13.7957	lda | for
0.375	-13.7957	represented | sentences
0.0981132	-13.7958	is, | a
0.571429	-13.7959	resulting | after
0.0655462	-13.796	should | a
0.0812183	-13.796	vectors | and
0.0831099	-13.7961	meeting | association for computational
0.28125	-13.7961	we | end-to-end
0.28125	-13.7961	k | +
0.0771028	-13.7961	shared | and
0.214286	-13.7961	table 2: | from
0.75	-13.7962	head-driven | grammar
0.454545	-13.7964	(r) | f-measure
0.177215	-13.7964	of | pattern.
0.22	-13.7965	limits | of
0.0984848	-13.7966	similar | by
0.116402	-13.7966	of | annotation.
0.32	-13.7966	= | 1).
0.132867	-13.7967	present | our
0.0376843	-13.7967	by | on
0.153153	-13.7967	a | correction
0.368421	-13.7967	over | dataset
0.0558252	-13.7968	no | in
0.135714	-13.7971	open | to
0.22449	-13.7971	taking | into
0.25641	-13.7972	where | within
0.163043	-13.7972	discarded | the
0.0954064	-13.7973	et | 2003;
0.19403	-13.7973	phrasal | in
0.141732	-13.7974	instance | that
0.149123	-13.7975	trying | the
0.454545	-13.7976	same | concept.
0.454545	-13.7976	search | typically
0.428571	-13.7976	grammar | learned
0.428571	-13.7976	human | general
0.036018	-13.7978	the | human
0.179487	-13.7979	many | same
0.296296	-13.7979	did | significant
0.0531561	-13.7979	to | how
0.229167	-13.798	their | corpus.
0.0733753	-13.7981	space | and
0.571429	-13.7981	comes | computational
0.139535	-13.7981	suggest | is
0.454545	-13.7982	• | select
0.080402	-13.7982	models | by
0.229167	-13.7983	matrix, | of
0.145299	-13.7984	for | equation
0.272727	-13.7985	onset | of
0.0625	-13.7985	to | parsing
0.19697	-13.7986	which | frequency
0.454545	-13.7987	generative | hidden
0.0960854	-13.7987	a | parameters
0.0790123	-13.7988	describe | to
0.25	-13.7988	therefore, | more
0.127389	-13.7988	for example, | figure
0.115789	-13.7989	when | system
0.214286	-13.7989	level | sentence
0.243902	-13.7991	0.5 | 0
0.25	-13.7991	any | phrase
0.296296	-13.7991	randomly | documents
0.4	-13.7992	function | return
0.307692	-13.7993	order | understand
0.0732218	-13.7994	use | our
0.168539	-13.7994	set | examples
0.5	-13.7994	obtained | translating
0.5	-13.7994	ontonotes | sense
0.122093	-13.7995	empirical methods | july
0.0485348	-13.7995	are | in | are
0.158416	-13.7995	in | classifier.
0.32	-13.7995	multilingual | language
0.185714	-13.7995	fairly | and
0.149123	-13.7996	• | between
0.134752	-13.7996	any | word
0.22	-13.7996	using | lm
0.0685185	-13.7997	and | and | 2006).
0.28125	-13.7997	converted | using
0.0125663	-13.7997	is | to | the
0.0715706	-13.7998	in | shown
0.454545	-13.7999	annotated | guidelines
0.0996169	-13.7999	candidate | as
0.138462	-13.8	j=1 | the
0.238095	-13.8	computer | engineering
0.368421	-13.8	proposed | very
0.172414	-13.8001	work | funded
0.75	-13.8001	loss | bigram
0.0822281	-13.8002	by | semantic
0.571429	-13.8003	particular, | explicit
0.0822281	-13.8004	length | to
0.0726141	-13.8004	sentence | from
0.0864553	-13.8004	of | properties
0.168539	-13.8005	participants | for
0.0810127	-13.8005	node | to
0.0267409	-13.8006	set of | the
0.296296	-13.8006	4.1 | was
0.234043	-13.8008	important, | the
0.145299	-13.8008	however, | were
0.109005	-13.8008	on | knowledge
0.22449	-13.8008	s-factor | the
0.15534	-13.8009	and | window
0.571429	-13.801	our | creation,
0.153846	-13.8011	more | because
0.333333	-13.8011	the | htr
0.129032	-13.8014	people | in
0.0491958	-13.8014	c©2011 | computational | of
0.133803	-13.8015	a | implementation
0.333333	-13.8015	any | relations
0.123457	-13.8016	meanings | the
0.243902	-13.8016	v | .
0.0954064	-13.8016	keep | the
0.0696798	-13.8016	multiple | and
0.296296	-13.8017	such | nouns
0.20339	-13.8017	tree | rule
0.12069	-13.8018	2002) | and
0.375	-13.8019	method | fully
0.214286	-13.8019	seems | more
0.5	-13.8019	in | output;
0.75	-13.802	accurate | system’s
0.75	-13.802	thesaurus | distributional
0.75	-13.802	concepts | multimodal
0.151786	-13.802	is | people
0.0771028	-13.8023	3 | as
0.25641	-13.8024	which | global
0.35	-13.8024	where | characters
0.375	-13.8024	also | constituent
0.428571	-13.8025	y | la
0.571429	-13.8025	are | manually.
0.571429	-13.8025	ccg | curran,
0.0924658	-13.8026	the | translating
0.5	-13.8026	train | separate
0.5	-13.8026	vectors | weighted
0.2	-13.8027	compute | two
0.454545	-13.8028	terms | contexts
0.0736842	-13.8028	other | on
0.0924658	-13.8028	done | and
0.35	-13.8028	however, | tags
0.111111	-13.8028	from | distribution
0.307692	-13.8029	an | dictionary.
0.0347744	-13.8029	our | is | of
0.107623	-13.803	like | that
0.12069	-13.803	our | improve
0.0641026	-13.8031	over | for
0.428571	-13.8031	genia | set
0.19403	-13.8031	repeated | and
0.333333	-13.8032	initially, | of
0.5	-13.8032	model, | variable
0.5	-13.8032	generalization | one
0.5	-13.8032	if | ith
0.109524	-13.8034	was | to | with
0.210526	-13.8035	since | work
0.089172	-13.8036	effects | the
0.127389	-13.8037	our | compared
0.126582	-13.804	helpful | the
0.571429	-13.804	for | two-way
0.116402	-13.8041	words, | are
0.0443764	-13.8041	the | x
0.428571	-13.8042	of | 2002.
0.428571	-13.8042	0 | dr
0.19697	-13.8042	training | labeled
0.35	-13.8042	performs | over
0.307692	-13.8045	performed | but
0.5	-13.8045	results | qualitative
0.454545	-13.8046	0.5 | points
0.454545	-13.8046	show | hard
0.114583	-13.8047	to | document.
0.121387	-13.8047	method, | of
0.571429	-13.8047	l | m.
0.146552	-13.8048	in | generative
0.0423317	-13.8048	it is | in
0.0984848	-13.8049	showed | and
0.75	-13.8049	users | tasks.
0.103448	-13.8049	that | perform
0.107981	-13.8049	is | next
0.061747	-13.805	however, | be
0.122807	-13.8051	is | path
0.5	-13.8051	to | phenomenon.
0.20339	-13.8052	threshold | on
0.243902	-13.8053	emacc | the
0.4	-13.8053	equal | same
0.113402	-13.8053	in | product
0.20339	-13.8054	judgments | for
0.296296	-13.8056	if | h
0.0456398	-13.8057	which | by
0.375	-13.8058	semantic | require
0.234043	-13.8059	often | word
0.75	-13.8059	argumentation | argument
0.229167	-13.806	only | even
0.0924658	-13.806	transfer | the
0.141732	-13.8062	of | wsj
0.243902	-13.8063	this | addressed
0.0571792	-13.8064	two | by
0.25641	-13.8064	address | using
0.0873494	-13.8064	in | bilingual
0.137405	-13.8065	possible | or
0.158416	-13.8066	an | question
0.170455	-13.8067	allowed | and
0.375	-13.8068	compare | metrics
0.0666667	-13.8069	in | extraction
0.454545	-13.8069	nullnullnull | nullnullnullnullnullnull
0.32	-13.8069	have | spelling
0.104803	-13.807	need to | in
0.20339	-13.807	simplest | to
0.137405	-13.8071	as | term
0.206897	-13.8071	opinion | target
0.088685	-13.8072	work | which
0.243902	-13.8072	forward | for
0.118644	-13.8073	relevant | is
0.375	-13.8073	not | explored
0.124224	-13.8074	we | queries
0.116402	-13.8074	(8) | the
0.164835	-13.8074	tagged | as
0.428571	-13.8075	linear order | the
0.163043	-13.8075	for | engineering
0.0726141	-13.8076	is | single
0.234043	-13.8077	translation | sentences.
0.307692	-13.8077	construction | this
0.188406	-13.8079	model | relations
0.166667	-13.808	a | maps
0.0984848	-13.8081	of | support
0.114583	-13.8081	investigate | in
0.428571	-13.8081	codes | training
0.0772834	-13.8082	with | 1
0.296296	-13.8082	pair | so
0.264706	-13.8082	function | .
0.175	-13.8083	for | open
0.19403	-13.8083	dt | jj
0.4	-13.8084	scoring | translation
0.571429	-13.8084	rewrite | rules
0.571429	-13.8084	a | formula.
0.22	-13.8084	nodes | which
0.145299	-13.8085	schemes | the
0.088685	-13.8085	in | texts
0.156863	-13.8085	it | words.
0.296296	-13.8086	(daum´e | al.,
0.428571	-13.8086	relation | represented
0.32	-13.8086	identifying | verb
0.164835	-13.8087	strongly | and
0.123457	-13.8088	and | created
0.296296	-13.809	point | its
0.153846	-13.809	system, | that
0.22449	-13.809	learning | multiple
0.5	-13.809	methodology | linguistic
0.571429	-13.8091	distinguish | senses
0.158416	-13.8091	with | reported
0.19403	-13.8091	and | transition-based
0.428571	-13.8092	π | π
0.428571	-13.8092	0.73 | 0.75
0.25641	-13.8094	3 | main
0.25	-13.8094	at | 1,
0.101215	-13.8094	this | model.
0.0828877	-13.8097	of | position
0.5	-13.8097	statistical | moses
0.206897	-13.8098	of | rating
0.117978	-13.8099	portland, | usa,
0.571429	-13.8099	para | on
0.571429	-13.8099	style | specific
0.264706	-13.8099	use | process
0.22449	-13.8101	for | 3).
0.0734341	-13.8102	source | for
0.28125	-13.8102	abstract | analysis
0.375	-13.8102	value | high
0.375	-13.8102	in | 2006.
0.188406	-13.8103	graphical | for
0.141732	-13.8106	an | error
0.19403	-13.8106	graph-based | a
0.75	-13.8107	training | (without
0.121387	-13.8108	decision | is
0.257143	-13.8108	classifier, | is
0.104803	-13.8109	the | linking
0.131944	-13.811	other | also
0.22	-13.811	relation | other
0.32	-13.8111	script | a
0.0865672	-13.8111	a | a | sequence
0.0691589	-13.8111	new | is
0.375	-13.8112	from | dialect
0.206897	-13.8112	paper | research
0.164835	-13.8113	to | annotator
0.108491	-13.8114	data | english
0.0354505	-13.8115	will | to
0.138462	-13.8116	by | classifier
0.375	-13.8116	benefit | only
0.454545	-13.8117	accurate | better
0.75	-13.8117	according | component-based
0.1	-13.8117	such | more
0.257143	-13.8118	on | issues
0.19403	-13.8118	set | across
0.571429	-13.8121	tutor | tutoring
0.571429	-13.8121	but | shorter
0.375	-13.8121	complex | phrases
0.128378	-13.8122	randomly | to
0.0835655	-13.8123	a | important
0.179487	-13.8125	emails | to
0.131944	-13.8126	we | problems
0.179487	-13.8127	reduces | and
0.138462	-13.8127	person | in
0.75	-13.8127	s., | g.,
0.333333	-13.8127	different | structures
0.153846	-13.8128	we | link
0.28125	-13.8128	tag | this
0.20339	-13.8128	are | basis
0.5	-13.8129	textual | linguistic
0.12069	-13.813	pattern | is
0.129252	-13.8131	a | adjective
0.142857	-13.8131	bleu | baseline
0.238095	-13.8132	multimodal | in
0.32	-13.8132	algorithm | weights
0.0988142	-13.8134	works | of
0.28125	-13.8135	each | averaged
0.571429	-13.8136	2004; | 2003)
0.571429	-13.8136	a: | und
0.571429	-13.8136	examples | discriminative
0.0813648	-13.8137	that | relation
0.125	-13.8137	categories | as
0.2	-13.8138	p | g
0.2	-13.8138	also | parser
0.117978	-13.8139	pairs | which
0.4	-13.8142	acknowledgments | discussions
0.0564617	-13.8142	the | point
0.107981	-13.8142	however, | these
0.28125	-13.8142	sprime | s
0.0704501	-13.8143	the | threshold
0.144068	-13.8144	candidate | each
0.152381	-13.8144	is | needs
0.170455	-13.8144	to | cluster.
0.103004	-13.8146	alternative | to
0.375	-13.8146	instances | web
0.188406	-13.8147	he | that
0.4	-13.8147	value, | with
0.4	-13.8147	still | much
0.0865672	-13.8147	the | external
0.072314	-13.8147	we | annotated
0.210526	-13.8147	because | sentence
0.285714	-13.8148	such | ones
0.196721	-13.8149	for | raw
0.20339	-13.8149	unit | as
0.19403	-13.8149	and | early
0.129252	-13.815	with | simple
0.22	-13.815	rather, | of
0.139535	-13.815	and | (for
0.454545	-13.8152	0.69 | 0.65
0.0638978	-13.8152	base | the
0.296296	-13.8152	model | sparse
0.229167	-13.8152	are | conditional
0.257143	-13.8152	summaries | not
0.129252	-13.8153	by | within
0.139535	-13.8154	tree | an
0.0203186	-13.8155	this | is | the
0.135338	-13.8155	evaluation | was
0.75	-13.8156	marsi | nivre
0.296296	-13.8156	one | learned
0.121387	-13.8157	in | queries
0.103448	-13.8157	comes | the
0.454545	-13.8158	up | percentage
0.333333	-13.8158	the | lost.
0.257143	-13.8158	outputs | from
0.131034	-13.8159	first | where
0.22	-13.8159	throughout | and
0.150943	-13.816	find | our
0.0833333	-13.8162	features | our
0.164835	-13.8162	processing | with
0.142857	-13.8162	the | negative.
0.333333	-13.8162	identities | and
0.238095	-13.8163	features | text.
0.0886076	-13.8164	the | approach,
0.129252	-13.8166	are | tags
0.166667	-13.8167	will | both
0.122699	-13.8168	we | address
0.257143	-13.8168	verb | predicate
0.5	-13.8168	approaches | consider
0.307692	-13.8169	= | ∝
0.0106506	-13.8169	is | to
0.0715746	-13.817	given | by
0.105727	-13.8172	binary | is
0.103448	-13.8173	to | knowledge,
0.571429	-13.8173	(3) | clue
0.285714	-13.8174	child | node
0.073913	-13.8174	shows | with
0.215686	-13.8175	optimization | on
0.5	-13.8175	training | sizes.
0.115183	-13.8176	, | )
0.0576671	-13.8177	a | entity
0.103448	-13.8177	of | core
0.243902	-13.8177	includes | such
0.164835	-13.8177	when | very
0.188406	-13.8179	were | because
0.141667	-13.8179	proposed | al.
0.375	-13.818	in | particular.
0.571429	-13.8181	how | speaker
0.272727	-13.8183	real | by
0.104348	-13.8184	that | available
0.136364	-13.8184	category | as
0.75	-13.8185	p(wn|wn−1f
0.296296	-13.8187	active | has
0.333333	-13.8188	pos tagging | pos
0.333333	-13.8188	event | annotations
0.215686	-13.8189	form | have
0.0595745	-13.819	approach | that
0.183099	-13.8192	7, | the
0.28125	-13.8194	but | perform
0.147826	-13.8194	english | were
0.75	-13.8195	kl(p||q)
0.75	-13.8195	suggesting | resource
0.375	-13.8195	from | came
0.571429	-13.8196	entity | closest
0.134328	-13.8197	to | interpretation
0.0753425	-13.8197	our | also
0.318182	-13.8197	evaluation | sentences.
0.0842697	-13.8197	is not | that
0.0756881	-13.8198	results are | in
0.210526	-13.8199	entirely | of
0.454545	-13.82	aids | and
0.117318	-13.82	that | relatively
0.0399493	-13.82	and | these
0.185714	-13.82	correction | for
0.131034	-13.8203	adopted | the
0.571429	-13.8203	thus, | add
0.103896	-13.8204	that | low
0.75	-13.8205	procedings | lrec
0.75	-13.8205	later | often
0.0663176	-13.8205	although | in
0.0798969	-13.8205	for | value
0.257143	-13.8205	elements | or
0.20339	-13.8205	additionally, | for
0.15534	-13.8205	is | link
0.307692	-13.8209	or | proper
0.0537143	-13.8209	methods | to
0.11399	-13.8209	this | framework
0.0629921	-13.821	as | in section
0.571429	-13.8211	espresso | patterns
0.106977	-13.8211	representation | for
0.0423546	-13.8213	feature | a
0.318182	-13.8214	work | ways
0.168539	-13.8215	more | pairs
0.0980392	-13.8215	data | data | for
0.0948905	-13.8215	are | according
0.196721	-13.8216	research | data
0.232558	-13.8216	approach | identify
0.0734341	-13.8218	a | query
0.571429	-13.8218	papers | acl
0.333333	-13.8219	with | old
0.4	-13.822	queries | top
0.307692	-13.8221	examples | any
0.185714	-13.8222	model | segmentation
0.285714	-13.8223	a | phrase-based | machine
0.0462329	-13.8223	the | translations
0.15534	-13.8223	the | e,
0.115183	-13.8224	we | probabilistic
0.0852941	-13.8224	intuition | the
0.75	-13.8225	consistency | among
0.75	-13.8225	neighbors | original
0.75	-13.8225	r | recall,
0.75	-13.8225	p(w|t)p(t|d)
0.4	-13.8225	techniques | english
0.22	-13.8225	linguistics | china
0.272727	-13.8226	analysis | social
0.144068	-13.8227	annotator | a
0.0438125	-13.8227	can | from
0.5	-13.8227	reviews | those
0.142857	-13.8228	extraction | event
0.0260683	-13.8229	some | the
0.159574	-13.8229	visual | to
0.152381	-13.823	instances | which
0.0468343	-13.823	syntactic | of
0.188406	-13.8231	first, | all
0.211538	-13.8231	create | data
0.22449	-13.8233	therefore | our
0.285714	-13.8234	other | verbs
0.5	-13.8234	score, | mean
0.5	-13.8234	lemma | tag
0.454545	-13.8235	modifications | extraction
0.166667	-13.8235	by | during
0.0421348	-13.8235	information | in
0.272727	-13.8236	link, | the
0.0900322	-13.8236	point | to
0.118644	-13.8237	left | in
0.285714	-13.8238	cp | a
0.285714	-13.8238	select | target
0.238095	-13.8238	their | standard
0.126582	-13.824	time, | to
0.307692	-13.8241	pmi | are
0.0397727	-13.8241	in | between | and
0.0299572	-13.8241	the | lexical
0.243902	-13.8241	for | cand
0.318182	-13.8243	decrease | as
0.25	-13.8244	consider | sentences
0.121951	-13.8244	we | randomly
0.0903226	-13.8245	therefore | and
0.318182	-13.8247	will | low
0.571429	-13.8248	expect | degree
0.0114589	-13.8248	the | from
0.307692	-13.8249	the | population.
0.119318	-13.825	that | (e.g.
0.232558	-13.825	they | possible
0.196721	-13.8255	are | key
0.0870871	-13.8255	m | and
0.168675	-13.8258	the | 2010.
0.5	-13.826	features, | standard
0.5	-13.826	examples | decision
0.206897	-13.8261	large | semantic
0.145299	-13.8262	focus | this
0.0955882	-13.8263	segmentation | a
0.117978	-13.8264	(2) where | the
0.375	-13.8264	sections | each
0.20339	-13.8264	mentions | or
0.307692	-13.8266	linking | that
0.156863	-13.8266	a | co-occurrence
0.333333	-13.8267	differences | significant.
0.145299	-13.8269	both | semantic
0.0580297	-13.8271	will | as
0.166667	-13.8272	data | shared
0.264706	-13.8272	s | pp
0.75	-13.8274	say | you
0.061194	-13.8275	metrics | the
0.35	-13.8275	consistent | human
0.272727	-13.8276	extracted | data.
0.296296	-13.8276	the | semiring,
0.20339	-13.8277	methods | also
0.146552	-13.8278	number | possible
0.571429	-13.8278	dialogues | users
0.120482	-13.8279	using | performance
0.0711382	-13.8279	our | have
0.0803109	-13.828	the | connected
0.25	-13.8281	particular | used
0.113402	-13.8282	a | count
0.238095	-13.8285	of | stages
0.318182	-13.8285	construct | semantic
0.571429	-13.8286	may | problems.
0.0621118	-13.8286	word | which
0.146552	-13.8287	for | sequences
0.177215	-13.8287	as | techniques
0.0362205	-13.8287	such | a
0.215686	-13.8288	name | this
0.215686	-13.8288	preposition | as
0.150943	-13.8291	by | argument
0.150943	-13.8291	we | avoid
0.4	-13.8294	3 | graphs
0.22449	-13.8295	each | (e.g.
0.185714	-13.8295	that | information,
0.166667	-13.8296	the | collins
0.257143	-13.8297	or | therefore
0.111675	-13.8298	determined | of
0.0665468	-13.83	entity | in
0.571429	-13.8301	test | game
0.0821918	-13.8303	is | state
0.0860534	-13.8304	show | can
0.272727	-13.8305	parsing | without
0.15534	-13.8305	is, | with
0.185714	-13.8305	our | effective
0.243902	-13.8309	system | comparable
0.149533	-13.8309	work | extraction
0.16129	-13.8309	better | more
0.4	-13.831	measure | under
0.25	-13.831	types | (e.g.,
0.067029	-13.831	5.1 | the
0.183099	-13.831	la | de
0.109453	-13.8312	a | root
0.139535	-13.8312	interesting | for
0.75	-13.8314	that, | returns
0.232558	-13.8314	similar | previous
0.4	-13.8315	designed | test
0.22	-13.8315	are | edit
0.20339	-13.8315	tion | the
0.0676417	-13.8315	provide | and
0.571429	-13.8316	metrics | various
0.264706	-13.8318	that | problems.
0.454545	-13.8319	an | adverb
0.0640394	-13.832	on | two
0.177215	-13.832	with | main
0.00345182	-13.832	the | the | the
0.20339	-13.832	they | accuracy
0.110553	-13.8321	that | discourse
0.191176	-13.8323	and | m.
0.11976	-13.8324	turns | the
0.285714	-13.8324	for | tests.
0.75	-13.8324	kb | represents
0.571429	-13.8324	tagged | entities
0.118644	-13.8325	we | just
0.272727	-13.8326	used | 2007)
0.105023	-13.8328	for | task | in
0.1	-13.833	a | who
0.571429	-13.8331	also | augmented
0.571429	-13.8331	useful | recognizing
0.0865672	-13.8331	of | predicate
0.164835	-13.8332	an | model,
0.375	-13.8333	target-dependent | of
0.375	-13.8333	hierarchical | phrase-based
0.75	-13.8334	exist | pair
0.75	-13.8334	user | amani
0.75	-13.8334	started on | on
0.0478821	-13.8334	are | more
0.180556	-13.8334	where | nodes
0.0696325	-13.8334	the | done
0.215686	-13.8334	this | exploit
0.129252	-13.8335	to | phenomena
0.116667	-13.8335	approach | more
0.132353	-13.8336	semantic | information
0.454545	-13.8337	able | infer
0.0606509	-13.8338	a | noun
0.2	-13.8338	given | (1)
0.571429	-13.8339	defined | equations
0.571429	-13.8339	segmentation | character.
0.215686	-13.8339	proposed | features
0.11	-13.834	suggest | to
0.110553	-13.8341	core | of
0.4	-13.8342	estimates | are
0.188406	-13.8343	sentences | information
0.0924658	-13.8343	the | is then
0.22449	-13.8343	directly | our
0.243902	-13.8344	representation | dependency
0.0541176	-13.8345	was | on
0.232558	-13.8345	we | example.
0.5	-13.8346	formula | i
0.285714	-13.8347	very | like
0.571429	-13.8347	harbin | of
0.375	-13.8348	women’s | of
0.0976562	-13.835	a | parameter
0.125828	-13.8352	domain | from
0.191176	-13.8354	of | error.
0.0942029	-13.8354	then | an
0.0376927	-13.8356	should | the
0.127517	-13.8357	presented | with
0.100418	-13.8358	the | lda
0.285714	-13.8359	em | al.,
0.0909091	-13.8359	in | we describe
0.0942029	-13.8359	it is | which
0.15534	-13.8359	this | consistent
0.0496032	-13.836	for | information
0.0650685	-13.836	user | a
0.5	-13.836	dico | dialogue
0.5	-13.836	adding | significantly
0.4	-13.8363	improved | was
0.4	-13.8363	in | 16,
0.25	-13.8363	experiment | one
0.102564	-13.8365	this | accuracy
0.215686	-13.8365	lexical | pos
0.196721	-13.8367	very | only
0.238095	-13.8367	4 | sets
0.125	-13.8368	other | has
0.170732	-13.8369	propose | novel
0.166667	-13.8372	currently | for
0.17284	-13.8372	functions | as
0.0852941	-13.8372	4 | as
0.243902	-13.8373	only | entities
0.073991	-13.8373	relation | is
0.5	-13.8373	∗department | tokyo,
0.5	-13.8373	0.35 | 0.42
0.5	-13.8373	yv | yp
0.166667	-13.8374	of | nonterminals
0.318182	-13.8374	first | ranking
0.0411241	-13.8377	a | then
0.215686	-13.8377	baseline | sentence
0.180556	-13.8378	that | ordering
0.10084	-13.8378	the | model | its
0.264706	-13.8378	yet | with
0.416667	-13.8379	submission | accepted
0.416667	-13.8379	rank> | (column
0.153846	-13.8379	two | single
0.307692	-13.838	our model | only
0.140496	-13.838	of | years
0.25	-13.838	step, | each
0.5	-13.838	local | τsprime
0.137405	-13.8381	existing | on
0.134328	-13.8383	this | significant
0.375	-13.8383	documents | space
0.20339	-13.8383	examples | figure
0.206897	-13.8384	at | per
0.416667	-13.8385	further | make
0.227273	-13.8385	promising | on
0.571429	-13.8385	parameter | maximize
0.130137	-13.8386	baseline | features
0.193548	-13.8386	edge | between
0.5	-13.8386	book | mary
0.5	-13.8386	intuitively, | when
0.5	-13.8386	τb | gap
0.0323665	-13.8387	the | event
0.0564827	-13.8388	we | lexical
0.183099	-13.8388	phrase | dependency
0.0528684	-13.8388	rules | to
0.333333	-13.8391	though | most
0.20339	-13.8391	to | sequence.
0.191176	-13.8392	fact, | are
0.0938628	-13.8392	while | from
0.5	-13.8393	corrections | teacher
0.75	-13.8394	purpose | classification.
0.177215	-13.8395	performance | compared
0.105505	-13.8395	property | of
0.0942029	-13.8395	to | directly
0.416667	-13.8396	has | empirical
0.15625	-13.8396	2000) | the
0.206897	-13.8398	observations, | the
0.0597668	-13.8399	prediction | the
0.285714	-13.84	relationships | not
0.0602056	-13.84	the | character | the
0.571429	-13.84	investigate | extraction
0.571429	-13.84	null | (nullnullnull)
0.183099	-13.84	it | obtain
0.0680529	-13.8401	that | context
0.152381	-13.8402	are | 10
0.0203469	-13.8402	the | sentence
0.318182	-13.8403	= | 
0.0680529	-13.8404	a | user
0.175	-13.8406	2. | using
0.0980392	-13.8406	in | difference
0.5	-13.8406	element | some
0.5	-13.8406	issues | arise
0.0736607	-13.8408	corpus | al.,
0.571429	-13.8408	should | summary.
0.375	-13.8408	lattice | all
0.333333	-13.8409	30 | each
0.123457	-13.841	evaluation | not
0.4	-13.8411	example | fig.
0.318182	-13.8412	from | database.
0.168675	-13.8413	we | below.
0.178082	-13.8413	higher | but
0.333333	-13.8413	lists | set
0.5	-13.8413	(miwa | 2010;
0.191176	-13.8415	manually | which
0.571429	-13.8415	example | procedure.
0.152381	-13.8416	when | these
0.175	-13.8417	and | verb.
0.257143	-13.8417	as | foreign
0.22	-13.8417	6 | models
0.206897	-13.8417	repeated | a
0.0984252	-13.8418	conditional | of
0.0475755	-13.8418	up | the
0.124224	-13.8419	the | rates
0.0995851	-13.8419	results | over
0.191176	-13.842	learners | and
0.22	-13.842	approach | three
0.5	-13.842	reduce | task
0.5	-13.842	phrases, | phrases,
0.5	-13.842	theories | theory
0.0980392	-13.8421	c | c | e
0.0833333	-13.8421	their | from
0.0690979	-13.8421	words | not
0.454545	-13.8422	sampling | plaintext
0.0347019	-13.8422	the | of each
0.105023	-13.8422	as | –
0.0798969	-13.8425	the | element | the
0.0965251	-13.8427	main | for
0.285714	-13.8427	also | identifying
0.103448	-13.8427	discriminative | and
0.22449	-13.8427	single | only
0.454545	-13.8428	news | those
0.454545	-13.8428	more than | precision
0.375	-13.8428	and | 2005)).
0.416667	-13.8429	cj; | when
0.0428571	-13.8431	of | words | in
0.178082	-13.8432	that | more | less
0.20339	-13.8432	record | a
0.196721	-13.8433	human | but
0.35	-13.8433	typical | can
0.5	-13.8433	α1 | node
0.75	-13.8434	gains | reordering
0.285714	-13.8435	are | sparse
0.168675	-13.8435	automatically | data
0.2	-13.8435	transition-based | a
0.128378	-13.8437	gave | the
0.0421818	-13.8438	translation | in
0.135338	-13.8438	ontology | and
0.35	-13.8438	predict | if
0.375	-13.8438	to | accent
0.207547	-13.8439	during | which
0.193548	-13.844	was | by | foundation
0.20339	-13.844	more | features.
0.138211	-13.8441	word | syntactic
0.117318	-13.8443	achieves | and
0.75	-13.8444	first | binary-valued
0.333333	-13.8445	the | unsuitable
0.285714	-13.8446	k=1 | where
0.416667	-13.8446	(1) | data;
0.149533	-13.8446	attention | of
0.5	-13.8446	φ | dn
0.4	-13.8448	to | trivially
0.126667	-13.8448	with | 5
0.375	-13.8448	model, | text
0.285714	-13.845	this | referring
0.188406	-13.8451	flow | the
0.416667	-13.8452	along | where
0.0845481	-13.8452	nodes | a
0.264706	-13.8452	be | here.
0.125828	-13.8453	composed | of
0.0972763	-13.8454	the | depth
0.227273	-13.8455	and | gale
0.104072	-13.8457	as | specific
0.133333	-13.8457	with | examples
0.132353	-13.8458	we | assume
0.22	-13.8459	used | research
0.5	-13.846	noun | adjective.
0.103604	-13.8461	created | a
0.571429	-13.8461	learned | word-aligned
0.571429	-13.8461	concepts | resulting
0.0730088	-13.8463	is | t
0.264706	-13.8463	percentages | in
0.106481	-13.8464	our | some
0.454545	-13.8465	last | word,
0.0536756	-13.8465	we | with | with
0.22	-13.8465	number | sentences.
0.0448613	-13.8466	the | show
0.0880503	-13.8466	(2010) | in
0.188406	-13.8467	place | a
0.196721	-13.8467	a | te
0.0340587	-13.8467	our | as
0.5	-13.8467	bow | lex
0.2	-13.8469	describes | with
0.0332584	-13.847	in | at
0.22	-13.8471	first, | more
0.0806452	-13.8473	relative | and
0.193548	-13.8474	carry | to
0.307692	-13.8475	5-gram | english
0.307692	-13.8475	express | with
0.307692	-13.8475	each | author
0.243902	-13.8475	appear | same
0.22449	-13.8475	for | resolution.
0.275862	-13.8476	the | hypothesis | hypothesis.
0.17284	-13.8476	in | algorithm,
0.0375289	-13.8476	a | into
0.454545	-13.8477	based | tuning
0.110553	-13.8478	includes | in
0.0954198	-13.8479	to | decoding
0.0708333	-13.8479	to | classification
0.416667	-13.848	features | reflect
0.0721868	-13.848	3 | on
0.22	-13.848	can | effectively
0.5	-13.848	scheme | polarity
0.4	-13.8481	more | distributional
0.0942029	-13.8481	to | fact that
0.142857	-13.8481	and | optimization
0.0991736	-13.8482	give | and
0.0648464	-13.8482	of | simple
0.0677966	-13.8482	found | a
0.227273	-13.8483	we therefore | that
0.157895	-13.8483	consider | words
0.107843	-13.8484	to | rules.
0.416667	-13.8485	90 | 75
0.296296	-13.8486	used | alignments
0.124183	-13.8486	of | children
0.121212	-13.8486	but | has
0.0788804	-13.8486	we | further
0.108911	-13.8486	methods | by
0.272727	-13.8487	that | error.
0.138211	-13.8488	tokens | as
0.285714	-13.8489	argumentation | of
0.207547	-13.8489	that | roughly
0.375	-13.8489	implemented | dialogue
0.0300146	-13.849	a | between
0.272727	-13.8491	those | target
0.17284	-13.8492	of | sequence.
0.0350526	-13.8493	feature | and
0.296296	-13.8494	statistical | sentences
0.0534262	-13.8495	the | mt
0.35	-13.8495	both | produced
0.285714	-13.8496	from | parsing.
0.0720339	-13.85	2009; | et
0.238095	-13.8501	will | next
0.121951	-13.8501	there | there
0.352941	-13.8502	5.3 | section
0.11976	-13.8502	we | want
0.157895	-13.8502	to | child
0.264706	-13.8502	use | tweets
0.0918728	-13.8503	the | phenomena
0.0882353	-13.8503	translation, | the
0.135338	-13.8504	a | constraint
0.080429	-13.8504	in | nodes
0.75	-13.8504	tome | principe
0.375	-13.8504	agreement | high
0.153846	-13.8505	hit | the
0.2	-13.8505	relations | two
0.114754	-13.8506	row | of
0.132353	-13.8506	to | english.
0.193548	-13.8506	used | 2008).
0.0509014	-13.8507	a | syntactic
0.211538	-13.8508	we | tweets.
0.0578512	-13.8509	• | is
0.375	-13.8509	part | office
0.166667	-13.851	and | analyses
0.0991736	-13.8511	separate | of
0.75	-13.8515	previous | (huang
0.571429	-13.8515	indicates | level
0.157895	-13.8517	a | sufficient
0.196721	-13.8518	all | could
0.2	-13.8518	languages | different
0.4	-13.8519	syntactic | arguments.
0.318182	-13.8519	the | wmt11.
0.454545	-13.852	useful | identify
0.108374	-13.852	expected | to
0.227273	-13.852	train | system
0.134328	-13.8522	provided | with
0.0744921	-13.8523	significant | and
0.178082	-13.8523	our | apply
0.571429	-13.8523	both | considerable
0.166667	-13.8524	have | much
0.4	-13.8524	and | summed
0.257143	-13.8524	others, | to
0.264706	-13.8524	use | string
0.416667	-13.8525	annotators | per
0.0671642	-13.8525	the | strategy
0.207547	-13.8526	our | limited
0.227273	-13.8526	annotation | gold
0.139344	-13.8526	vp | a
0.307692	-13.8529	prev | prev
0.170732	-13.8529	3.2 | with
0.4	-13.8529	international | information
0.207547	-13.8529	methods | models
0.227273	-13.8529	do | require
0.123377	-13.8531	for | computing
0.571429	-13.8531	classifier | coarse
0.257143	-13.8531	general | content
0.057377	-13.8532	in | annotation
0.112299	-13.8533	slightly | of
0.114754	-13.8534	a | semantic | semantic
0.375	-13.8534	pos | chunk
0.375	-13.8534	(page | al.,
0.5	-13.8534	disagreement | disagreement
0.352941	-13.8535	tested | data.
0.285714	-13.8535	constraints | section
0.0562827	-13.8535	all | from
0.4	-13.8535	proposed | labeling
0.333333	-13.8535	phrase-table | the
0.333333	-13.8535	c, | =
0.215686	-13.8536	makes | which
0.296296	-13.8538	task | search
0.571429	-13.8538	developed | creation
0.454545	-13.8539	row | represents
0.285714	-13.8539	feature | discriminative
0.352941	-13.854	common | across
0.0411437	-13.8541	search | the
0.20339	-13.8541	information | query
0.5	-13.8541	lead | higher
0.232558	-13.8542	using | computed
0.211538	-13.8542	supervised | our
0.0718816	-13.8544	we | pair
0.352941	-13.8545	don’t | this
0.75	-13.8545	“4) | or
0.124183	-13.8545	additional | on
0.232558	-13.8545	limited | or
0.211538	-13.8545	expand | in
0.1	-13.8546	which | source
0.571429	-13.8546	being | other.
0.571429	-13.8546	learntlow | than
0.571429	-13.8546	hard | both
0.35	-13.8547	extractive | sentences
0.0968992	-13.8548	smt | and
0.257143	-13.8548	the | documents)
0.352941	-13.8549	stanford | 2006).
0.0566535	-13.8549	to | rules
0.264706	-13.8549	analysis | data.
0.243902	-13.8551	example | shows
0.175	-13.8552	used | words.
0.175	-13.8552	a | transformation
0.129496	-13.8552	the | distinction
0.416667	-13.8553	allows | component
0.045611	-13.8553	for each | a
0.175676	-13.8553	3: | =
0.318182	-13.8553	document | words.
0.285714	-13.8554	rarely | a
0.152381	-13.8554	as | had
0.275862	-13.8555	proposed | best
0.121212	-13.8555	detection | et
0.35	-13.8556	relations | within
0.211538	-13.8556	cells | to
0.211538	-13.8556	it | know
0.108374	-13.8557	which | will
0.124183	-13.8558	in | aspects
0.20339	-13.8558	advantage | can
0.111111	-13.8559	that | events
0.132353	-13.856	used | data.
0.0529344	-13.856	evaluation | a
0.0722101	-13.8561	and | high
0.126667	-13.8561	as | supervised
0.5	-13.8561	grouped | same
0.0731707	-13.8562	the | remaining
0.25	-13.8562	handle | that
0.571429	-13.8562	score(x,y) | =
0.454545	-13.8563	ranks | human
0.196721	-13.8563	online | al.,
0.227273	-13.8563	obj | of
0.170732	-13.8564	classification, | and
0.0980392	-13.8565	to | optimal
0.207547	-13.8565	to | comparison.
0.285714	-13.8566	task | identify
0.135338	-13.8566	will | at
0.0987654	-13.8566	where | all
0.115385	-13.8568	the | variant
0.211538	-13.8568	5 | 13
0.15625	-13.8569	show | significantly
0.571429	-13.8569	automatic | summaries.
0.275862	-13.857	rst | is
0.375	-13.857	should | what
0.207547	-13.8571	negative | neutral
0.183099	-13.8571	in | learning,
0.16129	-13.8571	annotations | be
0.0401316	-13.8573	the | user
0.134328	-13.8575	domain | are
0.227273	-13.8575	one | above
0.333333	-13.8575	2009); | al.,
0.35	-13.8575	hw, | mw,
0.454545	-13.8576	address | linguistic
0.0479323	-13.8576	to | some
0.139344	-13.8577	to | effects
0.571429	-13.8577	exhaustive | parsing
0.0670391	-13.8578	by | different
0.132353	-13.8582	labels | by
0.112299	-13.8582	found that | to
0.333333	-13.8584	patterns | role
0.148148	-13.8584	∗ | the
0.275862	-13.8585	0 | -1
0.132353	-13.8585	also | they
0.0175173	-13.8585	we | with
0.75	-13.8586	|w|
0.75	-13.8586	aim | reducing
0.178082	-13.8586	define | between
0.0620032	-13.8587	n | and
0.22	-13.8587	minor | in
0.0545229	-13.8587	single | of
0.185714	-13.8589	clustering | which
0.185714	-13.8589	adjective | is
0.0935252	-13.859	is | thus
0.148148	-13.8592	can | (e.g.,
0.416667	-13.8593	search | click
0.0976562	-13.8593	the | memory
0.571429	-13.8593	subject | much
0.163043	-13.8593	set | corresponding
0.454545	-13.8594	learned | generate
0.137097	-13.8595	to | dynamic
0.5	-13.8595	explicit | order
0.275862	-13.8596	as | crf
0.120482	-13.8596	logical | and
0.75	-13.8596	though | restrict
0.25	-13.8596	also | relevant
0.285714	-13.8597	document | local
0.227273	-13.8597	in | utterances.
0.102222	-13.8597	our experiments | the | of
0.2	-13.8597	z | =
0.0625	-13.86	class | and
0.227273	-13.86	separate | on
0.0612245	-13.86	the | increase
0.0938628	-13.8601	the | kernel
0.0568336	-13.8601	to | similarity
0.0925267	-13.8603	if | 0
0.0734967	-13.8603	surface | the
0.125	-13.8604	from | alignment
0.0782828	-13.8605	map | the
0.318182	-13.8605	 | 
0.318182	-13.8605	7 conclusion | work
0.0316872	-13.8605	this is | the
0.211538	-13.8605	different | how
0.75	-13.8606	quality | ssrs
0.264706	-13.8606	array | and
0.0319731	-13.8608	how | the
0.123377	-13.8608	for | 2006).
0.571429	-13.8608	if | identical.
0.196721	-13.8609	by | definition
0.5	-13.8609	function | strictly
0.416667	-13.861	replacement | metrics
0.080429	-13.8611	next | of
0.22	-13.8611	different | across
0.25	-13.8613	corpus | up
0.175676	-13.8613	use | since
0.0638655	-13.8614	standard | a
0.180556	-13.8614	of | i,
0.0701031	-13.8615	we | parsing
0.139344	-13.8615	however, | when
0.5	-13.8616	y(x) | set
0.152381	-13.8618	an | weight
0.318182	-13.8618	after | some
0.0369334	-13.8618	we | different
0.275862	-13.8619	s | t.
0.0995851	-13.8619	word | different
0.0528129	-13.862	similar | in
0.0830946	-13.8621	reported | of
0.145455	-13.8621	are | individual
0.127517	-13.8622	messages | and
0.116022	-13.8623	2010), | to
0.093985	-13.8624	for | phrases
0.454545	-13.8625	approach, | but
0.227273	-13.8625	such | common
0.352941	-13.8626	both | relatively
0.75	-13.8627	comparison | limited
0.211538	-13.8628	using | framework
0.154639	-13.8629	ccg | and
0.0625	-13.863	we | about
0.285714	-13.8632	use | em
0.318182	-13.8632	presented | extraction
0.416667	-13.8633	high | whereas
0.416667	-13.8633	null2 | null1
0.203704	-13.8634	in | passive
0.0560626	-13.8637	the | combined
0.137097	-13.8638	a | yield
0.4	-13.8638	methods | distributional
0.0654867	-13.8638	this | approach
0.0636516	-13.8639	of | list
0.146789	-13.864	terms | or
0.285714	-13.864	one | change
0.571429	-13.864	39 | al.,
0.0649123	-13.864	are | when
0.144144	-13.8642	as | recognition
0.5	-13.8643	university | columbia
0.454545	-13.8644	express | they
0.0991736	-13.8644	significance | the
0.157895	-13.8644	take | into
0.0793651	-13.8645	some | by
0.75	-13.8647	people | words.
0.178082	-13.8647	has | recent
0.0983607	-13.8647	suggest | of
0.116279	-13.8649	identifying | to
0.164706	-13.8649	t | w
0.232558	-13.865	better | their
0.5	-13.865	extracted | rules,
0.5	-13.865	similarity, | an
0.25	-13.8651	have | information.
0.203704	-13.8654	version | this
0.102679	-13.8655	, | for
0.170732	-13.8657	on | million
0.5	-13.8657	result | accurate
0.5	-13.8657	each | negative,
0.75	-13.8658	0.73 | 5:
0.0194693	-13.8658	we | the | with
0.190476	-13.8658	model | markov
0.190476	-13.8658	predicates | are
0.142857	-13.8659	can | useful
0.10177	-13.8659	problem | with
0.0267221	-13.866	results | and
0.108374	-13.8663	user | that
0.238095	-13.8663	a | block
0.571429	-13.8663	1: | diagram
0.215686	-13.8663	lapata | for
0.275862	-13.8664	simple | (e.g.,
0.0676983	-13.8664	case | is
0.0600601	-13.8664	one | is | and
0.0835735	-13.8664	between | words
0.0348485	-13.8664	in | only
0.168675	-13.8667	transition | for
0.375	-13.8667	(and | if
0.137097	-13.8668	can | those
0.132353	-13.8668	this | prior
0.185714	-13.8671	only | noun
0.149533	-13.8672	methods | used
0.211538	-13.8672	related | relation
0.122581	-13.8673	of | knowledge,
0.0420354	-13.8673	new | of
0.275862	-13.8675	overall | baseline
0.296296	-13.8675	b | k
0.333333	-13.8676	their | limited
0.4	-13.8677	discusion | the
0.142857	-13.8677	few | as
0.5	-13.8677	prior | product
0.5	-13.8677	rate | up
0.0131854	-13.8678	the | and | for
0.75	-13.8678	readily | other
0.15625	-13.8678	property | is
0.130435	-13.8679	computational linguistics | 1
0.296296	-13.8679	the | updated.
0.0756501	-13.8679	very | for
0.126667	-13.868	t | d
0.139344	-13.8681	is | focus
0.133333	-13.8682	systems | translation
0.133333	-13.8682	a | entry
0.106796	-13.8682	elements | and
0.185714	-13.8683	classification | different
0.352941	-13.8684	an | literature
0.227273	-13.8684	enhanced | and
0.10084	-13.8684	cover | the
0.2	-13.8684	from | individual
0.5	-13.8684	speaker | object
0.131387	-13.8685	automatic | by
0.75	-13.8688	spanish | portuguese
0.304348	-13.8688	approaches | comparable
0.178082	-13.8688	consider | sentence | the
0.352941	-13.8689	result | about
0.133333	-13.8689	workshop | statistical
0.154639	-13.869	field | to
0.416667	-13.869	translations | spanish
0.5	-13.8691	approach | used.
0.137097	-13.8694	semantic | its
0.141667	-13.8694	word | lexical
0.164706	-13.8694	number | terms
0.227273	-13.8694	triggering | a
0.16129	-13.8694	thus, | by
0.25	-13.8695	the | consonant
0.175676	-13.8695	is | typical
0.296296	-13.8696	the | readers.
0.183099	-13.8696	performance | also
0.173333	-13.8697	type of | are
0.149533	-13.8698	because | one
0.75	-13.8699	prednisone | augmentin
0.185714	-13.8699	with | edges
0.318182	-13.8701	discourse | contrast
0.0609375	-13.8702	a | parse
0.5	-13.8705	noun | (or
0.5	-13.8705	in | experiment).
0.5	-13.8705	prior | required
0.022991	-13.8706	the | given
0.138211	-13.8706	the | fails
0.054254	-13.8706	those | and
0.215686	-13.8707	the | path,
0.146789	-13.8709	additionally, we | the
0.0690335	-13.8709	and | form
0.75	-13.8709	significant | χ2
0.193548	-13.8709	most | english
0.375	-13.8709	0.74 | 0.64
0.101322	-13.8712	study | that
0.1875	-13.8712	using | labels
0.454545	-13.8713	text | institute
0.107843	-13.8714	three | which
0.318182	-13.8715	approach | 2006).
0.232558	-13.8717	lexical | resources
0.168675	-13.872	instance | or
0.232558	-13.872	that | zero
0.0823864	-13.872	necessary | the
0.4	-13.8721	models | always
0.257143	-13.8721	these | questions
0.0732265	-13.8722	we | label
0.100418	-13.8722	already | to
0.352941	-13.8723	information | distributional
0.0819209	-13.8723	we | corpora
0.0616114	-13.8724	was | from
0.0851064	-13.8725	of | required
0.304348	-13.8726	selecting | most
0.333333	-13.8726	other | reported
0.571429	-13.8726	discussions | social
0.571429	-13.8726	morphological | tagger,
0.5	-13.8726	3.2 | n-gram
0.0991736	-13.8727	the | advanced
0.352941	-13.8728	monolingual | reordering
0.227273	-13.8728	children | are
0.222222	-13.8729	included | not
0.178082	-13.873	the | scene
0.4	-13.8732	graphs, | as
0.124183	-13.8732	as | discourse
0.133333	-13.8732	created | for
0.178082	-13.8732	in | pp
0.166667	-13.8733	further | which
0.125828	-13.8733	that | ∈
0.062603	-13.8733	frequency | of
0.5	-13.8733	human | judgments.
0.243243	-13.8734	a | component,
0.571429	-13.8734	noun | monosemous
0.183099	-13.8734	large | data.
0.148148	-13.8734	but | would
0.0547264	-13.8737	for | over
0.264706	-13.8737	5 | evaluated
0.243243	-13.8738	our | distinct
0.180556	-13.8738	our | had
0.222222	-13.8739	methods | where
0.5	-13.874	consists | 100
0.416667	-13.8741	each | iteration.
0.571429	-13.8742	still | processing
0.571429	-13.8742	experiment | corpora.
0.0791557	-13.8742	of | points
0.352941	-13.8743	by | changing
0.454545	-13.8744	better | quality.
0.454545	-13.8744	spectral | speech
0.130435	-13.8744	at | its
0.193548	-13.8744	recent | language
0.1875	-13.8745	instead | use
0.243243	-13.8745	function | time
0.0625	-13.8746	the | summary
0.0280343	-13.8746	work | and
0.5	-13.8746	context | effective
0.5	-13.8746	matches | target
0.0603715	-13.8748	sentence | with
0.296296	-13.8749	compute | pairs
0.207547	-13.875	at | without
0.375	-13.875	(with | not
0.75	-13.8751	associations | certain
0.75	-13.8751	setting | resources
0.75	-13.8751	i(u|q)
0.5	-13.8753	ability | document
0.5	-13.8753	although | idea
0.168675	-13.8755	an | improvement
0.375	-13.8755	accuracy | pmi
0.051087	-13.8756	the | reordering
0.173333	-13.8757	translation | rules
0.047048	-13.8759	table | results
0.140496	-13.876	sets, | of
0.75	-13.8761	entry | row
0.25	-13.8761	evaluate | at
0.25	-13.8761	on | approaches.
0.138211	-13.8762	for | change
0.0773067	-13.8762	results | data
0.227273	-13.8763	z, | the
0.111111	-13.8763	argmax | the
0.264706	-13.8763	vector, | is
0.140496	-13.8764	if | target
0.304348	-13.8764	the | pushing
0.125	-13.8764	furthermore, | that
0.222222	-13.8766	+ | bow
0.571429	-13.8766	any | conditions
0.117647	-13.8767	time | on
0.112903	-13.8767	is | coreference
0.5	-13.8767	gemeinsames | merkmal
0.0717391	-13.8768	the | mapping
0.175676	-13.877	while | method
0.75	-13.8771	relations | terms.
0.4	-13.8771	do | nor
0.0607477	-13.8772	that | target
0.571429	-13.8773	approach | clinical
0.0843373	-13.8773	a | unsupervised
0.416667	-13.8776	gains | features
0.130435	-13.8776	of | completely
0.333333	-13.8777	curve | figure
0.222222	-13.8779	role | argument
0.15625	-13.8779	with | rate
0.154639	-13.878	formula | of
0.304348	-13.8781	must | when
0.375	-13.8781	study | improve
0.5	-13.8781	|r|
0.285714	-13.8782	sl | are
0.4	-13.8782	so | additional
0.4	-13.8782	manchester, | uk
0.207547	-13.8782	data | automatic
0.0791557	-13.8782	high | a
0.146789	-13.8784	to | fragments
0.304348	-13.8785	their | representation
0.296296	-13.8786	innovative | building
0.375	-13.8787	experiments | wsj
0.416667	-13.8788	available | web.
0.416667	-13.8788	probability | t,
0.5	-13.8788	sd | mcccj
0.266667	-13.8789	adding | more
0.178082	-13.8789	as | makes
0.296296	-13.879	system | ranked
0.6	-13.879	(the, | de)
0.6	-13.879	japanese | sensing
0.6	-13.879	(eisenstein | 2010).
0.6	-13.879	student | tutor.
0.6	-13.879	ned | induced
0.6	-13.879	synonym | conceptresolver
0.6	-13.879	fr | es
0.6	-13.879	+ | i;t
0.1	-13.8791	question | is
0.375	-13.8792	f-score | gold
0.0761671	-13.8793	we | unsupervised
0.318182	-13.8794	work | classifiers
0.0967742	-13.8795	we | articles
0.175676	-13.8795	baseline | feature
0.6	-13.8798	weather | vas
0.6	-13.8798	if | eaten.
0.6	-13.8798	lends | well
0.6	-13.8798	pa | œ
0.318182	-13.8798	(brants | al.,
0.416667	-13.8799	based | frequency.
0.207547	-13.8799	while | lexical
0.193548	-13.88	object | with
0.275862	-13.8801	correctly | an
0.196721	-13.8801	in | article.
0.141593	-13.8801	lexical | each
0.101322	-13.8802	some | more
0.159574	-13.8803	translation | lexical
0.170732	-13.8804	3.4 | and
0.196721	-13.8804	can | range
0.0482759	-13.8804	4 | a
0.153061	-13.8805	it | result
0.6	-13.8806	1.21 | 1.30
0.6	-13.8806	(2) | imputed
0.352941	-13.8807	among | about
0.12766	-13.8808	to | output.
0.190476	-13.8808	that | lists
0.0441176	-13.8809	top | the
0.116959	-13.8809	for | complex
0.5	-13.8809	significantly | either
0.296296	-13.881	synchronous | that
0.0203868	-13.881	by | a
0.100437	-13.8812	svm | and
0.184615	-13.8813	is | involving
0.0754258	-13.8814	parameters | of
0.104762	-13.8815	that | word | that
0.6	-13.8815	utterances | transcribed
0.6	-13.8815	attachment | (uas),
0.6	-13.8815	turn | floor
0.6	-13.8815	h. | t.
0.318182	-13.8816	method | transliteration
0.5	-13.8816	of | intervals,
0.0680934	-13.8818	in | parse
0.16092	-13.8819	of | lemma
0.227273	-13.8819	english | contains
0.333333	-13.8819	the | alg.
0.0791557	-13.8819	learn | and
0.285714	-13.8822	an | node.
0.0420736	-13.8822	the | quality
0.137097	-13.8823	( | ,
0.6	-13.8823	medical | medicine
0.6	-13.8823	ki | who
0.6	-13.8823	obtains | reduction
0.6	-13.8823	ratio | presence
0.6	-13.8823	constraints | relaxation.
0.6	-13.8823	preservers | reversers
0.275862	-13.8824	use | contexts
0.130435	-13.8824	trained | trained
0.75	-13.8824	parsers | wsj
0.0637931	-13.8824	to | domain
0.318182	-13.8825	models | provide
0.266667	-13.8827	coefficient | =
0.138211	-13.8827	by | automatic
0.296296	-13.8827	phrase | moses
0.207547	-13.8828	hidden | models
0.257143	-13.8828	time | event
0.124183	-13.883	the | assumed
0.153061	-13.883	model | time
0.066055	-13.8831	we | parallel
0.6	-13.8831	match | counts
0.6	-13.8831	posterior | star
0.6	-13.8831	all | vl,
0.6	-13.8831	design | submodular
0.6	-13.8831	john | usa
0.0320763	-13.8831	system | to
0.0575843	-13.8832	categories | the
0.217391	-13.8833	created | all
0.125	-13.8833	states | in
0.352941	-13.8836	encouraged | to
0.107843	-13.8836	is | tags
0.454545	-13.8839	process, | sentence
0.097561	-13.8839	data | two
0.0656934	-13.884	a | classification
0.6	-13.884	pls | test
0.6	-13.884	effectiveness | items
0.6	-13.884	closely related | mining
0.6	-13.884	embedded | prepositional
0.118343	-13.8841	1 | example
0.184615	-13.8841	) | θ
0.153061	-13.8841	data | news
0.0424564	-13.8842	each | the | that
0.125	-13.8842	to | equation
0.0876623	-13.8843	all | training
0.75	-13.8844	corpus, | measured
0.75	-13.8844	state: | [slot2],
0.304348	-13.8844	improved | it
0.454545	-13.8845	annotations | labels
0.196721	-13.8845	sentence, | be
0.571429	-13.8845	tagging | tagged
0.114943	-13.8845	contained | of
0.0643478	-13.8846	for | any
0.285714	-13.8846	for | shift-reduce
0.304348	-13.8848	languages | syntactic
0.132812	-13.8848	the | maximizes
0.117647	-13.8849	which | new
0.203704	-13.8849	words | up
0.193548	-13.8849	the | ministry
0.5	-13.8851	outputs | highest
0.454545	-13.8852	classes | per
0.0835821	-13.8853	f | a
0.121795	-13.8854	that | coreference
0.266667	-13.8856	user | first
0.0814607	-13.8856	these two | the
0.0266787	-13.8857	that | from
0.6	-13.8857	preposition | writer
0.6	-13.8857	bhagat | 2008;
0.6	-13.8857	see | trend
0.6	-13.8857	generation | summarization,
0.207547	-13.886	here | model
0.304348	-13.8861	+ | (7)
0.0943396	-13.8862	structure | as
0.116279	-13.8863	is | token
0.0912281	-13.8863	we | errors
0.124183	-13.8863	coverage | a
0.454545	-13.8864	a | encountered.
0.184615	-13.8864	introduced | with
0.75	-13.8865	20 | 0.9
0.0669216	-13.8865	the | introduced
0.5	-13.8865	penalty | if
0.352941	-13.8866	4 | associated
0.352941	-13.8866	optimization | function
0.0873786	-13.8866	discuss | of
0.122581	-13.8868	forms | a
0.0437601	-13.8868	the | data,
0.25	-13.8869	these | expressions
0.207547	-13.8869	using | like
0.571429	-13.8869	biased | smt
0.243243	-13.887	able | information
0.173333	-13.887	1 | no
0.140351	-13.8872	lexical | their
0.211538	-13.8872	extracted | their
0.6	-13.8873	there | gender-specific
0.124183	-13.8875	= | >
0.142857	-13.8875	is | conditional
0.215686	-13.8875	have | evaluated
0.0727273	-13.8875	the | protein
0.109375	-13.8876	on | set,
0.1875	-13.8876	by | adding
0.75	-13.8876	analysis | wsi
0.113636	-13.8877	constructed | a
0.105263	-13.8877	to | manual
0.142857	-13.8877	types | each
0.0670498	-13.8877	is | to | other
0.196721	-13.8878	maintaining | the
0.171053	-13.8878	words | frequent
0.153061	-13.8878	boundaries | a
0.0323152	-13.8879	words | a
0.0717391	-13.8879	words | which
0.5	-13.8879	task | answers.
0.190476	-13.888	also | applied
0.416667	-13.8881	vector | high
0.6	-13.8882	local | word-pos
0.6	-13.8882	item | xis
0.318182	-13.8883	will | existing
0.454545	-13.8884	multiple | labels
0.0735632	-13.8884	which | all
0.215686	-13.8884	by | applying
0.352941	-13.8886	methods | ability
0.375	-13.8886	provides | systems
0.375	-13.8886	constraint | if
0.0733945	-13.8887	is | less
0.180556	-13.8887	decoding | an
0.141593	-13.8888	which | original
0.333333	-13.8889	reverb | than
0.6	-13.889	missing | 5%
0.6	-13.889	transitions | global
0.6	-13.889	components | integrated
0.217391	-13.8891	with | taking
0.1875	-13.8892	structure | dependency
0.232558	-13.8892	in | significant.
0.211538	-13.8892	define | two
0.0437601	-13.8893	in | given
0.207547	-13.8895	scenes | and
0.123377	-13.8895	these results | that
0.124183	-13.8896	examples of | in
0.75	-13.8897	(n | 1)
0.0896552	-13.8898	on | while
0.0322866	-13.8898	the | time
0.6	-13.8899	(section | provided
0.6	-13.8899	sentences | markup.
0.6	-13.8899	showed | em
0.6	-13.8899	productdisplay | factor
0.222222	-13.89	embedded | a
0.180556	-13.8902	where | pair
0.159091	-13.8903	to | incremental
0.0503751	-13.8904	different | for
0.416667	-13.8904	represents | meaning
0.243243	-13.8904	metric | at
0.178082	-13.8904	information | when
0.352941	-13.8905	are | treatment
0.352941	-13.8905	topics | related
0.352941	-13.8905	say | have
0.0803324	-13.8905	with | score
0.0787402	-13.8905	the | percentage of
0.175676	-13.8905	their | syntactic
0.162791	-13.8906	the | treelet
0.6	-13.8907	shallow-transfer | systems
0.6	-13.8907	associated | green.
0.6	-13.8907	on | drafts
0.6	-13.8907	tasks | ie
0.75	-13.8908	(not | due
0.157895	-13.8909	set | x
0.217391	-13.891	sentence | positive
0.0865385	-13.8911	in | 2008).
0.0454545	-13.8913	as | and | are
0.211538	-13.8913	additional | between
0.157895	-13.8914	a | slot
0.5	-13.8914	mining | news
0.2	-13.8915	structure | into
0.571429	-13.8917	0.89 | 0.97
0.285714	-13.8918	pronoun | or
0.0790191	-13.8918	is | find
0.0732265	-13.8918	a | weight
0.318182	-13.8919	if | assign
0.0442774	-13.8921	text | the | of
0.207547	-13.8921	evaluation | described
0.454545	-13.8922	systems | results.
0.193548	-13.8922	“michael | the
0.6	-13.8924	0, | >
0.6	-13.8924	have | dicto
0.6	-13.8924	alignment | alignment-oriented
0.333333	-13.8926	considered | gold
0.162791	-13.8927	but | time
0.232558	-13.8928	show | without
0.193548	-13.8928	the | minus
0.140351	-13.8929	models | other
0.173333	-13.8929	model | meaning
0.193548	-13.893	traditional | as
0.217391	-13.8931	judgements | of
0.0750605	-13.8931	head | and
0.6	-13.8932	smoothing | unseen
0.6	-13.8932	words | (rule
0.0327869	-13.8934	we | it
0.352941	-13.8935	it | guaranteed
0.352941	-13.8935	features | attributes
0.454545	-13.8935	extraction | extended
0.119497	-13.8935	in | testing
0.178082	-13.8936	each | several
0.13913	-13.8937	gene | in
0.0416048	-13.8938	(2) | the
0.266667	-13.8939	comes | with
0.75	-13.8939	einen | pds
0.145455	-13.894	of | cluster.
0.136	-13.894	possible | this
0.333333	-13.894	probabilistic | lexical
0.454545	-13.8941	obtained | probabilities
0.0631399	-13.8941	e | r | e
0.6	-13.8941	|i
0.6	-13.8941	imbalanced | class
0.6	-13.8941	83 | model
0.6	-13.8941	boys | girls
0.6	-13.8941	or | propbank,
0.6	-13.8941	calculated | counting
0.6	-13.8941	take | as | example.
0.207547	-13.8942	setting | that
0.121019	-13.8942	but | same
0.217391	-13.8943	with | arc
0.171053	-13.8943	there | improvement
0.140351	-13.8943	we | assign
0.0525687	-13.8943	position | the
0.0366762	-13.8944	large | of
0.190476	-13.8944	was | higher
0.416667	-13.8945	factorized | and
0.175676	-13.8945	for | recent
0.050885	-13.8947	as | all
0.0801105	-13.8948	denote | of
0.6	-13.8949	0.31 | 0.36
0.6	-13.8949	strings | punctuations
0.266667	-13.895	relations | event
0.318182	-13.895	because | simple
0.0754258	-13.8952	task | that
0.0494737	-13.8952	– | and
0.0944882	-13.8953	and | had
0.132812	-13.8953	relation | which
0.0497354	-13.8953	human | of
0.236842	-13.8954	choices | for
0.203704	-13.8954	grammatical | with
0.171053	-13.8955	process, | in
0.0818713	-13.8955	null | nullnullnull
0.375	-13.8955	models | them.
0.16092	-13.8956	task | evaluation
0.5	-13.8956	created | verb
0.0818713	-13.8957	of | concepts
0.236842	-13.8958	prior | where
0.571429	-13.8958	pr(w′|w)
0.6	-13.8958	reliable | multiple
0.153061	-13.8959	one | phrase
0.352941	-13.896	define | meaning
0.75	-13.896	whereas | simply
0.171053	-13.896	dynamic | for
0.120253	-13.8961	we | extend
0.175676	-13.8962	ordering | is
0.257143	-13.8962	27 | and
0.109375	-13.8963	an | lexical
0.227273	-13.8963	moreover, | word
0.227273	-13.8963	they | correct
0.318182	-13.8963	(2006) | translation
0.5	-13.8963	cues | speech
0.0547074	-13.8964	the | 2008).
0.304348	-13.8964	among | also
0.333333	-13.8964	tree | proposed
0.126761	-13.8965	we | incorporate
0.266667	-13.8965	) | data
0.0750605	-13.8965	in | state
0.375	-13.8965	research | problem
0.571429	-13.8966	but | area
0.6	-13.8966	since | earliest
0.6	-13.8966	sum | scaled
0.6	-13.8966	unsupervised | (table
0.25	-13.8967	of | mention.
0.304348	-13.8968	words | hence
0.416667	-13.8969	found | cluster
0.175676	-13.897	set | contains
0.5	-13.897	arabic | analyzer
0.0893471	-13.8971	we | measures
0.222222	-13.8972	when | larger
0.0313152	-13.8973	to | such
0.196429	-13.8973	even | training
0.151515	-13.8973	wikipedia | with
0.13913	-13.8974	pattern | that
0.6	-13.8975	clearly | e.g.,
0.6	-13.8975	annotator | text.
0.137097	-13.8978	two | most
0.0857143	-13.8978	the | boundaries
0.275862	-13.8979	n-best | each
0.130435	-13.8979	we | two | each
0.454545	-13.898	sentence | explicit
0.454545	-13.898	document | conference
0.146789	-13.898	weights | on
0.0697674	-13.898	except | the
0.0662879	-13.898	predicate | the
0.184615	-13.8981	words | considered
0.180556	-13.8981	or | even
0.0919118	-13.8981	it | both
0.318182	-13.8981	often | very
0.285714	-13.8982	in | discussion.
0.141593	-13.8982	relations | not
0.227273	-13.8982	text | large
0.0489297	-13.8983	and | english
0.6	-13.8983	t(e|
0.232558	-13.8984	reviews, | of
0.173333	-13.8985	complex | such
0.454545	-13.8986	has | consistent
0.184615	-13.8986	to | classes,
0.1875	-13.8987	is, | not
0.175676	-13.8987	+ | pos
0.132812	-13.8987	2010 | of
0.375	-13.8987	some | native
0.136	-13.8988	source | into
0.147059	-13.8988	and | information,
0.0461255	-13.8989	relation | and
0.571429	-13.899	derivation | exactly
0.0682731	-13.899	head | of
0.75	-13.8992	labels | instance.
0.333333	-13.8992	discriminative | alignment
0.153061	-13.8992	assessment | and
0.454545	-13.8993	are | targeting
0.232558	-13.8994	improvement | at
0.0500544	-13.8995	test | in
0.175676	-13.8997	applied | which
0.133858	-13.8998	results | set.
0.162791	-13.8998	x | set
0.193548	-13.8999	= | β
0.0963855	-13.9001	local | in
0.0708155	-13.9002	result | and
0.222222	-13.9003	relation | if
0.203704	-13.9006	is | tweet
0.0922509	-13.9006	verb | as
0.196429	-13.9007	of | la
0.217391	-13.9008	in | my
0.134921	-13.9008	for | problems
0.085443	-13.9008	of | five
0.6	-13.9009	i|
0.6	-13.9009	complexity | significant
0.6	-13.9009	non-relevant | asr
0.6	-13.9009	linear | again
0.0379589	-13.901	the | sets
0.196429	-13.901	informed | the
0.131783	-13.901	resolution | a
0.11875	-13.901	the | turns
0.304348	-13.9011	only | e.g.
0.243243	-13.9011	the | geographical
0.0221893	-13.9011	in | which
0.0551905	-13.9013	the | coreference
0.0708155	-13.9013	step | and
0.5	-13.9013	like | models,
0.571429	-13.9014	articles | deposition
0.0884354	-13.9015	however, | it
0.227273	-13.9015	10 | (%)
0.121019	-13.9015	and | evidence
0.102804	-13.9015	(3) | in
0.157303	-13.9016	encoded | a
0.0881356	-13.9017	the | a.
0.0619765	-13.9017	with | two
0.232558	-13.9017	their | scores
0.6	-13.9017	discuss | give
0.6	-13.9017	scanning, | target
0.0338401	-13.9018	we | more
0.112994	-13.9018	word | information
0.0562414	-13.9019	can | can
0.243243	-13.9021	where | 5
0.119497	-13.9024	discussion | for
0.118012	-13.9024	it | so
0.75	-13.9024	metrics | classes.
0.75	-13.9024	ˆpmle(e|q)
0.375	-13.9024	loss | best
0.125	-13.9024	japanese | of
0.352941	-13.9025	given | trigger
0.454545	-13.9025	represented | state
0.6	-13.9026	det | state
0.6	-13.9026	you | write
0.6	-13.9026	was | leaving
0.6	-13.9026	model | gothic
0.113636	-13.9027	detection | for
0.1875	-13.9027	that | abstract
0.203704	-13.9027	which | performance.
0.232558	-13.9027	is | cue
0.416667	-13.9028	emotion | love
0.222222	-13.9029	by | probabilistic
0.193548	-13.9029	work | grammar
0.375	-13.9029	benefit | our
0.352941	-13.903	evidence | such
0.203704	-13.903	as | arguments.
0.285714	-13.9031	to | competing
0.0118853	-13.9031	we | the | in
0.375	-13.9034	use | history
0.5	-13.9034	similar | method,
0.75	-13.9035	label | actually
0.75	-13.9035	algorithms | matsumoto,
0.203704	-13.9035	some | even
0.178082	-13.9035	that | association
0.0932836	-13.9035	that, | to
0.6	-13.9035	paper | writing
0.154639	-13.9036	two | standard
0.212766	-13.9036	information | relation
0.266667	-13.9037	annotated | such as
0.203704	-13.9038	+ | ≤
0.0366919	-13.9038	data | for
0.184615	-13.9039	to | semantics.
0.212766	-13.9042	a | zi
0.304348	-13.9042	it | outputs
0.0688935	-13.9044	approach | from
0.333333	-13.9044	structure | related
0.117284	-13.9046	models | words
0.178082	-13.9047	work | well
0.571429	-13.9047	articles | interest
0.333333	-13.9049	portion | our
0.144144	-13.9049	actually | a
0.352941	-13.905	twitter | social
0.275862	-13.905	the | committee
0.0691824	-13.905	values | to
0.109948	-13.9051	build | in
0.454545	-13.9051	words | indicative
0.168831	-13.9052	also | much
0.166667	-13.9052	the | ti
0.212766	-13.9052	word | window
0.6	-13.9052	form | supervision.
0.6	-13.9052	solution | bayesian
0.6	-13.9052	within | common
0.6	-13.9052	employing | where
0.0585586	-13.9052	and | large
0.2	-13.9054	will | those
0.148515	-13.9056	in | segment
0.0767263	-13.9056	rule | is
0.5	-13.9056	wordnet | (miller
0.212766	-13.9058	the | respectively;
0.333333	-13.9058	workshops | al.,
0.193548	-13.9059	the | transliterations
0.232558	-13.906	with | special
0.6	-13.906	particularly | complex
0.6	-13.906	beam | larger
0.164706	-13.9061	10 | number
0.25	-13.9063	variable | or
0.333333	-13.9063	compute | distance
0.333333	-13.9063	gtrec | gtrec
0.454545	-13.9064	i | design
0.416667	-13.9064	what | present
0.107692	-13.9064	other | word
0.207547	-13.9065	problem, | on
0.0664137	-13.9066	sense | a
0.227273	-13.9066	they | features.
0.140351	-13.9066	were | would
0.222222	-13.9067	each | it.
0.0696203	-13.9067	same | that
0.266667	-13.9068	stylistic | and
0.6	-13.9069	viewed | form
0.6	-13.9069	holds | rule
0.217391	-13.907	time | system
0.175676	-13.907	was | so
0.571429	-13.9071	will | receive
0.114286	-13.9072	study | on
0.108808	-13.9072	are | higher
0.151515	-13.9072	appear | as
0.375	-13.9072	if | npk
0.0126252	-13.9073	for | in
0.164706	-13.9073	experiments | word
0.227273	-13.9073	simplest | is
0.333333	-13.9073	this | finds
0.275862	-13.9074	methods | selection
0.173333	-13.9074	is | implemented
0.117284	-13.9076	and | require
0.236842	-13.9076	3.2, | the
0.227273	-13.9076	grammar | it
0.037417	-13.9076	set | for
0.454545	-13.9077	among | approaches
0.6	-13.9077	consensus | component
0.6	-13.9077	a | (sect.
0.190476	-13.9077	the | terminals
0.0555556	-13.9078	5 | for
0.121795	-13.9078	for | candidates
0.151515	-13.9079	evaluation | both
0.222222	-13.9079	reviews | from
0.222222	-13.9079	that | entity.
0.162791	-13.908	is | other.
0.232558	-13.908	first, | only
0.291667	-13.908	workers | or
0.352941	-13.9081	combinatory | grammar
0.25	-13.9081	our | labels.
0.454545	-13.9083	hiberno-english, | of
0.0813953	-13.9084	take | a
0.5	-13.9084	accurate | should
0.0590015	-13.9085	is | by | by
0.171053	-13.9086	data | supervised
0.6	-13.9086	now | kb
0.0599369	-13.9087	the | metrics
0.134921	-13.9087	evaluation | test
0.571429	-13.9088	 | np
0.375	-13.9088	utterance | when
0.75	-13.9089	how | works.
0.454545	-13.909	be | expensive.
0.0798898	-13.909	distribution | for
0.25	-13.9092	1 | sequence
0.0838509	-13.9092	where | not
0.0462963	-13.9093	other | as
0.184615	-13.9094	i=1 | where
0.243243	-13.9094	only | time.
0.133858	-13.9095	currently | to
0.6	-13.9095	paraphrase | (dolan
0.352941	-13.9096	our | closest
0.454545	-13.9096	selection | examples
0.154639	-13.9096	category | that
0.232558	-13.9096	distance | between | in
0.275862	-13.9097	test | hypothesis
0.333333	-13.9097	step, | can
0.00926063	-13.9098	is | the | of
0.162791	-13.9098	contribution | that
0.107143	-13.9098	structures | a
0.193548	-13.9098	of | successful
0.416667	-13.9099	paper | sections
0.318182	-13.9099	these | trigger
0.5	-13.9099	good | according
0.285714	-13.91	uses | about
0.15	-13.9101	item | is
0.266667	-13.9102	the | second-best
0.266667	-13.9102	of | related.
0.0759494	-13.9102	is | output
0.109948	-13.9103	use | based
0.120253	-13.9103	for | observed
0.118012	-13.9103	had | for
0.129771	-13.9103	the | formed
0.171053	-13.9103	states | are
0.25	-13.9103	however, | multiple
0.6	-13.9103	dialogue | operate
0.6	-13.9103	effect | increases
0.142857	-13.9104	from | scores
0.571429	-13.9104	thus, | identification
0.318182	-13.9104	given | candidates
0.275862	-13.9105	probabilities | rules
0.416667	-13.9105	following | string
0.5	-13.9106	41st | the
0.134921	-13.9107	our | linguistic
0.184615	-13.9108	final | each
0.203704	-13.9108	and | ).
0.164706	-13.9108	the | leftmost
0.032057	-13.9108	this | from
0.0863787	-13.9109	the | linked
0.124138	-13.9111	derivation | in
0.416667	-13.9111	several | clusters
0.0915751	-13.9112	a | difference
0.153061	-13.9112	results | improve
0.6	-13.9112	optimizing | according
0.275862	-13.9113	questions: | the
0.0802292	-13.9113	thus, | in
0.5	-13.9113	results | 5%
0.243243	-13.9115	(3) | where
0.375	-13.9115	n=1 | n
0.352941	-13.9116	so | directly
0.352941	-13.9116	one | factor
0.132812	-13.9116	a | attribute
0.416667	-13.9117	solid | edges
0.203704	-13.9117	4 | present
0.25	-13.9117	features | recall
0.137931	-13.9119	evidence | is
0.257143	-13.9119	an | disambiguation
0.102804	-13.9119	returned | the
0.5	-13.912	google | view,
0.5	-13.912	open | project
0.352941	-13.9121	thus | higher
0.352941	-13.9121	œ | k
0.285714	-13.9121	event | within
0.75	-13.9121	resource | kind
0.75	-13.9121	variables | values.
0.0419527	-13.9122	are | than
0.0454959	-13.9122	labels | the
0.257143	-13.9123	actions | or
0.275862	-13.9125	obtain | corpus
0.0956175	-13.9125	out | that
0.333333	-13.9125	few | examples
0.173333	-13.9126	possible | one
0.291667	-13.9126	not | available.
0.222222	-13.9127	label | using
0.0712695	-13.9127	to | statistical
0.0944882	-13.9128	showed | to
0.571429	-13.9129	candidates | sentences.
0.6	-13.9129	ceaf | best
0.0638298	-13.9129	b | a
0.0621849	-13.9129	proposed | in
0.0765306	-13.913	a | l
0.207547	-13.9131	can | extract
0.207547	-13.9131	of | semantics,
0.375	-13.9131	laboratory | university,
0.0922509	-13.9132	result, | the
0.108808	-13.9134	on | would
0.222222	-13.9137	are | discriminative
0.304348	-13.9138	uses | data.
0.6	-13.9138	6.1 | 2010
0.6	-13.9138	authors | lda
0.6	-13.9138	haipahaada | sundaradikhataa
0.6	-13.9138	yue | cambridge
0.104762	-13.9141	n | v
0.454545	-13.9142	relations | nouns
0.75	-13.9142	binding | gene
0.266667	-13.9144	variable | an
0.118012	-13.9144	general | with
0.0511945	-13.9145	the | categories
0.164706	-13.9146	to | vectors.
0.6	-13.9146	arrested | request
0.416667	-13.9147	judges | how
0.0963855	-13.9147	translation | system
0.104265	-13.9148	of | arabic
0.0991379	-13.9148	calculated | and
0.0966387	-13.9149	were | this
0.181818	-13.9149	these | like
0.291667	-13.9151	initialize | =
0.115607	-13.9152	queries | a
0.416667	-13.9153	any | action
0.75	-13.9153	resources | wikipedia,
0.571429	-13.9153	sum | every
0.571429	-13.9153	framework, | learning
0.120253	-13.9154	two | feature
0.121019	-13.9154	frequently | and
0.6	-13.9155	parser | seconds
0.291667	-13.9155	5 | compared
0.08	-13.9156	the | otherwise
0.352941	-13.9157	vbd | nnp
0.243243	-13.9157	later | et
0.0863787	-13.9157	4.2 | and
0.375	-13.9158	model | brown
0.375	-13.9158	mellon | forbes
0.416667	-13.9159	tweets | messages
0.130769	-13.9159	that | coverage
0.148515	-13.916	only | pairs
0.0627178	-13.916	function | in
0.0654206	-13.916	or | which
0.147059	-13.916	most | was
0.275862	-13.9161	consequently, | is
0.136752	-13.9161	alignment | phrase
0.2	-13.9161	allow | or
0.571429	-13.9162	• | f.
0.0919118	-13.9162	arguments | in
0.103774	-13.9163	point | is
0.6	-13.9164	as | beforehand
0.6	-13.9164	mine | search
0.375	-13.9164	related | (e.g.
0.128788	-13.9166	can | input
0.157303	-13.9167	accuracy | each
0.136752	-13.9167	model | parsing
0.0821114	-13.9168	this | test
0.291667	-13.9168	university | college
0.132812	-13.9169	for | matching
0.571429	-13.917	frame | conceptual
0.121622	-13.917	as | through
0.416667	-13.9171	very | task.
0.25	-13.9171	tree | function
0.0915751	-13.9171	annotation | by
0.5	-13.9171	from | paths.
0.291667	-13.9172	semantic | entailment
0.275862	-13.9173	while | lower
0.117284	-13.9174	and | highly
0.375	-13.9174	in | jdmwe
0.454545	-13.9175	the | mturk.
0.454545	-13.9175	chain | 1
0.2	-13.9175	model | w
0.258065	-13.9176	reduction | error
0.0974576	-13.9177	the | assigns
0.145631	-13.9178	nodes | from
0.333333	-13.9178	rejected | by
0.5	-13.9178	generously | the
0.217391	-13.9179	product | by
0.166667	-13.9179	first | document
0.106599	-13.9179	we | i.e.,
0.0555556	-13.918	were | on
0.454545	-13.9181	introduced | order
0.6	-13.9181	may | physical
0.352941	-13.9182	relations | sentences.
0.0484037	-13.9182	6 | the | of
0.131783	-13.9182	extract | that
0.266667	-13.9183	main | when
0.196429	-13.9184	or | likely
0.133858	-13.9184	where | system
0.173333	-13.9184	choice | by
0.222222	-13.9185	have | overall
0.384615	-13.9185	machine | synchronous
0.75	-13.9186	grammar | sections
0.571429	-13.9186	novel | synchronous
0.571429	-13.9186	questions | satisfaction
0.571429	-13.9186	argmaxynull | null
0.318182	-13.9186	the | interval.
0.0539153	-13.9186	of | n
0.192982	-13.9187	not | just
0.0400281	-13.9187	we | where
0.162791	-13.9187	from | end
0.102804	-13.9187	against | in
0.0346093	-13.9188	was | a
0.077957	-13.9189	content | a
0.13913	-13.9189	the | disambiguation.
0.6	-13.919	lateen | dependency
0.6	-13.919	parts: | part
0.6	-13.919	maximum | (mle)
0.0411985	-13.9192	of | pages | 2011.
0.120253	-13.9192	such | other
0.179104	-13.9193	through | can
0.0725995	-13.9195	that | improve
0.75	-13.9196	categories | tags,
0.384615	-13.9196	items | both
0.352941	-13.9197	those | foundation.
0.196429	-13.9198	of | responses.
0.0594679	-13.9198	a | specific
0.0539846	-13.9199	a | data.
0.151515	-13.9199	where | defined
0.304348	-13.9199	could | up
0.123288	-13.9199	attributes | and
0.352941	-13.9203	model | normal
0.243243	-13.9203	parsing | 2008).
0.102326	-13.9203	are | 3.
0.184615	-13.9206	sms | to
0.258065	-13.9206	detection | words
0.145631	-13.9207	distinguish | a
0.75	-13.9207	others | tend
0.0739857	-13.9207	denotes | the
0.114943	-13.9207	2003), | the
0.6	-13.9207	because | proved
0.6	-13.9207	non-projective | training
0.6	-13.9207	pos, | only
0.6	-13.9207	url | web
0.375	-13.9207	moses | our
0.102804	-13.9207	sentences | an
0.5	-13.9207	each token | then
0.0690377	-13.9209	in | 2,
0.168831	-13.921	from | provide
0.0915751	-13.9212	on | new
0.333333	-13.9212	one | issue
0.116564	-13.9213	or | some
0.384615	-13.9213	f1 | 61.5
0.318182	-13.9214	information | (in
0.100917	-13.9215	trigger | a
0.168831	-13.9215	as | 2009),
0.159091	-13.9215	of | files
0.0499457	-13.9216	this | data
0.6	-13.9216	synonyms | examples
0.166667	-13.9217	each | consists
0.304348	-13.9217	candidate | noun
0.333333	-13.9217	results, | section
0.0528256	-13.9218	this | words
0.75	-13.9218	p(y|x,θ)
0.207547	-13.9218	the | procedure,
0.416667	-13.922	presented | way
0.454545	-13.9221	0.7 | 100.0
0.454545	-13.9221	decoding | models.
0.5	-13.9221	weighting | weights
0.5	-13.9221	research | 2005).
0.5	-13.9221	phosphor | pr
0.119497	-13.9222	this | content
0.110497	-13.9223	system | different
0.0607553	-13.9223	3. | the
0.0603175	-13.9223	there | be
0.384615	-13.9224	very | corpora.
0.6	-13.9224	correlations | computed
0.144231	-13.9224	annotators | as
0.052071	-13.9225	experiments | in
0.192982	-13.9226	2: | do
0.159091	-13.9226	a | eq.
0.131783	-13.9226	when | its
0.173333	-13.9226	looking | that
0.171053	-13.9227	patterns, | and
0.0137645	-13.9228	a | of | in
0.318182	-13.9228	learning | optimization
0.0604575	-13.9229	and | described
0.132812	-13.9229	and | efficient
0.222222	-13.923	it | agreement
0.040404	-13.9231	the | t
0.291667	-13.9231	precision | computed
0.6	-13.9233	each | link.
0.6	-13.9233	process | learning.
0.135593	-13.9234	size | training
0.212766	-13.9235	earthquake | the
0.184615	-13.9236	our | graph
0.0451671	-13.9236	the | we use
0.333333	-13.9236	relation | identified
0.333333	-13.9236	uses | local
0.258065	-13.9236	shows | correlation
0.0262681	-13.9237	the | performance
0.416667	-13.9238	however, | worse
0.184615	-13.9238	of | 1996).
0.136752	-13.9238	only | both
0.0429846	-13.9238	for | only
0.175676	-13.9239	and | derivations
0.120805	-13.924	demonstrates | the
0.375	-13.924	compare | measures
0.454545	-13.9241	argue | better
0.333333	-13.9241	3.4 | each
0.0301689	-13.9241	this | is | and
0.6	-13.9242	for | pos-tagger.
0.6	-13.9242	jump | jumps
0.6	-13.9242	filtering | svm
0.103286	-13.9242	the | subsets
0.0213105	-13.9243	the | which | to
0.0886525	-13.9243	new | on
0.5	-13.9243	choice | noise
0.416667	-13.9244	has | (for
0.175676	-13.9244	there | about
0.041958	-13.9245	while | in
0.153846	-13.9245	θ | 
0.147059	-13.9245	notice | in
0.1875	-13.9246	2008 | and
0.1875	-13.9246	it | words,
0.227273	-13.9247	only | considered
0.051462	-13.9247	with | set
0.141593	-13.9249	an | (e.g.,
0.208333	-13.9249	our | support
0.0966387	-13.925	effect | in
0.6	-13.925	subject | complement.
0.75	-13.9251	9. | trees
0.153846	-13.9252	relation | semantic
0.140351	-13.9252	have | higher
0.181818	-13.9252	analysis | all
0.571429	-13.9253	predictions | according
0.144231	-13.9253	as | string
0.184615	-13.9254	can | as:
0.25	-13.9254	have | right
0.217391	-13.9255	include | as well
0.184615	-13.9257	on | trees.
0.0506187	-13.9258	system | as
0.0622837	-13.9259	and | structure
0.6	-13.9259	determiners | type
0.454545	-13.926	sentence, | those
0.10989	-13.926	and | rank
0.088968	-13.926	the | θ
0.333333	-13.926	order | during
0.318182	-13.926	present | them
0.258065	-13.9262	classifier | determine
0.120253	-13.9263	an | similar
0.212766	-13.9263	constraints | this
0.0201211	-13.9263	and | for | and
0.384615	-13.9263	with | parser’s
0.0755668	-13.9263	use | word
0.0912409	-13.9264	probabilities | in
0.5	-13.9265	methodology | automatic
0.243243	-13.9266	we | concrete
0.145631	-13.9267	decoding | as
0.0838509	-13.9267	in | paper.
0.181818	-13.9267	transitions | to
0.114458	-13.9267	for | few
0.6	-13.9268	use | fillers,
0.6	-13.9268	modeller | input
0.222222	-13.9269	provide | system
0.104478	-13.9269	basis | of
0.25	-13.9269	used | had
0.0467381	-13.9269	a | probability
0.571429	-13.9269	les | sont
0.571429	-13.9269	(peel) | (tear
0.571429	-13.9269	backoff | bon−1(wi−11
0.571429	-13.9269	chaos | presidential
0.571429	-13.9269	[a | [b
0.571429	-13.9269	vu | boelelaan
0.571429	-13.9269	overlapping | w∈ri
0.571429	-13.9269	taskease | dialoguelength
0.571429	-13.9269	bl | ci+st
0.571429	-13.9269	sf | jf
0.571429	-13.9269	voitures | cars]
0.571429	-13.9269	es+er+pc | es+pc+epr
0.0742574	-13.9271	as | most
0.111732	-13.9272	however, | model
0.266667	-13.9272	regular | or
0.207547	-13.9272	by | complete
0.75	-13.9273	historical | newspapers
0.243243	-13.9273	(np | (nn
0.123288	-13.9273	this | approaches
0.0615641	-13.9274	for | lexical
0.384615	-13.9274	6.2 | classification
0.266667	-13.9275	and | snyder
0.333333	-13.9275	rules | parsing
0.1	-13.9276	the | choosing
0.15	-13.9277	that | alternative
0.173333	-13.9277	such | even
0.112994	-13.9278	examples | with
0.333333	-13.928	crf | training
0.0458801	-13.928	means | the
0.236842	-13.9281	parameters | development
0.155556	-13.9282	information | university
0.192982	-13.9282	unsupervised | training
0.176471	-13.9282	= | γ
0.291667	-13.9282	order | every
0.266667	-13.9283	experimental | methods
0.102326	-13.9283	features | different
0.375	-13.9283	list | verbs
0.6	-13.9285	finite | clause
0.111111	-13.9286	two | first
0.266667	-13.9287	key | information
0.5	-13.9287	best | selection.
0.304348	-13.9288	results | versus
0.0948617	-13.9289	by | single
0.196429	-13.929	we | correlations
0.416667	-13.9292	some | japanese
0.258065	-13.9292	section | concludes | work.
0.059375	-13.9293	take | of
0.104478	-13.9294	is | point
0.153846	-13.9294	s | q
0.0820669	-13.9294	frequent | and
0.5	-13.9294	problem | estimation
0.196429	-13.9295	improvement | translation
0.0234924	-13.9295	such | and
0.09375	-13.9296	sentiment | is
0.184615	-13.9297	a | speech.
0.384615	-13.9297	if | arcs
0.0813953	-13.9297	the | units
0.333333	-13.9299	subjectivity | words
0.0325933	-13.9299	a | but
0.352941	-13.93	will | consider
0.275862	-13.9301	δ | in
0.275862	-13.9301	about | them
0.0578635	-13.9301	manually | the
0.304348	-13.9301	segmentation | alignment
0.07277	-13.9302	20 | of
0.10989	-13.9302	indicating | of
0.25	-13.9302	length | per
0.101382	-13.9303	and | representations
0.148515	-13.9303	synset | a
0.222222	-13.9304	frequency | our
0.304348	-13.9305	without | context
0.75	-13.9306	explicitly | information.
0.0880282	-13.9306	present | on
0.454545	-13.9307	atomic | are
0.0167535	-13.9307	a | and | a
0.0916031	-13.9308	sources | the
0.0703297	-13.9309	such | can
0.152174	-13.9309	n | all
0.13913	-13.9309	when | translation
0.153846	-13.931	of | p.
0.318182	-13.9311	detected | with
0.6	-13.9312	large | datasets.
0.114458	-13.9312	is | mentions
0.0637523	-13.9313	the | word, | the
0.227273	-13.9313	issue | by
0.0609555	-13.9313	domain | and
0.454545	-13.9314	learn | meaning
0.0492505	-13.9315	might | the
0.352941	-13.9316	syntactic | through
0.5	-13.9316	9 | 35
0.75	-13.9317	setup | 2003;
0.196429	-13.9318	expected | by
0.0397767	-13.9319	approach | in
0.0966387	-13.9319	a | modeling
0.304348	-13.9319	expressions: | usa,
0.0508083	-13.9319	relations | and
0.384615	-13.9319	achieve | improvements
0.13913	-13.9319	our | models.
0.454545	-13.932	role | discourse
0.454545	-13.932	motivation | task
0.454545	-13.932	resolution | problem
0.0572687	-13.932	see | of
0.0825688	-13.932	a | individual
0.135593	-13.9321	are | patterns
0.151515	-13.9322	30 | in
0.304348	-13.9323	segments | from
0.304348	-13.9323	affect | wordnet
0.5	-13.9323	empty | during
0.122449	-13.9325	an | time
0.103286	-13.9325	words | different
0.352941	-13.9326	because | transition
0.454545	-13.9327	rb | rb
0.454545	-13.9327	(l) | (l)
0.454545	-13.9327	0.45 | 0.54
0.454545	-13.9327	remote | casings
0.454545	-13.9327	éplucher | apart)
0.454545	-13.9327	e3 | e3
0.0788732	-13.9327	an | feature
0.16092	-13.9329	has been | work
0.275862	-13.9329	task, | training
0.275862	-13.9329	which | achieve
0.203704	-13.933	for | 6.
0.5	-13.9331	agreement | phrasal
0.0767196	-13.9333	in | (e.g.
0.144231	-13.9333	context | language
0.171053	-13.9334	as | identifying
0.416667	-13.9335	provided | sets
0.132812	-13.9335	previous | this
0.128788	-13.9336	way | by
0.304348	-13.9337	many | without
0.291667	-13.9337	interestingly, | on
0.6	-13.9338	p(w|z)
0.6	-13.9338	t | steps
0.6	-13.9338	p(w|h)
0.0528967	-13.9339	context | a
0.102804	-13.9339	most | this
0.189655	-13.934	however, | less
0.147059	-13.9342	on | testing
0.126866	-13.9342	of | tasks,
0.0773333	-13.9344	english | as
0.0858086	-13.9345	tasks | to
0.5	-13.9345	negation | (e.g.,
0.266667	-13.9346	algorithm | previous
0.416667	-13.9347	returned | query
0.6	-13.9347	belief | route
0.6	-13.9347	study | usage
0.091954	-13.9348	a | represents
0.333333	-13.9348	develop | set
0.159091	-13.9349	for | around
0.208333	-13.935	size | but
0.125874	-13.9352	subjects | of
0.1625	-13.9352	indeed, | and
0.116564	-13.9353	by | search
0.384615	-13.9353	studies | corpus
0.5	-13.9353	coordination | has
0.189655	-13.9354	probability | w
0.375	-13.9355	rules | logical
0.0809249	-13.9355	correct | to
0.6	-13.9356	abstract | fine-grained
0.6	-13.9356	usually | several
0.0600649	-13.9356	8 | the
0.352941	-13.9357	decoding | while
0.352941	-13.9357	of | building,
0.10989	-13.9357	in | produced
0.258065	-13.9357	example | i
0.0564175	-13.9358	this | information
0.145631	-13.9358	system | described
0.333333	-13.9358	approach | applicable
0.375	-13.936	section | evaluates
0.75	-13.9361	aravind | bonnie
0.75	-13.9361	prior | assignments
0.25	-13.9361	that | lemma
0.0825688	-13.9361	previous work | the
0.352941	-13.9362	extract | monolingual
0.0533672	-13.9362	each | which
0.0874126	-13.9362	if | words
0.384615	-13.9364	creation | language
0.0974576	-13.9365	a | set,
0.0231504	-13.9365	word | and
0.155556	-13.9366	reasonable | a
0.275862	-13.9366	a | window.
0.15	-13.9367	of | items.
0.0759162	-13.9369	for | analysis
0.243243	-13.937	finally, | approach
0.384615	-13.937	score | pair.
0.0982906	-13.9371	for | random
0.0950413	-13.9371	2. | with
0.181818	-13.9371	denote | with
0.102941	-13.9372	or | if
0.352941	-13.9373	within | sense
0.122449	-13.9374	for | resolution
0.6	-13.9374	the | male.
0.0307757	-13.9375	the | three
0.5	-13.9375	study | community
0.5	-13.9375	0.05 | 0.17
0.384615	-13.9376	moreover, | applied
0.291667	-13.9376	and | speaker.
0.0569343	-13.9378	method | for
0.0574374	-13.9378	the | 2007).
0.131783	-13.9378	dictionary, | the
0.0569343	-13.9379	and | accuracy
0.0461538	-13.938	argument | the
0.107143	-13.938	above | is
0.130769	-13.938	used | 2010).
0.106061	-13.938	from | lexical
0.0958333	-13.938	the | ...
0.136752	-13.9381	when | feature
0.173913	-13.9381	that | preposition
0.384615	-13.9381	charniak | model
0.113095	-13.9382	feature | number
0.6	-13.9382	questionnaire | developed
0.352941	-13.9383	relations | polarity
0.75	-13.9383	subtree | contains
0.0846906	-13.9385	this | pairs
0.0477642	-13.9385	task | to | the
0.217391	-13.9386	challenge | et
0.184615	-13.9387	that | writing
0.179104	-13.9387	could | at
0.333333	-13.9387	want | whether
0.196429	-13.9388	model | base
0.236842	-13.9389	incoming | in
0.136752	-13.9389	stands | the
0.0986547	-13.9389	source | which
0.5	-13.939	supervised | expensive
0.6	-13.9391	dialogue | preferences
0.125	-13.9391	rules | which
0.141509	-13.9392	on | chinese
0.266667	-13.9393	on | seems
0.0866667	-13.9393	of | points | the
0.75	-13.9394	group | unrelated
0.0534351	-13.9394	for | first
0.141509	-13.9394	indicating | is
0.153846	-13.9394	showed | can
0.12	-13.9395	they | only
0.111732	-13.9396	only | was
0.056338	-13.9396	the | effect
0.0735294	-13.9398	(or | the
0.384615	-13.9398	wiki | with
0.222222	-13.9399	e. | and
0.333333	-13.9402	systems | automatically
0.208333	-13.9402	suffix | a
0.5	-13.9404	provide | implementation
0.189655	-13.9404	we | adjacent
0.75	-13.9405	on | procesing.
0.75	-13.9405	evidence | formulas
0.0625	-13.9405	every | of
0.0804598	-13.9405	is | .
0.153846	-13.9405	was | while
0.100917	-13.9406	this | make
0.6	-13.9409	pipeline | actual
0.0477157	-13.941	our | an
0.0518697	-13.941	possible | and
0.375	-13.941	et | cao
0.106061	-13.941	the | definitions
0.333333	-13.9412	coreference | performance
0.5	-13.9412	terms | strong
0.1875	-13.9413	mbr | combination
0.159091	-13.9413	both | and | set
0.0771277	-13.9413	automatic | of | and
0.0970464	-13.9414	adaptation | and
0.120805	-13.9414	we | expressions
0.0982143	-13.9415	for | segmentation
0.0651341	-13.9415	level | of
0.384615	-13.9415	s | c.
0.75	-13.9416	against | behavior
0.0782123	-13.9416	a | common
0.236842	-13.9417	90% | for
0.153846	-13.9417	number | n
0.5	-13.9419	interest | nlp
0.0479666	-13.9419	for | into
0.352941	-13.942	choice | such
0.266667	-13.942	f-score | +
0.171053	-13.942	using | tags
0.070615	-13.9421	methods | al.,
0.0232389	-13.9421	between | and | the
0.164557	-13.9424	for | hidden
0.352941	-13.9425	two | subsets
0.352941	-13.9425	microsoft | wa
0.196429	-13.9426	gale | the
0.0219166	-13.9426	a | is | in
0.0946502	-13.9427	algorithm | al.,
0.1875	-13.9427	results | much
0.75	-13.9427	august | publication:
0.75	-13.9427	bond | ionic
0.75	-13.9427	oov | aug-1gram/ant-filt
0.75	-13.9427	mercy | tents
0.75	-13.9427	(parker | 2009)
0.75	-13.9427	92% | 100%
0.75	-13.9427	arendasy | holling
0.75	-13.9427	58 | 56
0.75	-13.9427	descriptions | gre3d7
0.75	-13.9427	ssb | kinc
0.75	-13.9427	mˇeigu´o | zˇongtˇong
0.75	-13.9427	wtm | o(ir
0.75	-13.9427	page | browser
0.75	-13.9427	neg | neut
0.75	-13.9427	countmatch(n-gram) | (n-gram)
0.75	-13.9427	p(e) | 1)j
0.75	-13.9427	∩a) | p(c/¬a)∗p(¬a)
0.75	-13.9427	pmi | (is,
0.75	-13.9427	parenleftbigc | 12sr
0.75	-13.9427	0.92 | 0.33
0.75	-13.9427	_伞 | _钱
0.75	-13.9427	σ | σ2
0.75	-13.9427	programa | y
0.75	-13.9427	s77 | s103
0.75	-13.9427	z1 | z2
0.75	-13.9427	variation | (vi)
0.75	-13.9427	(†) | str(a,i)∈l(ψ)
0.75	-13.9427	p(θ+i | p(θ−i
0.75	-13.9427	nps | bare
0.75	-13.9427	abstraction | wiki
0.75	-13.9427	informatics | hsr
0.75	-13.9427	facial | (facs)
0.75	-13.9427	72 | 84
0.75	-13.9427	0.10 | 0.30 0.40
0.75	-13.9427	“boy | “tall
0.75	-13.9427	blast | typology
0.75	-13.9427	“crf | crf-bp”
0.75	-13.9427	(alexandrescu | kirchhoff,
0.75	-13.9427	2% | due
0.75	-13.9427	+hv; | (gi
0.75	-13.9427	cells | θcij
0.75	-13.9427	level) | (at
0.75	-13.9427	cookbook | crisis.
0.75	-13.9427	11.85 | 13.27
0.75	-13.9427	dfc1 | d3g2
0.75	-13.9427	c1, | c3,
0.75	-13.9427	the | growth followed
0.75	-13.9427	cn1 | ˆk,
0.75	-13.9427	c | tracer
0.75	-13.9427	55.6 | 63.8
0.75	-13.9427	qi | q′i
0.75	-13.9427	border | states
0.75	-13.9427	d(aj | 0|aj−1
0.75	-13.9427	min | bracketrightbig
0.75	-13.9427	combine([d+n1]::[nns1], | nns1]
0.75	-13.9427	这些 | _多少
0.75	-13.9427	ui | σ2u),
0.75	-13.9427	zuv,n# | zuv,#m
0.75	-13.9427	grupo | tecnología
0.75	-13.9427	kılıç | fish’
0.75	-13.9427	(freitag, | finn
0.75	-13.9427	“in-chart” | “out-chart”
0.75	-13.9427	semeval-2 | 14
0.75	-13.9427	t5 | t7
0.75	-13.9427	cmu-denkowski | umd-eidelman
0.75	-13.9427	bill | clinton,
0.75	-13.9427	bus | names,
0.75	-13.9427	tot | 53.4
0.75	-13.9427	hnew | h
0.75	-13.9427	f(a) | i∈e
0.75	-13.9427	factorial | (ghahramani
0.75	-13.9427	pargram | (butt
0.75	-13.9427	warg | wpred,
0.75	-13.9427	fom | beam-predict
0.75	-13.9427	o-dep | f-dep
0.75	-13.9427	bill | cornelia
0.75	-13.9427	86 | words−js
0.75	-13.9427	reap | eskenazi,
0.75	-13.9427	nulldnullsummationdisplay | nksummationdisplay
0.75	-13.9427	jsem | b´yt
0.75	-13.9427	∀f,e | ∀e
0.75	-13.9427	 is |  meat.
0.75	-13.9427	2600 | 's
0.75	-13.9427	labeled | [mann
0.75	-13.9427	p(q|d) | p(f|d,e)
0.75	-13.9427	yd | ysi
0.75	-13.9427	britain | prolonged
0.75	-13.9427	mcclosky | miwa
0.75	-13.9427	e1-1 | t1-1
0.75	-13.9427	links: | adja
0.75	-13.9427	vpos | vphr
0.75	-13.9427	fct | funding)
0.75	-13.9427	2001; | hwa,
0.75	-13.9427	sorensen, | nguyen
0.75	-13.9427	1computer | department,
0.75	-13.9427	steve | holt
0.75	-13.9427	video | sony
0.75	-13.9427	logit | fixed
0.75	-13.9427	 period |  news
0.75	-13.9427	evama4 | mahilaaen5
0.75	-13.9427	(i,t) | 〈wt,ft
0.75	-13.9427	stimuli | multimedia
0.75	-13.9427	kik | dj.
0.75	-13.9427	ghent | 281
0.75	-13.9427	cand | cand,1st
0.75	-13.9427	d98 | d89
0.75	-13.9427	(α=0.05) | (α=0.01);
0.75	-13.9427	_？ | _雨伞
0.75	-13.9427	researcher, | porter=porter
0.75	-13.9427	ί | ο
0.75	-13.9427	rafael | fusionopolis
0.75	-13.9427	8,000 | 56,000
0.75	-13.9427	/b/−/d/ | erence
0.75	-13.9427	.75 | .78
0.75	-13.9427	no. | length(si)
0.75	-13.9427	barker | included
0.75	-13.9427	ljubljana, | deparment
0.75	-13.9427	y∈y | p(x,y,z)
0.75	-13.9427	“[a0 | commission]
0.75	-13.9427	mind/brain | rovereto
0.75	-13.9427	d87 | d85
0.25	-13.9427	might | would
0.6	-13.9427	necessary | match
0.375	-13.9427	somewhat | right:
0.375	-13.9427	threw | pass
0.0986547	-13.9428	manually | a
0.2	-13.943	standard | phrase
0.318182	-13.9433	on | extent
0.258065	-13.9433	used | 2002)
0.225	-13.9434	finally, | future
0.243243	-13.9434	shows | high
0.222222	-13.9435	when | type
0.222222	-13.9435	more | through
0.173913	-13.9435	previous | two
0.0259114	-13.9436	used | the | of
0.140351	-13.9436	after | an
0.6	-13.9436	our approach | conceptually
0.0855263	-13.9437	parsing | al.,
0.106599	-13.9438	we | via
0.416667	-13.944	nullnullnullnullnullnullnullnullnull | nullnullnullnullnullnullnull
0.258065	-13.944	(lm) | the
0.258065	-13.944	translation | often
0.105	-13.9443	frequency | for
0.0650096	-13.9443	are | semantic
0.0982143	-13.9444	the | wide
0.150538	-13.9444	because | use
0.6	-13.9445	uas | english
0.1	-13.9446	selecting | of
0.0434783	-13.9446	similarity | and
0.0197795	-13.9446	the | for | is
0.2	-13.9447	system | multiple
0.133333	-13.9448	semantic | different
0.318182	-13.9448	code | from
0.243243	-13.9449	each | turn
0.0504009	-13.9449	trees | the
0.164557	-13.9451	that | spoken
0.192982	-13.9452	arabic | are
0.416667	-13.9452	these | convey
0.208333	-13.9452	assume | one
0.0990991	-13.9452	of | method.
0.111111	-13.9453	the | compositionality
0.6	-13.9453	kingdom | .
0.6	-13.9453	e1 | is,
0.6	-13.9453	expressive | they
0.1625	-13.9456	communicative | and
0.106599	-13.9456	as | less
0.105528	-13.9456	input | with
0.258065	-13.9456	prec | prec
0.5	-13.9456	trbm | features
0.291667	-13.9458	information | mutual
0.291667	-13.9458	determine | information
0.113772	-13.9459	model | these
0.136752	-13.9459	for | experiment
0.243243	-13.946	cosine | vector
0.0680851	-13.946	by | same
0.258065	-13.946	grammatical | language
0.384615	-13.9461	further | average
0.039604	-13.9461	and | they
0.222222	-13.9462	many | sentences
0.0328185	-13.9463	is | more
0.1625	-13.9463	news | by
0.212766	-13.9464	all | errors
0.416667	-13.9464	algorithm | trees.
0.222222	-13.9465	that | remains
0.0738916	-13.9466	over | with
0.217391	-13.9466	or | context.
0.0982143	-13.9466	from | large
0.333333	-13.9466	case | annotation
0.181818	-13.9466	this | designed
0.091954	-13.9468	extended | of
0.0950413	-13.9468	results are | table
0.225	-13.9471	when | up
0.140187	-13.9471	mentions | as
0.142857	-13.9471	and | yields
0.333333	-13.9471	figure 4: | as
0.6	-13.9471	identified | factors
0.204082	-13.9471	from | what
0.106599	-13.9472	process | as
0.083871	-13.9475	corpus | for | and
0.140187	-13.9475	to | regular
0.25	-13.9475	from | transfer
0.275862	-13.9476	more | towards
0.304348	-13.9476	but | determine
0.0912548	-13.9476	is | (see
0.142857	-13.9476	a | explicit
0.318182	-13.9476	pairs | input
0.416667	-13.9477	improve | grammar
0.184615	-13.9477	designed | with
0.0685225	-13.9477	on | other hand,
0.0749354	-13.9478	is | previous
0.5	-13.9478	0.61 | 0.67
0.0662651	-13.948	on | was
0.6	-13.948	accuracy | high,
0.189655	-13.948	an | sentence,
0.0765172	-13.9481	the | predicates
0.0615385	-13.9482	for | do
0.171053	-13.9483	learning | their
0.171053	-13.9483	binary | on
0.208333	-13.9483	corpus | generated
0.0707763	-13.9484	we | values | the
0.243243	-13.9485	an | learner
0.230769	-13.9485	language | error
0.5	-13.9486	(luong | al.,
0.0966387	-13.9487	a | candidates
0.6	-13.9489	code | assigned
0.6	-13.9489	obtained | incorporating
0.112426	-13.9491	(ii) | of
0.159091	-13.9491	2 | first
0.333333	-13.9491	research | activity
0.2	-13.9492	fact | only
0.0495595	-13.9493	indicates | the
0.117647	-13.9496	have | many
0.0767196	-13.9496	for | very
0.137931	-13.9496	objects | of
0.037271	-13.9496	there is | the
0.291667	-13.9497	al., | pang
0.077135	-13.9498	proceedings | annual | portland,
0.6	-13.9498	article | recent
0.0782123	-13.9498	an | translation
0.196429	-13.9499	next, | is
0.150538	-13.95	model | very
0.113095	-13.9501	the | holds
0.0249649	-13.9502	the | then
0.291667	-13.9502	function | w
0.173913	-13.9503	values | using
0.123288	-13.9504	i) | the
0.25	-13.9505	using | penn
0.0858086	-13.9505	discriminative | the
0.0487013	-13.9505	one | as
0.0336391	-13.9506	a | semantic
0.0607083	-13.9506	set | each
0.304348	-13.9508	textual | semantic
0.304348	-13.9508	change | at
0.5	-13.9508	operational | for
0.179104	-13.9509	the | sites
0.115854	-13.951	c©2011 | institute
0.258065	-13.951	items | same
0.051374	-13.9511	with | more
0.0707763	-13.9512	reference | and
0.275862	-13.9513	form | →
0.192982	-13.9513	to | entries.
0.352941	-13.9514	datasets | test
0.212766	-13.9514	but | improve
0.266667	-13.9516	study | performance
0.0798817	-13.9516	which | was
0.5	-13.9516	3 | 4 show
0.114458	-13.9516	20 | 60
0.0394089	-13.9517	a | set | a
0.304348	-13.9517	based | optimization
0.168831	-13.9518	for | speakers
0.164557	-13.952	these | similarity
0.0950413	-13.952	is | statistical
0.222222	-13.9521	d | number
0.0912548	-13.9521	edge | and
0.141509	-13.9522	filtering | a
0.5	-13.9523	twss | is
0.181818	-13.9523	by | classes
0.384615	-13.9524	measure | computed
0.0435154	-13.9524	previous | and
0.110497	-13.9526	variable | a
0.171053	-13.9527	all | p
0.318182	-13.9528	.28 | .27
0.318182	-13.9528	(zelle | mooney,
0.2	-13.9528	at | –
0.275862	-13.953	for | insert
0.184615	-13.953	selected | sentences
0.217391	-13.9531	have | weight
0.0515588	-13.9531	size | of
0.2	-13.9531	extraction | comparable
0.155556	-13.9532	reported | are
0.117647	-13.9532	by | distribution
0.142857	-13.9533	two | while
0.186441	-13.9533	we | heuristics
0.6	-13.9534	particular, | nearly
0.6	-13.9534	extends | state
0.6	-13.9534	p(t|w)
0.0855263	-13.9534	to | weights
0.184615	-13.9535	with | approach.
0.1625	-13.9536	degree | which
0.333333	-13.9536	used | character
0.236842	-13.9537	the | opponents
0.212766	-13.9537	work | patterns
0.0938776	-13.9537	call | the
0.110497	-13.9537	these | when
0.0779944	-13.9537	likely | a
0.5	-13.9538	introduced | new | task
0.2	-13.954	the | abstraction
0.179104	-13.9541	so | different
0.0912548	-13.9541	on | 1
0.384615	-13.9541	sketch | sketch
0.0640301	-13.9542	documents | to
0.6	-13.9543	mentions, | head
0.6	-13.9543	assigning | strings
0.6	-13.9543	type | premise
0.171053	-13.9545	we | treat
0.101449	-13.9546	the | srl
0.2	-13.9546	t1 | of
0.236842	-13.9547	one | representation
0.204082	-13.9547	problem | to be
0.184615	-13.9549	sentences | them
0.0931174	-13.955	to | structures
0.108696	-13.955	with | –
0.0731707	-13.9551	for | shown
0.150538	-13.9552	text | one
0.384615	-13.9553	are | ambiguous.
0.0859107	-13.9553	in | larger
0.181818	-13.9553	method | feature
0.114458	-13.9553	the | verb,
0.157303	-13.9554	these | systems
0.132231	-13.9554	input | an
0.0883392	-13.9556	is | identify
0.144231	-13.9556	adding | for
0.145631	-13.9557	using | search
0.0982143	-13.9557	l | s
0.444444	-13.9558	polysemous | polysemous
0.111765	-13.956	and | indicates
0.117647	-13.956	that | performed
0.0825397	-13.956	cluster | to
0.179104	-13.956	approach | some
0.10582	-13.9561	is | mt
0.133333	-13.9563	of | stack
0.444444	-13.9564	deictic | have
0.10989	-13.9566	of | f1
0.0803571	-13.9567	2 | 6
0.5	-13.9568	scores | reranking
0.212766	-13.9569	(2008) | from
0.384615	-13.957	“this | is
0.6	-13.957	after | scanning
0.444444	-13.9571	minimal | adjoining
0.444444	-13.9571	arguments | match
0.352941	-13.9572	include | alignment
0.166667	-13.9572	suggest | that
0.28	-13.9572	created | annotation
0.304348	-13.9572	active | supervision
0.243243	-13.9572	thank | us
0.186441	-13.9572	achieve | by
0.186441	-13.9572	using | larger
0.108696	-13.9573	system | results
0.196429	-13.9575	one | node
0.0745501	-13.9575	means | and
0.230769	-13.9575	two | reference
0.0738916	-13.9576	♣ | ♣
0.236842	-13.9576	entire | are
0.291667	-13.9576	where | made
0.5	-13.9576	any | example.
0.444444	-13.9578	bicycles | expensive
0.444444	-13.9578	ρkj(0) | esummationdisplay
0.111111	-13.9579	number | rules
0.6	-13.9579	linguists | their
0.266667	-13.958	which | otherwise
0.0470942	-13.9581	table | shows
0.230769	-13.9582	iteration, | to
0.352941	-13.9583	on | integration
0.236842	-13.9583	patterns | syntactic
0.0179248	-13.9584	the | is | that
0.444444	-13.9584	(zhao | 2010).
0.120805	-13.9585	of | 15
0.230769	-13.9585	means | some
0.0995475	-13.9587	each word | a
0.384615	-13.9588	w, | set
0.137931	-13.9589	briefly | and
0.230769	-13.9589	with | paraphrases
0.153846	-13.9591	papers | to
0.5	-13.9591	effectiveness | domain
0.266667	-13.9592	goes | to
0.230769	-13.9592	factor | by
0.155556	-13.9593	particular | on
0.384615	-13.9593	segment | points
0.291667	-13.9593	where | corresponds
0.0511905	-13.9594	that | these
0.0769231	-13.9594	also | with | and
0.106952	-13.9595	network | a
0.275862	-13.9596	in | chain.
0.0502283	-13.9597	data | with
0.0868056	-13.9597	an | use
0.6	-13.9597	treewidth | a
0.176471	-13.9598	features | head
0.291667	-13.9598	when | observed
0.10929	-13.9599	from | syntactic
0.333333	-13.9601	parser | projected
0.125926	-13.9603	for | metrics
0.12	-13.9603	english | which
0.352941	-13.9604	linear | has
0.352941	-13.9604	relevance | search
0.0665323	-13.9604	following | a | of
0.444444	-13.9604	∼ | α
0.141509	-13.9605	that | represented
0.0423453	-13.9606	and | target
0.0782123	-13.9606	in | total
0.110497	-13.9607	a | temporal
0.105263	-13.9607	no | between
0.416667	-13.9608	conversation | at
0.10929	-13.9609	in | 2005).
0.28	-13.961	a | test.
0.0545703	-13.961	large | to
0.258065	-13.961	on | sides
0.258065	-13.961	can | distributional
0.196429	-13.9611	for | similarity.
0.164557	-13.9611	two | rules
0.243243	-13.9613	although | task
0.416667	-13.9614	its | unique
0.0709382	-13.9614	another | a
0.147368	-13.9615	parameter | that
0.0762943	-13.9617	parsers | the
0.243243	-13.962	do | i
0.243243	-13.962	from | (and
0.0629496	-13.962	the | workshop
0.0934959	-13.962	the | comes
0.0132917	-13.9621	each | the
0.1625	-13.9621	his | for
0.141509	-13.9622	portland, | 24
0.225	-13.9623	describe | how
0.212766	-13.9623	difference | statistically
0.315789	-13.9623	function | large
0.6	-13.9624	type | environment.
0.444444	-13.9624	that | nuclear
0.173913	-13.9625	related | research
0.0419693	-13.9627	we | performance
0.0912548	-13.9628	used | corpus
0.147368	-13.9628	estimation | a
0.111111	-13.9628	entity | or
0.150538	-13.9629	think | a
0.5	-13.9629	which | nothing
0.176471	-13.963	and | bootstrapping
0.0882353	-13.963	the | β
0.117647	-13.9631	that | structures
0.416667	-13.9633	whole | into
0.25	-13.9633	on | involving
0.6	-13.9633	(2000) | while
0.243243	-13.9635	annotate | as
0.012638	-13.9635	for | and | the
0.5	-13.9636	results | bold.
0.153846	-13.9638	high | not
0.173913	-13.9639	using | 2008).
0.208333	-13.9639	estimate | word
0.204082	-13.9639	they | specific
0.116129	-13.964	detecting | of
0.384615	-13.964	top | age
0.0767123	-13.9642	2010; | et
0.6	-13.9642	robot | world
0.291667	-13.9642	actual | not
0.108108	-13.9643	the | relation,
0.0573152	-13.9645	a | (e.g.,
0.444444	-13.9645	but | induction
0.444444	-13.9645	duc | summarization
0.384615	-13.9646	unsupervised | problem
0.125926	-13.9647	score | using
0.053455	-13.9648	the | candidates
0.113095	-13.9648	in | (i.e.
0.0570571	-13.9649	verb | and
0.144231	-13.9649	the | speaking
0.0510689	-13.9651	for each | we
0.0286792	-13.9651	has | to
0.6	-13.9651	greek | italian
0.291667	-13.9651	experiments | report
0.0973451	-13.9652	in | (section
0.0562771	-13.9653	test | for
0.0652174	-13.9653	of | random
0.0273785	-13.9654	in | are | and
0.160494	-13.9655	it | noted
0.0549451	-13.9655	in | possible
0.0230415	-13.9656	however, | the
0.28	-13.9657	partial | over
0.186441	-13.9657	parsing | all
0.189655	-13.9657	preposition | with
0.0391608	-13.9658	in | the same
0.444444	-13.9658	building | work
0.0591054	-13.9659	standard | to
0.0789474	-13.966	extraction | for
0.0931174	-13.966	that | role
0.0844156	-13.966	usually | of
0.166667	-13.9662	obtained | data
0.158537	-13.9662	time | it
0.0312919	-13.9662	similarity | the
0.0535248	-13.9662	that | some
0.304348	-13.9664	integrated | language
0.147368	-13.9664	for | (i.e.
0.204082	-13.9664	work | identifying
0.124088	-13.9665	parsing | from
0.160494	-13.9665	for | (a
0.315789	-13.9666	their | before
0.092	-13.9667	similar | al.,
0.304348	-13.9668	were | via
0.291667	-13.9668	because | contains
0.0577508	-13.9669	word | or
0.0726392	-13.9671	the | log
0.212766	-13.9671	the | participle
0.416667	-13.9671	multilingual | has
0.138889	-13.9671	problems | as
0.444444	-13.9672	correlates | than
0.168831	-13.9673	some | into
0.113095	-13.9673	different | training
0.225	-13.9674	larger | can
0.117647	-13.9675	find | by
0.384615	-13.9675	senses | type
0.0701357	-13.9675	i | for
0.266667	-13.9676	studies | language
0.0365186	-13.9676	candidate | the
0.2	-13.9676	wordnet | or
0.0865052	-13.9677	are | by | with
0.333333	-13.9677	methods | computing
0.444444	-13.9678	explicit | structure
0.0747423	-13.9679	w | a
0.6	-13.9679	|v | |
0.6	-13.9679	heuristics | mwes
0.315789	-13.9681	has | community.
0.0964912	-13.9681	even | that
0.352941	-13.9684	characters | they
0.416667	-13.9684	identifiers | pos
0.444444	-13.9685	noise | speech
0.0712589	-13.9685	are | any
0.291667	-13.9686	distributed | with
0.106952	-13.9688	kernel | and
0.135593	-13.9688	the | big
0.6	-13.9688	without | symbol
0.150538	-13.9688	sets | which
0.416667	-13.969	collins, | mooney,
0.416667	-13.969	ce | which
0.171429	-13.9691	for | gain
0.153846	-13.9691	by | factor
0.0469388	-13.9691	model | from
0.101942	-13.9693	for | significant
0.131148	-13.9694	in | lm
0.6	-13.9697	next | again
0.6	-13.9697	0.36 | bleu
0.6	-13.9697	it | fail.
0.125	-13.9697	and | combining
0.186441	-13.9697	others | a
0.5	-13.9697	some | perhaps
0.444444	-13.9699	transition-based | parse
0.176471	-13.97	data | contains
0.164557	-13.9701	for | —
0.0847458	-13.9703	the | aspect
0.0588235	-13.9703	phrase | is
0.384615	-13.9704	probabilities | hypotheses
0.166667	-13.9705	other | 2010).
0.6	-13.9706	considered | sentences.
0.0635514	-13.9707	the | definition
0.333333	-13.9707	frequency | size
0.416667	-13.9709	target | (i.e.,
0.0519802	-13.9709	the | in | set.
0.384615	-13.971	specific | certain
0.113095	-13.971	to | alternative
0.0865052	-13.9711	with | english
0.352941	-13.9711	propose | joint
0.156627	-13.9711	two | no
0.158537	-13.9712	for | especially
0.142857	-13.9712	article | as
0.5	-13.9712	conditional | joint
0.219512	-13.9713	a | exponential
0.101449	-13.9716	the | multilingual
0.0380254	-13.9716	model | as
0.153846	-13.9717	each | w
0.083871	-13.9719	probabilistic | of
0.0758808	-13.9719	the | well.
0.0440647	-13.9719	the | words | are
0.186441	-13.9719	this | dynamic
0.258065	-13.972	that | relate
0.5	-13.972	for | (xue
0.5	-13.972	result, | obtain
0.12	-13.9721	model | most
0.0647059	-13.9723	and | extracted
0.6	-13.9724	v | hyperedges.
0.6	-13.9724	combining | phrase-based
0.28	-13.9725	development | annotation
0.141509	-13.9725	we use | feature
0.110465	-13.9726	automatic | on
0.212766	-13.9726	accuracy | dependency
0.0756757	-13.9726	with | performance
0.0330496	-13.9726	they | to
0.416667	-13.9728	(ruppenhofer | al.,
0.186441	-13.9728	on | highly
0.266667	-13.9729	hierarchical | models
0.0902256	-13.9729	source | from
0.0859107	-13.9729	to | coreference
0.0502924	-13.9731	in | no
0.204082	-13.9732	outputs | are
0.109827	-13.9733	we | weighted
0.444444	-13.9733	any | otherwise
0.217391	-13.9735	possible | over
0.135593	-13.9735	an | following
0.129032	-13.9736	like | by
0.108696	-13.9736	illustrated | the
0.352941	-13.9738	same | tend
0.384615	-13.9739	therefore, | dependency
0.156627	-13.974	liu | al.,
0.0597015	-13.9741	different | that
0.179104	-13.9742	and | coherence
0.0478723	-13.9742	the | see
0.0294118	-13.9743	are | on
0.304348	-13.9743	huang | al.
0.5	-13.9743	experiment | effectiveness
0.104712	-13.9744	pos | word
0.0960699	-13.9745	models, | to
0.28	-13.9746	was | annotator
0.28	-13.9746	2 | collection
0.0741688	-13.9747	we | domain
0.208333	-13.9747	visual | is
0.0927419	-13.9748	are | standard
0.225	-13.9749	task | annotation
0.150538	-13.975	pair | phrase
0.113772	-13.9751	events | as
0.0964912	-13.9752	when | all
0.6	-13.9752	shown | translating
0.123188	-13.9752	using | set.
0.225	-13.9753	ways | are
0.0956522	-13.9753	features | two
0.266667	-13.9754	parsers | can
0.130081	-13.9754	(3) | for
0.0679406	-13.9754	for example, | be
0.125	-13.9754	for | mentions
0.192982	-13.9755	the | scripts
0.217391	-13.9755	work | uses
0.0923695	-13.9755	data, | is
0.0872727	-13.9755	but | this
0.134454	-13.9755	do | at
0.144231	-13.9755	reasons | of
0.158537	-13.9756	languages | by
0.28	-13.9759	system | useful
0.28	-13.9759	make | parsing
0.242424	-13.9761	this | demonstrated
0.136364	-13.9762	in | he
0.169014	-13.9763	to | dialect
0.183333	-13.9763	we | zero
0.28	-13.9764	compound | its
0.25	-13.9766	english | extracted
0.315789	-13.9767	runs | only
0.176471	-13.9767	should | other
0.0341834	-13.977	they | and
0.6	-13.977	mi | 2008).
0.125926	-13.9771	of | in | section.
0.258065	-13.9771	u | l
0.444444	-13.9773	opinion | (or | opinion
0.5	-13.9773	system | source-context
0.266667	-13.9774	for | 1,000
0.333333	-13.9774	factored | language
0.304348	-13.9775	of | metaphorical
0.352941	-13.9776	have | errors.
0.0373711	-13.9776	their | to
0.28	-13.9776	can | valid
0.112426	-13.9778	temporal | to
0.0690423	-13.9779	variable | the
0.352941	-13.9781	as | person,
0.352941	-13.9781	states | user
0.25	-13.9781	cannot | their
0.0859107	-13.9781	• | from
0.212766	-13.9782	compared | best
0.0606586	-13.9782	the | of | tags
0.179104	-13.9783	adjectives | as
0.304348	-13.9784	when | individual
0.333333	-13.9784	as | string,
0.291667	-13.9784	effective | when
0.219512	-13.9784	however, | improvement
0.0687361	-13.9785	computed | and
0.145833	-13.9785	out | were
0.25	-13.9785	shows | relative
0.119205	-13.9787	can be | as | =
0.0336842	-13.9787	set of | and
0.444444	-13.9787	words, | part-of-speech
0.219512	-13.9787	strategies | from
0.171429	-13.9788	general | such
0.25	-13.9789	an | otherwise
0.230769	-13.9789	this | quality.
0.5	-13.9789	now | describe
0.137615	-13.979	a | estimation
0.0794118	-13.9791	to | graph
0.352941	-13.9792	human | asked
0.416667	-13.9792	as | concerned,
0.0895522	-13.9792	disambiguation | and
0.0682819	-13.9795	abstract | in
0.258065	-13.9795	set | pairs,
0.104167	-13.9797	of | attribute
0.25	-13.9797	4 | respectively.
0.0747423	-13.9797	their | by
0.0308506	-13.9798	for | word
0.384615	-13.9798	class | .
0.208333	-13.9798	with | relations.
0.6	-13.9798	propagation | propagate
0.6	-13.9798	person | org
0.291667	-13.9798	how | extract
0.304348	-13.9799	improves | when
0.258065	-13.9799	4 | 1.
0.0865052	-13.98	randomly | and
0.173913	-13.98	that | noise
0.0189621	-13.98	in | in | in
0.103627	-13.9802	to | task,
0.111111	-13.9802	supported | research
0.352941	-13.9803	since | end
0.106383	-13.9803	shown | 2.
0.225	-13.9804	we | φ
0.102439	-13.9805	in | token
0.141509	-13.9805	of | involves
0.0588235	-13.9805	and | words.
0.230769	-13.9806	epi | and
0.6	-13.9807	total | (in
0.444444	-13.9807	first, | obtain
0.304348	-13.9808	having | number
0.0956522	-13.9809	did not | of
0.0853242	-13.981	concept | of
0.315789	-13.9811	since | mentions
0.138889	-13.9811	the | minimize
0.25	-13.9812	scores | while
0.144231	-13.9812	close | a
0.137615	-13.9814	studies | for
0.444444	-13.9814	lda | models.
0.28	-13.9815	or | indicate
0.333333	-13.9815	model | plate
0.126866	-13.9815	constraints | be
0.6	-13.9816	(np | (sbar
0.6	-13.9816	2.3 | distribution
0.204082	-13.9817	selectional | and
0.125926	-13.9818	to | levels
0.2	-13.9819	training | baseline
0.258065	-13.9819	wmt | for
0.444444	-13.9821	2008; | chen
0.444444	-13.9821	de | stanford
0.0227436	-13.9822	we | a | for
0.0386507	-13.9822	is | other
0.2	-13.9822	number | questions
0.0791789	-13.9823	as | would
0.258065	-13.9823	true | or
0.0323371	-13.9824	in | different
0.133333	-13.9825	how | each
0.0459433	-13.9826	standard | of
0.141509	-13.9827	we | a list
0.0787172	-13.9827	provide | in
0.5	-13.9827	it | visually
0.28	-13.9828	network | random
0.384615	-13.9828	cause | between
0.352941	-13.983	might | any
0.147368	-13.9831	made | by
0.315789	-13.9835	features | arabic
0.6	-13.9835	disambiguation | difficult
0.6	-13.9835	column | highest
0.0700935	-13.9835	to | query
0.5	-13.9835	partial | combination
0.352941	-13.9836	1 | minimal
0.134454	-13.9836	is | person
0.150538	-13.9836	dialogue | system
0.105263	-13.9837	manually | for
0.130081	-13.9838	documents | or
0.0589198	-13.9839	first | as
0.129032	-13.9839	strategy | for
0.0420792	-13.984	in | source
0.333333	-13.984	allow | user
0.28	-13.9841	9 | all
0.118421	-13.9841	automatic | from
0.092	-13.9842	for example, | by
0.0695067	-13.9842	d | a
0.204082	-13.9842	3.2 | features
0.291667	-13.9842	können:can | morgen:tomorrow
0.11465	-13.9843	to | efficient
0.189655	-13.9843	annotated | information
0.2	-13.9844	learner | by
0.179104	-13.9846	1 introduction | of | has
0.2	-13.9847	complexity | by
0.0756757	-13.9849	cluster | of
0.153846	-13.9849	tags | word
0.0439856	-13.985	grammar | the
0.291667	-13.9851	as an | over
0.0809969	-13.9851	above, | the
0.136364	-13.9854	use | evaluation
0.304348	-13.9855	however, | appear
0.444444	-13.9855	unigram | learned
0.148936	-13.9857	found | was
0.28	-13.9858	then | hypothesis
0.384615	-13.9858	standard | phrase-based
0.384615	-13.9858	constraints | single
0.196078	-13.9858	to | segmentation,
0.136364	-13.986	for | bigram
0.230769	-13.986	current | set
0.118421	-13.9863	 | of
0.0215827	-13.9864	which | a
0.133333	-13.9866	1 | both
0.352941	-13.9868	of | comunication
0.0333681	-13.9869	the | noun
0.444444	-13.9869	from | predictions.
0.258065	-13.9871	designed | language
0.6	-13.9872	72 | model
0.0504202	-13.9873	the | larger
0.180328	-13.9873	for | t,
0.166667	-13.9874	section | details
0.156627	-13.9875	results | even
0.208333	-13.9875	many | where
0.134454	-13.9875	first, | be
0.444444	-13.9876	develop | method
0.046371	-13.9877	to | best
0.158537	-13.9881	systems | all
0.6	-13.9881	analyzed | relevant
0.236842	-13.9883	even | sentence
0.304348	-13.9883	expressions: | real
0.138889	-13.9884	will | its
0.145833	-13.9884	particular, | are
0.085034	-13.9884	the | parses
0.117647	-13.9885	in | frame
0.230769	-13.9885	learnt | to
0.180328	-13.9887	one | might
0.416667	-13.9888	learned | (e.g.,
0.0467005	-13.9888	a | context
0.142857	-13.9888	used | standard
0.219512	-13.9889	an | tool
0.169014	-13.989	set | manually
0.444444	-13.989	too | (i.e.
0.0742706	-13.989	what | in
0.0537931	-13.9892	left | the
0.28	-13.9893	finally, | present
0.042735	-13.9893	the | overall
0.212766	-13.9894	are | variable
0.152174	-13.9894	have | tasks
0.196078	-13.9895	k1 | the
0.196078	-13.9895	each | side
0.5	-13.9897	1993; | 2003;
0.158537	-13.9898	40 | 100
0.0603448	-13.9898	we | a | system
0.108571	-13.9898	towards | and
0.384615	-13.9899	# | *.
0.154762	-13.99	pos | using
0.416667	-13.9901	define | argmax
0.333333	-13.9902	systems | about
0.120567	-13.9903	at | when
0.147368	-13.9904	so | more
0.5	-13.9904	dificulty | word
0.416667	-13.9908	gold | auto
0.6	-13.9909	suggests | combining
0.6	-13.9909	g. | a.
0.266667	-13.991	for | middle
0.269231	-13.991	between | indicate
0.0869565	-13.9912	can | over
0.192982	-13.9913	user | different
0.0856164	-13.9913	each | their
0.333333	-13.9913	tokens | most
0.138889	-13.9914	as | low
0.269231	-13.9914	more | initial
0.0948276	-13.9914	second | on
0.208333	-13.9914	most | errors
0.0932203	-13.9915	the | strongly
0.186441	-13.9915	all | particular
0.14433	-13.9916	in | distribution.
0.0503597	-13.9916	all | on
0.384615	-13.9917	related | studied
0.0588235	-13.9917	found | in
0.2	-13.9918	5.1 | et
0.6	-13.9919	list | called
0.6	-13.9919	maybe | it
0.124088	-13.992	italian | the
0.0453668	-13.992	of | input
0.113208	-13.9921	turn | in
0.192982	-13.9922	yang | and
0.145833	-13.9922	by | against
0.0555556	-13.9925	best | for
0.104167	-13.9926	mention | is
0.083612	-13.9926	errors | a
0.180328	-13.9926	the | passage
0.101449	-13.9927	with | time
0.333333	-13.9928	shown | equation
0.6	-13.9928	z, | x
0.291667	-13.9928	& | number
0.5	-13.9928	1 | 250
0.384615	-13.9929	its | far
0.357143	-13.9931	assigned | weight
0.189655	-13.9931	just | are
0.0990566	-13.9933	once | to
0.107955	-13.9933	and | introduced
0.120567	-13.9933	comparative | and
0.102041	-13.9934	example | an
0.173913	-13.9935	factors | that
0.0956522	-13.9935	parameter | to
0.5	-13.9936	methodology | applied
0.5	-13.9936	ability | local
0.6	-13.9937	text | routed
0.0270645	-13.9937	is | by | of
0.158537	-13.994	their | these
0.0611511	-13.994	in | verb
0.242424	-13.9941	null | null3
0.0338164	-13.9941	with | from
0.0563166	-13.9942	a | how
0.357143	-13.9942	identify | within
0.304348	-13.9945	model | forward
0.304348	-13.9945	top | class
0.0319613	-13.9945	the | very
0.04814	-13.9945	a | english
0.0815047	-13.9946	are | syntactic
0.189655	-13.9946	i | do
0.6	-13.9947	pivot | phrases
0.136364	-13.9947	constraint | is
0.103627	-13.9949	2001; | al.,
0.108571	-13.995	is | predict
0.171429	-13.9952	an | word.
0.137615	-13.9952	one | may
0.444444	-13.9952	labeled | points.
0.0691244	-13.9953	for | dialogue
0.160494	-13.9956	to make | we
0.154762	-13.9958	an | linear
0.357143	-13.9959	results | wsd
0.5	-13.9959	observed | twitter
0.416667	-13.996	mentioned | some
0.0646465	-13.996	model | this
0.225	-13.9961	+ | denotes
0.0807453	-13.9961	to | texts
0.0730479	-13.9962	difference | in
0.28	-13.9963	|x
0.118881	-13.9964	its | word
0.117647	-13.9965	for | testing
0.242424	-13.9967	often | semantic
0.176471	-13.9967	selecting | as
0.103093	-13.9967	not | well
0.0285604	-13.9967	a | one
0.0866426	-13.9969	in | cluster
0.333333	-13.997	many | considered
0.183333	-13.997	and | database.
0.0769231	-13.9971	and | state
0.0479303	-13.9973	the | news
0.147368	-13.9973	distance | two
0.130081	-13.9974	over | training
0.291667	-13.9974	improvement | approach
0.103093	-13.9975	on | specific
0.0809969	-13.9975	use | training
0.11465	-13.9976	considered | as
0.106383	-13.9976	multiple | are
0.384615	-13.9977	set | outgoing
0.119718	-13.9978	bleu | system
0.14433	-13.9979	3 | were
0.225	-13.9979	while | found
0.0206771	-13.9979	of | with | of
0.444444	-13.998	final | matching
0.444444	-13.998	y | z.
0.0764873	-13.9981	for | s
0.0728643	-13.9982	syntactic | for
0.180328	-13.9982	3 | order
0.357143	-13.9982	tasks | labeling
0.357143	-13.9982	split | cluster
0.125984	-13.9983	improvements | for
0.0838926	-13.9983	in | above
0.352941	-13.9984	kyoto | japan
0.0696056	-13.9984	show | with
0.6	-13.9984	bayesian | homophonic
0.158537	-13.9987	and | (cf.
0.119718	-13.9987	results | also
0.192982	-13.9988	work, | which
0.196078	-13.999	proposed | words
0.333333	-13.9991	pr | c
0.152941	-13.9992	and | denoted
0.140187	-13.9993	word | tag
0.126984	-13.9993	case | from
0.177419	-13.9993	distribution | data
0.092	-13.9994	required | of
0.6	-13.9994	michael | comments.
0.444444	-13.9994	ranker | scores
0.28	-13.9998	statistical | parsing
0.269231	-13.9998	using | pruning
0.0756757	-13.9998	up | a
0.177419	-13.9999	details | are
0.177419	-13.9999	second, | from
0.111111	-14.0001	a | regression
0.118881	-14.0001	shown | an
0.177419	-14.0001	length | all
0.0688073	-14.0001	that | single
0.186441	-14.0002	contributions | this
0.6	-14.0003	> | g
0.130081	-14.0004	study | as
0.242424	-14.0005	computational | dan
0.132231	-14.0007	with | total
0.384615	-14.0007	consider | instance
0.384615	-14.0007	probabilities | likelihood
0.150538	-14.0007	for | highly
0.208333	-14.0008	better | use
0.059727	-14.0009	• | for
0.357143	-14.001	words | contexts,
0.352941	-14.0011	their | feedback.
0.416667	-14.0012	identified | per
0.6	-14.0012	high | speech.
0.25	-14.0014	consists | target
0.384615	-14.0014	standard | compare
0.189655	-14.0014	a | utterance.
0.103627	-14.0015	2008), | et
0.2	-14.0015	process | one
0.291667	-14.0015	class | argument
0.180328	-14.0016	from | b
0.0995261	-14.0016	move | the
0.315789	-14.0017	the | figures.
0.333333	-14.0017	any | make
0.186441	-14.0017	on | 2002).
0.136364	-14.0018	2009), | in
0.2	-14.0019	non-terminals | the
0.0960699	-14.002	cannot | in
0.0573248	-14.002	main | and
0.6	-14.0022	specific | content.
0.6	-14.0022	precision | fraction
0.230769	-14.0022	conditions | be
0.106742	-14.0024	believe | is
0.225	-14.0025	nearest | in
0.145833	-14.0025	beginning | a
0.0782609	-14.0025	for | specific
0.0187427	-14.0025	the | features
0.384615	-14.0026	restricted | noun
0.357143	-14.0026	ranking | extracted
0.219512	-14.0026	set | annotations
0.109827	-14.0028	unsupervised | is
0.180328	-14.0028	indicate | by
0.269231	-14.0028	articles | they
0.208333	-14.0028	observed | between
0.0901961	-14.0029	we | term
0.291667	-14.0029	tree | derivation
0.137615	-14.003	posterior | and
0.137615	-14.003	to | posterior
0.5	-14.003	with | inc.
0.00769424	-14.0031	the | of | to
0.6	-14.0031	links | pages.
0.0552326	-14.0032	our | set
0.133929	-14.0032	data | its
0.131148	-14.0033	in | interpretation
0.173913	-14.0036	of | coherent
0.0884615	-14.0037	probability | with
0.0620438	-14.0037	human | in
0.147368	-14.0037	because | other
0.132743	-14.0037	current | on
0.416667	-14.0038	whose | according
0.416667	-14.0038	content | parameters
0.291667	-14.0038	defined | |
0.115385	-14.0039	an | dialogue
0.138889	-14.004	however, | information
0.0691244	-14.004	what | to
0.0857143	-14.0041	we | per
0.147368	-14.0042	noted | is
0.333333	-14.0043	with | monotone
0.444444	-14.0043	a | 0.8.
0.0736842	-14.0044	if | are | to
0.384615	-14.0044	techniques | sentiment
0.235294	-14.0044	errors, | are
0.196078	-14.0045	patterns | an
0.142857	-14.0049	learning | each
0.357143	-14.0049	had | when
0.416667	-14.0051	assigned | higher
0.208333	-14.0051	learner | on
0.0932203	-14.0052	strings | the
0.0780347	-14.0052	recall | in
0.214286	-14.0052	data | make
0.180328	-14.0053	recall, | in
0.333333	-14.0053	in | targets.
0.164384	-14.0054	in | graph-based
0.106742	-14.0054	we | implementation
0.0485327	-14.0055	analysis | in
0.111801	-14.0055	was | other
0.315789	-14.0057	turn | time
0.196078	-14.0058	this | aspect
0.0604982	-14.0058	we | single
0.166667	-14.0059	s | which
0.242424	-14.0059	based | part
0.6	-14.006	tsuruoka | al.
0.0901961	-14.0061	estimated | of
0.242424	-14.0063	teacher | to
0.158537	-14.0064	set | p
0.196078	-14.0064	applied | dependency
0.154762	-14.0065	benefit | feature
0.103093	-14.0066	we | estimate
0.384615	-14.0068	parsing | text,
0.0694444	-14.0069	words | words | and
0.129032	-14.0071	many | these
0.357143	-14.0071	without | pairs
0.444444	-14.0071	system | 54.9
0.315789	-14.0072	applications | semantic
0.186441	-14.0072	uses | training
0.214286	-14.0072	= | i.e.
0.137615	-14.0073	performance | trained
0.130081	-14.0074	do | necessarily
0.384615	-14.0074	1 | approximate
0.0744681	-14.0077	used | data
0.126984	-14.0077	of | nn
0.0854093	-14.0078	- | in
0.0511222	-14.0078	any | to
0.107955	-14.0078	has | over
0.219512	-14.0078	of | participants.
0.104167	-14.0079	and | of | text.
0.333333	-14.008	entry | one
0.0792683	-14.008	the | basis
0.00770882	-14.0082	in | the | of
0.269231	-14.0083	would | difficult
0.416667	-14.0084	la | les
0.0705596	-14.0084	svm | the
0.180328	-14.0085	are | hierarchical
0.105263	-14.0085	lexical | which
0.133929	-14.0085	to | biomedical
0.0641283	-14.0086	experiments | on
0.101523	-14.0086	and | summary
0.384615	-14.0086	90% | on
0.230769	-14.0087	experiment | language
0.6	-14.0088	0.00 | 0.60
0.6	-14.0088	cn | bcn
0.171429	-14.009	0.2 | precision
0.0660981	-14.0092	as | text
0.164384	-14.0093	graph, | and
0.357143	-14.0094	find | strong
0.107955	-14.0097	outside | the
0.0869565	-14.0097	relationship | and
0.444444	-14.0099	49 | 50
0.5	-14.0101	jari | tapio
0.5	-14.0101	large | gigaword
0.5	-14.0101	(2009), | was
0.315789	-14.0102	a | image.
0.291667	-14.0103	property | if
0.166667	-14.0104	show | significant
0.0355208	-14.0104	by | from
0.106742	-14.0105	input | by
0.0710784	-14.0106	representing | the
0.0627376	-14.0106	to | sentiment
0.315789	-14.0107	function | 3
0.0948276	-14.0107	c©2011 association | science
0.14433	-14.0108	moreover, | be
0.122302	-14.0108	pages | 24
0.28	-14.0108	bleu | target
0.141414	-14.0111	3 | evaluation
0.186441	-14.0111	performance | due
0.162162	-14.0112	while | both
0.315789	-14.0112	st | uf
0.0694444	-14.0113	designed | the
0.204082	-14.0113	her | in
0.108571	-14.0114	of | necessary
0.0485327	-14.0115	not | it
0.5	-14.0116	incremental | recognition
0.5	-14.0116	bears | from
0.154762	-14.0117	4 | models
0.171429	-14.0117	both | given
0.357143	-14.0117	or | preferences
0.132743	-14.0117	to | google
0.106145	-14.0117	performance | training
0.183333	-14.0117	general, | are
0.0848057	-14.0118	select | and
0.177419	-14.0119	are | far
0.444444	-14.012	data, | see
0.357143	-14.0122	due | sparsity
0.0752089	-14.0124	for | alignment
0.147368	-14.0124	was | any
0.113208	-14.0124	that | position
0.333333	-14.0127	merge | as
0.333333	-14.0127	graph | summary
0.128	-14.0127	d | n
0.444444	-14.0127	if | said
0.214286	-14.0127	features | slightly
0.384615	-14.0129	social | enron
0.416667	-14.013	by | questions.
0.269231	-14.013	analysis | improve
0.0848057	-14.013	indicator | the
0.235294	-14.013	used | texts
0.145833	-14.0131	recently | to
0.0827815	-14.0132	a | sentence,
0.0327031	-14.0132	that | are | to
0.269231	-14.0134	if | true
0.444444	-14.0134	composite | its
0.0316395	-14.0134	two | of | and
0.28	-14.0135	different | ranking
0.183333	-14.0135	based | extracted
0.101523	-14.0136	are | correct
0.133929	-14.0136	for | entailment
0.0891473	-14.0138	probability | word
0.225	-14.0139	φ | θ
0.0963303	-14.0139	in | simply
0.145833	-14.014	modeling | with
0.0232739	-14.014	with | for
0.103627	-14.0141	further | as
0.384615	-14.0141	found | effective
0.384615	-14.0141	work | broad
0.235294	-14.0141	included | one
0.192308	-14.0142	words | thus
0.0411184	-14.0143	3 | a
0.128	-14.0143	section | describes the
0.102564	-14.0144	this | sense
0.064	-14.0145	in | even
0.357143	-14.0145	obtained | combined
0.6	-14.0145	larger | indicating
0.103627	-14.0146	query | for
0.102041	-14.0146	resources | for
0.258065	-14.0146	support | it
0.315789	-14.0147	extraction | very
0.384615	-14.0147	event | action
0.444444	-14.0148	was | united
0.444444	-14.0148	(su | al.,
0.174603	-14.0148	first, | model
0.0297619	-14.0149	by | with
0.092827	-14.015	second | with
0.0700483	-14.0151	in | across
0.3	-14.0151	et | 2009a),
0.0371107	-14.0152	for | words
0.0981308	-14.0152	as | form
0.113924	-14.0153	is | various
0.0811688	-14.0154	shown | are
0.0423143	-14.0154	the | shows
0.120567	-14.0155	of | s.
0.183333	-14.0155	d | h
0.104167	-14.0157	to | approach.
0.416667	-14.0157	there | been | efforts
0.0857143	-14.0157	is | after
0.162162	-14.0159	much | this
0.132743	-14.0159	arguments | are
0.0626186	-14.016	e | e | t
0.092827	-14.0161	these | different
0.315789	-14.0162	on | broad
0.0273185	-14.0162	phrase | the
0.164384	-14.0163	letter | a
0.6	-14.0164	distinguish | cases.
0.230769	-14.0164	statistics | training
0.384615	-14.0166	focus | form
0.102041	-14.0167	for | -
0.0454097	-14.0167	in | new
0.258065	-14.0167	gains | on
0.291667	-14.0168	content | documents
0.056872	-14.0169	show | for
0.148936	-14.0169	with | version
0.3	-14.017	similar | 2010),
0.131579	-14.0171	we | the | above.
0.0698795	-14.0172	a | reference
0.1125	-14.0172	results | bleu
0.124031	-14.0173	lists | to
0.116438	-14.0174	a | starting
0.142857	-14.0174	many | been
0.2	-14.0174	each | full
0.192308	-14.0175	of | role.
0.28	-14.0175	4 | annotations
0.333333	-14.0175	alignment | models.
0.136364	-14.0177	accuracy | number
0.0727273	-14.0177	• | are
0.107345	-14.0178	are | contain
0.3	-14.0179	as | aac
0.5	-14.018	weighted | log
0.138889	-14.0181	since | information
0.0282508	-14.0181	or | in
0.109195	-14.0182	computational linguistics, | oregon, june
0.166667	-14.0184	also | possible
0.0803859	-14.0184	which | were
0.186441	-14.0185	accomplished | the
0.0967742	-14.0188	is | q
0.048951	-14.0188	in | syntactic
0.177419	-14.019	pairs | corpus.
0.291667	-14.0191	here, | features
0.0825083	-14.0194	in | us
0.196078	-14.0195	those | using
0.0508055	-14.0196	the | comparable
0.0658174	-14.0196	a | previous
0.092827	-14.0196	mean | in
0.258065	-14.0196	system | baseline.
0.5	-14.0196	provided | total
0.416667	-14.0197	s101 | s110
0.416667	-14.0197	palestinian | bush
0.416667	-14.0197	(if | (exist
0.14433	-14.0198	with | relevant
0.444444	-14.0198	opinion | reviews
0.444444	-14.0198	list | mentions
0.0747922	-14.0198	relationship | the
0.291667	-14.02	= | xi
0.064257	-14.0201	quality | and
0.0219096	-14.0201	the | a | or
0.214286	-14.0202	(liu | et
0.0536723	-14.0203	of | sets
0.110429	-14.0203	sense | word
0.209302	-14.0204	of | failure
0.0211617	-14.0205	the | were | the
0.154762	-14.0205	it | directly
0.137615	-14.0205	on | a single
0.269231	-14.0207	annotations | were
0.269231	-14.0207	to | scenarios.
0.333333	-14.0207	lexical | contextual
0.315789	-14.0208	database | words
0.235294	-14.0209	attempts | by
0.107955	-14.0211	2009), | et
0.6	-14.0212	as: | productdisplay
0.6	-14.0212	computed | counting
0.106742	-14.0212	equation | a
0.156627	-14.0213	a | meanings
0.0604982	-14.0213	in | noun
0.158537	-14.0215	community | a
0.133929	-14.0215	as | categories
0.164384	-14.0216	filter | that
0.128	-14.0216	to | n-grams
0.128	-14.0216	which | possible
0.052774	-14.0216	framework | the
0.230769	-14.0216	simply | it
0.242424	-14.0217	but | fact
0.204082	-14.0217	this | easy
0.235294	-14.022	stop | is
0.25	-14.0221	inference | sampling
0.0426667	-14.0222	the | models.
0.214286	-14.0222	m | h
0.0839161	-14.0223	system | this
0.209302	-14.0224	d7 | bh
0.0916667	-14.0225	each | score
0.125984	-14.0227	our | improvement
0.444444	-14.0227	94 | 10
0.158537	-14.0228	each | computed
0.0202985	-14.0228	is | by | the
0.0760563	-14.0229	document | for
0.103093	-14.023	the | varying
0.101523	-14.0231	of | evaluation.
0.126984	-14.0231	an | structure
0.106742	-14.0231	3.1 | is
0.6	-14.0232	almost | per
0.315789	-14.0233	translation | derivation
0.333333	-14.0234	summarizes | on
0.0995025	-14.0236	< | and
0.5	-14.0236	referring | objects
0.183333	-14.0238	of | disambiguation.
0.156627	-14.024	overlapping | and
0.188679	-14.024	increases | with
0.141414	-14.0241	represent | with
0.0599647	-14.0241	any | is
0.444444	-14.0241	is | paraphrase.
0.0877863	-14.0244	6 | that
0.0990099	-14.0245	we | token
0.133929	-14.0245	our | words.
0.0771513	-14.0246	features | all
0.315789	-14.0248	during | dialogue
0.357143	-14.0248	since | factors
0.214286	-14.025	1 | <
0.0614525	-14.0251	for | event
0.2	-14.0251	strongly | that
0.186441	-14.0251	this | explicitly
0.137615	-14.0255	section | concludes
0.0758427	-14.0256	language | translation
0.0163594	-14.0259	the | of | which
0.0703883	-14.0259	of | question
0.0954545	-14.026	we | study
0.166667	-14.0261	publication | the
0.0766962	-14.0261	which | into
0.6	-14.0261	strategy | strategies.
0.045	-14.0262	the | hypothesis | the
0.444444	-14.0262	5 | among
0.0857143	-14.0263	large | that
0.0209985	-14.0264	the | feature
0.225	-14.0265	this | tested
0.104396	-14.0265	the | contribute
0.151163	-14.0266	model | higher
0.0516796	-14.0267	experiments | a
0.0593368	-14.0267	the | polarity
0.219512	-14.0267	they | make
0.315789	-14.0269	temporal | text
0.0795107	-14.027	tagged | the
0.6	-14.027	λ, | λ
0.291667	-14.027	we | metadata
0.0954545	-14.0271	bound | the
0.188679	-14.0271	lattices | and
0.384615	-14.0271	combination | methods
0.0355828	-14.0272	we | results
0.0739726	-14.0273	they are | of
0.25	-14.0273	may | important
0.115646	-14.0275	for | (i.e.,
0.177419	-14.0275	with | equal
0.186441	-14.0275	one | user
0.164384	-14.0276	are | (i.e.
0.0912863	-14.0276	to | vectors
0.0920502	-14.0276	in | segmentation
0.5	-14.0276	french | japanese
0.444444	-14.0277	how | collect
0.204082	-14.0277	not | complete
0.136364	-14.0278	queries | for
0.0725389	-14.0278	translations | and
0.103093	-14.028	(2008) | to
0.235294	-14.0281	one | level
0.0284585	-14.0283	the | of | has
0.0592334	-14.0284	opinion | and
0.0493976	-14.0286	the | low
0.14	-14.0287	allows | to | with
0.0603908	-14.0287	simple | a
0.186441	-14.0287	mean | each
0.28	-14.0288	1 introduction | lexical
0.0743802	-14.0288	our | these
0.156627	-14.0288	question | can
0.6	-14.029	npk | has
0.136364	-14.0291	often | not
0.0700483	-14.0292	for all | and
0.0672646	-14.0292	the | parsers
0.192308	-14.0293	and | collins
0.5	-14.0293	estimated | algorithm.
0.357143	-14.0294	clusters | high
0.384615	-14.0295	last | given
0.166667	-14.0296	by | match
0.0205592	-14.0296	we | for | and
0.0884615	-14.0298	and | points
0.174603	-14.0299	based | proposed
0.123077	-14.0299	on | lower
0.183333	-14.03	theoretical | in
0.28	-14.0301	or | group
0.0877863	-14.0302	it | most
0.149425	-14.0302	al., | liu
0.209302	-14.0302	et | lin
0.384615	-14.0302	for each | phrase,
0.0416313	-14.0305	in | sentences
0.444444	-14.0305	all | users.
0.164384	-14.0306	experiments | evaluate
0.0698795	-14.0306	search | in
0.235294	-14.0307	input | rules
0.177419	-14.0307	therefore, | an
0.154762	-14.0308	is | novel
0.219512	-14.031	information | often
0.444444	-14.0313	differ | way
0.444444	-14.0313	theorem | can
0.444444	-14.0313	0.54 | 0.57
0.149425	-14.0314	think | to
0.107345	-14.0315	and | time.
0.180328	-14.0316	context | their
0.09375	-14.0317	on | classification
0.186441	-14.0317	less | words
0.5	-14.0317	five | positive
0.0976744	-14.0319	latent | in
0.333333	-14.032	demonstrate | proposed
0.444444	-14.032	over | topics,
0.291667	-14.0322	at | prior
0.357143	-14.0323	– | full
0.111111	-14.0323	that | languages.
0.164384	-14.0324	different | sentences
0.192308	-14.0324	then | classifier
0.444444	-14.0327	(i | i
0.6	-14.0328	thus, | lower
0.0776119	-14.033	classification | for
0.103261	-14.0331	detailed | in
0.333333	-14.0331	la | .
0.103825	-14.0331	i | by
0.138614	-14.0333	of | σ
0.384615	-14.0333	wsd | results
0.384615	-14.0333	verification | parallel
0.174603	-14.0333	three | sets
0.0277465	-14.0336	text | the
0.259259	-14.0337	so | context
0.25	-14.0338	• | :
0.116438	-14.0339	with | multiple
0.204082	-14.0339	proposes | to
0.315789	-14.034	dependency | syntax
0.121212	-14.034	errors | that
0.214286	-14.034	values | they
0.188679	-14.0341	systems | phrase
0.188679	-14.0341	any | where
0.444444	-14.0341	different | selection.
0.177419	-14.0341	see | over
0.5	-14.0341	applying | inference
0.228571	-14.0345	features | achieve
0.0920502	-14.0346	to | e.g.,
0.259259	-14.0346	= | τ
0.3	-14.0348	therefore | some
0.444444	-14.0349	positions | words
0.5	-14.0349	conversational | between
0.14	-14.0351	we | versions
0.12931	-14.0351	as | language.
0.384615	-14.0352	learner | high
0.196078	-14.0353	also | average
0.118056	-14.0355	polarity | in
0.0867925	-14.0355	is | among
0.0286753	-14.0357	two | a
0.214286	-14.0357	the | that.
0.106145	-14.0357	discussed | section
0.141414	-14.0358	bag-of-words | the
0.104972	-14.0359	if | i
0.0685579	-14.0359	an | all
0.269231	-14.0359	have | grammatical
0.0867925	-14.0359	provides | in
0.122137	-14.036	a | dataset.
0.192308	-14.0361	similarly, | with
0.14	-14.0362	inside | of
0.0361057	-14.0363	that | at
0.444444	-14.0363	evaluation | pyramid
0.444444	-14.0363	-3 | 0
0.219512	-14.0363	statement | to
0.269231	-14.0364	some | categories
0.384615	-14.0364	name | correct
0.192308	-14.0365	be | together
0.188679	-14.0365	set | phrases
0.242424	-14.0366	on | effects
0.0503778	-14.0367	a | while
0.0723514	-14.0367	constraints | and
0.259259	-14.0367	clause | with
0.219512	-14.0367	= | 1)
0.106145	-14.0368	2 | 10
0.183333	-14.0368	data | generated
0.384615	-14.037	original | model.
0.0456906	-14.0371	this | each
0.28	-14.0374	available | corpus.
0.183333	-14.0374	that | capable
0.357143	-14.0375	calculate | feature
0.259259	-14.0375	want | find
0.0406174	-14.0376	it | by
0.204082	-14.0376	were | average
0.6	-14.0377	multimodal | graphs
0.0493976	-14.0378	set | by
0.219512	-14.0378	this | parser.
0.126984	-14.0379	is | definition
0.333333	-14.0379	= | l=1
0.315789	-14.0381	· | +
0.219512	-14.0381	specification | and
0.0674157	-14.0382	the | typically
0.5	-14.0382	they | tags.
0.209302	-14.0383	jun’ichi | of
0.180328	-14.0383	vocabulary | by
0.084507	-14.0384	term | is
0.291667	-14.0384	with | closest
0.0632411	-14.0386	appears | the
0.3	-14.0387	of | improved.
0.3	-14.0387	experimentation | to
0.28	-14.0388	senses | target
0.0645833	-14.0388	previous | on
0.0515873	-14.0389	but | as
0.107784	-14.0389	model | score
0.124031	-14.0389	argument | are
0.0827586	-14.0389	dataset | in
0.291667	-14.0389	news | french
0.291667	-14.0389	11 | 7
0.5	-14.039	research | open
0.444444	-14.0392	hidden | variable
0.444444	-14.0392	task | polarity
0.357143	-14.0393	optimization | learning
0.124031	-14.0393	are | models.
0.162162	-14.0394	on | modeling
0.25	-14.0395	f-measure | by
0.0625	-14.0396	2004; | et
0.188679	-14.0396	translations | other
0.228571	-14.0397	we | count | number
0.6	-14.0397	bleu | conjoined
0.116438	-14.0399	nist | and
0.444444	-14.0399	0.66 | 0.65
0.444444	-14.0399	females | to
0.444444	-14.0399	orthographic | more
0.115646	-14.0401	level | as
0.192308	-14.0402	consider | such
0.3	-14.0402	into | whose
0.0833333	-14.0403	by | both
0.214286	-14.0403	parsing | trained
0.357143	-14.0404	return | section
0.5	-14.0406	similar | different.
0.0477816	-14.0408	which | which
0.0462185	-14.0409	(1) | and
0.228571	-14.0409	large-scale | as
0.196078	-14.041	use | small
0.0801282	-14.0413	generation | to
0.138614	-14.0414	a | beyond
0.444444	-14.0414	hypothesis | under
0.444444	-14.0414	different | representations.
0.5	-14.0414	2, | length
0.5	-14.0414	word | aligns
0.5	-14.0414	comments | anonymous
0.0764706	-14.0414	not | only
0.0309229	-14.0415	data | a
0.0395349	-14.0415	that | one
0.0527778	-14.0418	of | system.
0.0692124	-14.0418	to | errors
0.121212	-14.0419	overall, | of
0.121212	-14.0419	rule | from
0.111111	-14.0419	such | using
0.188679	-14.0421	their | human
0.124031	-14.0421	location | and
0.14433	-14.0422	architecture | in
0.315789	-14.0422	set | constraints.
0.114865	-14.0422	a | reading
0.204545	-14.0424	such | ,
0.0588235	-14.0425	on | same
0.106145	-14.0425	to | answer
0.0358056	-14.0427	vector | the
0.0561622	-14.0427	we will | of
0.0634921	-14.0427	to | small
0.209302	-14.0428	this | explain
0.444444	-14.0428	c | original
0.0791139	-14.0431	of | produced
0.25	-14.0431	how | parsing
0.242424	-14.0433	work | grants
0.114865	-14.0433	for | amount
0.188679	-14.0433	best | systems
0.092511	-14.0434	modules | the
0.214286	-14.0434	between | common
0.25	-14.0435	only | discourse
0.128205	-14.0435	as | objective
0.6	-14.0436	van | &
0.0698795	-14.0437	list | a
0.0954545	-14.0437	mapping | a
0.0816327	-14.0439	it is | an
0.196078	-14.0439	therefore | can
0.219512	-14.0439	section | briefly
0.0161167	-14.044	for | we | the
0.138614	-14.0441	) | |
0.196078	-14.0442	1, | or
0.0564263	-14.0443	, | and
0.092511	-14.0443	are | document
0.444444	-14.0443	sets | results.
0.444444	-14.0443	judged | better
0.225	-14.0444	turkers | the
0.225	-14.0444	z | that
0.1125	-14.0445	choice | to
0.0641822	-14.0448	knowledge | in
0.0671141	-14.0449	a | natural
0.269231	-14.0452	compare | learning
0.185185	-14.0452	we | (cf.
0.228571	-14.0454	a | wsi
0.6	-14.0456	decoder | multiple-beam
0.259259	-14.0456	restriction | is
0.185185	-14.0458	by | having
0.384615	-14.0458	example | de
0.026084	-14.0458	model | in
0.174603	-14.0458	present | learning
0.0883534	-14.046	possible | are
0.14433	-14.0462	weights | are
0.0995025	-14.0462	a | appears
0.219512	-14.0464	propositions | be
0.269231	-14.0465	• | lattice
0.0657895	-14.0465	statistical | translation
0.384615	-14.0465	method | detecting
0.2	-14.0465	many | system
0.0747922	-14.0465	many | that
0.0335946	-14.0466	defined | the
0.6	-14.0466	rules | remove
0.6	-14.0466	2: | temporal
0.185185	-14.0467	nevertheless, | in
0.118881	-14.0467	very | from
0.0260399	-14.0468	was | and
0.169231	-14.0469	adjacent | of
0.00903845	-14.0469	we | the | and
0.228571	-14.0469	according | some
0.2	-14.0469	information | important
0.28	-14.047	define | 0
0.0509804	-14.047	of | proposed
0.0446872	-14.0471	the | least
0.133929	-14.0471	our | 1.
0.0406977	-14.0472	our | which
0.444444	-14.0472	following | estimated
0.259259	-14.0473	useful | model
0.039627	-14.0473	the | state | the
0.0607735	-14.0474	the | conference
0.28	-14.0475	also | items
0.6	-14.0475	axioms | (table
0.219512	-14.0475	based | information.
0.109756	-14.0477	and | expected
0.0848708	-14.0477	connected | the
0.333333	-14.0478	instance | corpus
0.185185	-14.0479	8 | 7
0.444444	-14.0479	will | exploit
0.169231	-14.048	model | content
0.357143	-14.048	λ | r
0.3	-14.0481	annotated | agreement
0.333333	-14.0483	0.62 | 0.60
0.0788644	-14.0483	features | words
0.384615	-14.0484	blog | comment
0.25	-14.0485	simplest | for
0.6	-14.0485	dyer | 15213,
0.14433	-14.0486	0.6 | 0.8
0.149425	-14.0487	cases | by
0.169231	-14.0488	cluster | word
0.0883534	-14.049	they | same
0.0509804	-14.0491	system | that
0.097561	-14.0492	our | while
0.333333	-14.0494	omitted | for
0.067757	-14.0494	some | with
0.0520548	-14.0495	the | already
0.6	-14.0495	takes | integer
0.0887097	-14.0496	was | this
0.269231	-14.0496	with | parser,
0.269231	-14.0496	as | similarity.
0.384615	-14.0496	shown | highest
0.0665188	-14.0497	news | and
0.5	-14.0497	evaluated | automatically
0.0813559	-14.0499	we | during
0.242424	-14.0501	order | representation
0.100503	-14.0503	not | do
0.0980392	-14.0503	a | a sequence
0.0596745	-14.0505	a | generated
0.242424	-14.0505	sibling | in
0.6	-14.0505	indicates | summaries
0.192308	-14.0506	usage | is
0.114094	-14.0506	were | same
0.144444	-14.0507	how | into
0.28	-14.0507	random | performance
0.357143	-14.051	r1 | x2
0.162162	-14.0511	to | move
0.235294	-14.0511	training | consisting
0.047191	-14.0512	(3) | the
0.0833333	-14.0512	and | maximum
0.0618956	-14.0512	this is | for
0.269231	-14.0514	both | arguments
0.0326029	-14.0515	the | those | the
0.118056	-14.0515	can | accuracy
0.384615	-14.0515	exposed | nodes
0.0565428	-14.0516	the | f
0.357143	-14.0516	those | within
0.111111	-14.0518	of | 2).
0.169231	-14.0519	on | accuracy.
0.0491018	-14.052	the second | of
0.188679	-14.052	translation | scores
0.110429	-14.0521	the | singular
0.0662252	-14.0521	that | even
0.5	-14.0521	exploit | similarity
0.116438	-14.0522	an | alignment
0.333333	-14.0522	each | sets,
0.180328	-14.0525	the | stores
0.147727	-14.0525	inference | by
0.6	-14.0525	duc | general
0.6	-14.0525	the | semeval-2007,
0.123077	-14.0525	space | with
0.0656455	-14.0526	which | at
0.114094	-14.0526	n | p
0.0522696	-14.0526	4 | of | and
0.113333	-14.0528	7 | with
0.192308	-14.0529	cluster | same
0.242424	-14.0529	sind | are
0.196078	-14.053	consider | different
0.204545	-14.0532	in | triggers
0.0901639	-14.0533	x | for
0.196078	-14.0533	are | intended
0.219512	-14.0533	from | —
0.14	-14.0534	workers | to
0.0890688	-14.0534	this | structure
0.3	-14.0535	previous | best
0.3	-14.0535	unlike | previous
0.0550459	-14.0535	models | as
0.0970874	-14.0537	for | discourse
0.0520548	-14.0538	the | whose
0.144444	-14.0538	we | eq.
0.28	-14.0539	y | q
0.169231	-14.0539	to | align
0.0712468	-14.0539	performance | by
0.357143	-14.0539	precision | 0.94
0.357143	-14.0539	instead | knowledge
0.0671296	-14.054	→ | the
0.118519	-14.0541	a | online
0.25	-14.0543	been | studies
0.28	-14.0544	since | gold
0.28	-14.0544	highly | more
0.155844	-14.0545	4 | at
0.6	-14.0545	f-score | very
0.0348624	-14.0546	that | word
0.0813559	-14.0548	language | or
0.242424	-14.0549	is | position.
0.0607735	-14.0549	of | tags
0.0876494	-14.0549	furthermore, | and
0.110429	-14.0551	people | a
0.357143	-14.0551	although | space
0.114094	-14.0551	cases | as
0.242424	-14.0553	, | u
0.109091	-14.0554	means | are
0.259259	-14.0554	rules | glue
0.0301232	-14.0555	to | language
0.138614	-14.0556	into | use
0.114865	-14.0556	has | same
0.047836	-14.0558	text | in
0.0980392	-14.0558	and | temporal
0.188679	-14.0558	most | syntactic
0.192308	-14.056	task | systems
0.3	-14.056	(3) | function
0.157895	-14.056	will | sentence
0.072	-14.0562	in | node
0.097561	-14.0564	these | information
0.112583	-14.0565	to | network
0.151163	-14.0565	of | users.
0.269231	-14.0567	computed | values
0.127119	-14.0571	double | the
0.0652174	-14.0572	is | type
0.0527778	-14.0573	determine | the
0.235294	-14.0573	system | module
0.0778816	-14.0574	of | variables
0.111111	-14.0575	next | for
0.0954545	-14.0576	60 | of
0.0527046	-14.0577	we | a | it
0.169231	-14.0578	only | input
0.180328	-14.0579	arguments | not
0.071618	-14.0579	of | joint
0.0703518	-14.0579	can | other
0.3	-14.058	have | spoken
0.25	-14.058	trees | same
0.209302	-14.0583	where | side
0.0887097	-14.0584	iteration | the
0.25	-14.0584	annotators | one
0.071618	-14.0585	we | a | algorithm
0.235294	-14.0585	corpus | us
0.062753	-14.0585	computing | the
0.192308	-14.0586	which | string
0.0778816	-14.0586	generate | a
0.142857	-14.0586	is | optimization
0.0808081	-14.0589	experiments, | and
0.130435	-14.059	selected | are
0.107143	-14.0591	bayesian | and
0.384615	-14.0592	would | obtain
0.111842	-14.0593	any | it
0.027027	-14.0593	5 | the
0.259259	-14.0593	related | classification
0.6	-14.0595	avery | is
0.0990099	-14.0595	algorithms | a
0.122137	-14.0596	by | candidate
0.384615	-14.0598	bootstrapping | small
0.333333	-14.0599	how | approaches
0.333333	-14.0599	given | eq.
0.16	-14.06	move | and
0.102703	-14.06	such | one
0.214286	-14.0601	learning | instances
0.146067	-14.0602	emission | the
0.204545	-14.0603	state | time
0.151163	-14.0603	 | 
0.0733696	-14.0603	for | sense
0.28	-14.0604	paths | are
0.5	-14.0604	chocolate | chocolate
0.5	-14.0604	greater | england
0.5	-14.0604	d96 | d93
0.5	-14.0604	genia | rna
0.5	-14.0604	slot4], | reward:
0.5	-14.0604	didn’t | didn’t
0.5	-14.0604	t18 | t19
0.5	-14.0604	p(θ,y,b|x,x
0.5	-14.0604	bilmes, | krause
0.5	-14.0604	headlines | factoids
0.5	-14.0604	cuexact | llr
0.5	-14.0604	73.3 | 73.5
0.5	-14.0604	tjong | sang,
0.5	-14.0604	fpr | fpr
0.5	-14.0604	b7 | cwcybnczcx
0.5	-14.0604	wpred | wj(ci,j+wpartial
0.5	-14.0604	ilp | ibm-2
0.5	-14.0604	skipped): | labeled
0.5	-14.0604	(cl) | 32.28
0.5	-14.0604	notp | 0
1	-14.0605	crf-std+ | crf-lex+
1	-14.0605	s49s50 | s87s111s114s100s32s76s101s110s103s116s104
1	-14.0605	hamā | padhā
1	-14.0605	f(κ0 | −ξi),
1	-14.0605	pidat): | pavrel:
1	-14.0605	[jkp+07]: | tomkins.
1	-14.0605	)()( | navgf
1	-14.0605	prefix | stable.
1	-14.0605	feat | 〈q,t,precedesequal,ap〉,
1	-14.0605	0.450 | 0.131
1	-14.0605	#1: | {hockey,baseball}
1	-14.0605	npj, | |rpj|
1	-14.0605	76.7% | 58.2%
1	-14.0605	to|ecur|do | ecur
1	-14.0605	aviv | aviv,
1	-14.0605	concomp | onehop
1	-14.0605	qj | +γtrace[(s−s0)t(s−s0)],
1	-14.0605	p&r= | k=0.87
1	-14.0605	hashtags | smileys.
1	-14.0605	(after | undesired
1	-14.0605	polytechnique | chemin
1	-14.0605	morgen:morning | können:may
1	-14.0605	-1 | -0.2
1	-14.0605	eigenvector | qpi
1	-14.0605	hiragana, | kanji,
1	-14.0605	nouns(si) | pronouns(si)
1	-14.0605	〈det〉the | 〈artifact〉d∧
1	-14.0605	evidence, | antithesis,
1	-14.0605	hoa | diane
1	-14.0605	pl | eγsil
1	-14.0605	geo:flowsthrough | geo:instate
1	-14.0605	release | $250
1	-14.0605	nel | 2011]
1	-14.0605	s54s48s37 | s56s48s37
1	-14.0605	22536 | gp41-induced
1	-14.0605	s50s48s48s48s48 | s50s48
1	-14.0605	[(r1, | r2),(r2,
1	-14.0605	p(f|e).
1	-14.0605	ana:t30 | ana:t32
1	-14.0605	z(x) | sharpness
1	-14.0605	mimicry | bargh,
1	-14.0605	(active | errors);
1	-14.0605	16.5% | 19.7%
1	-14.0605	(vp | lack)
1	-14.0605	oriented | (bod,
1	-14.0605	ele[he] | visitar[visit]
1	-14.0605	np | 89.10
1	-14.0605	sift | pixels
1	-14.0605	unit(r1) | passive(r2)
1	-14.0605	each) | (varying
1	-14.0605	rprime | min(ceprime(ng),cprime(ng))
1	-14.0605	73.2% | 56.3%
1	-14.0605	yprime α[x−1,1,yprime]·pj(y|yprime)
1	-14.0605	unine | savoy,
1	-14.0605	lexicalised | lexical;
1	-14.0605	relationships | bramsen
1	-14.0605	adj-nn | v-subj
1	-14.0605	â | ¯h
1	-14.0605	semigroup | null1(null)
1	-14.0605	costs: | dec.:
1	-14.0605	nserc, | (rgpin:
1	-14.0605	frick, | grohe
1	-14.0605	lambda | formulas
1	-14.0605	fail | resolved.
1	-14.0605	performs, | takes,
1	-14.0605	(percentage | returned),
1	-14.0605	70,6062006: | impact”
1	-14.0605	47.3 | align+ind
1	-14.0605	relevance) | fingertips
1	-14.0605	mechanism | warm-start
1	-14.0605	french, | annodis
1	-14.0605	/> | <prop
1	-14.0605	(difonzo | bordia,
1	-14.0605	latbleu | nbexpbleu
1	-14.0605	coreferenceresolution | cluster
1	-14.0605	2(e) | 2(f)
1	-14.0605	genecomplex | geneproduct
1	-14.0605	82.45 | 78.66
1	-14.0605	ui | widgets,
1	-14.0605	efron | thisted
1	-14.0605	to | aviv.’
1	-14.0605	e(i,j) | e(i,jprime),
1	-14.0605	stanford | 7.9m
1	-14.0605	root_domain | pure_science
1	-14.0605	zesch, | hochschulstrasse
1	-14.0605	dimlex | umbach,
1	-14.0605	mineditedcost | mineditedsug
1	-14.0605	99.4 | 97.6
1	-14.0605	e∈dn(f) | c,1.0
1	-14.0605	morgan | publishers.
1	-14.0605	r(x, | when
1	-14.0605	s-ns | s-p
1	-14.0605	0.3404 | ⟨digits⟩
1	-14.0605	hypertension | intracranial
1	-14.0605	effac´es | hawaienne
1	-14.0605	36.6 | 49.9
1	-14.0605	physics | malostransk´e
1	-14.0605	(i | 1,2)
1	-14.0605	“?�(fahui)” | “/
1	-14.0605	t0gether, | 5top,
1	-14.0605	“eparl+nc” | “eparl+nc+crawl2”
1	-14.0605	carlos | lig,
1	-14.0605	|}&1{| | ooj||ip
1	-14.0605	[i,j]h | sealed
1	-14.0605	ppt | ast.
1	-14.0605	santanu | tanik
1	-14.0605	closest-firstclusteringmethodto | thecoreferentialpre-clustersandgeneratethefinal
1	-14.0605	8.5 | 9.25
1	-14.0605	(dunning, | (tmi)
1	-14.0605	t0) | )2z
1	-14.0605	arg2-atr | arg2-sou
1	-14.0605	// | ///
1	-14.0605	1–270, | 493,892
1	-14.0605	(null106) | (sec/sent)
1	-14.0605	encoding | (bloom,
1	-14.0605	0.1598 | 0.3227
1	-14.0605	fr | postaux
1	-14.0605	t¨opfe | der
1	-14.0605	danger | orange,
1	-14.0605	zs(r) | ∀s,r∈
1	-14.0605	(maamouri | mada+tokan
1	-14.0605	adji | obji
1	-14.0605	questions | “why
1	-14.0605	ibm1, | ibm4.
1	-14.0605	χαρτόκουτο | κουτί
1	-14.0605	schwachheit | pwavrel:
1	-14.0605	account, | costs
1	-14.0605	avgip | avgilr
1	-14.0605	maintains | encountered.
1	-14.0605	i;h1;k | k;h2;j
1	-14.0605	edinburg | goldfarb-halpern
1	-14.0605	0.326 | 0.199
1	-14.0605	elementary | (tr2str),
1	-14.0605	one, | ten,
1	-14.0605	indo-european, | sino-tibetan.
1	-14.0605	laurie | weingart,
1	-14.0605	“round-trip” | targetlanguage
1	-14.0605	ω | ωi
1	-14.0605	mccoy | delaware;
1	-14.0605	(go | barbosa
1	-14.0605	genitive | non-human
1	-14.0605	terminology | tc37/sc4
1	-14.0605	wfars | almslmwn
1	-14.0605	reddington | tintarev
1	-14.0605	10000 | 990180
1	-14.0605	thresholds | 0.1).
1	-14.0605	investment | retirement
1	-14.0605	cppcc | chiang
1	-14.0605	hosts | symbiotic
1	-14.0605	ratings | timing:2,
1	-14.0605	slda | slda
1	-14.0605	speeches, | propaganda
1	-14.0605	,, | 
1	-14.0605	(katz | giesbrecht,
1	-14.0605	’program’, | sp.
1	-14.0605	bleuj | bleut
1	-14.0605	anchor | event]
1	-14.0605	normalizing | φi(t,j,s)
1	-14.0605	particles | roles;
1	-14.0605	subliminal | danger,
1	-14.0605	lem | dep
1	-14.0605	+0.8 | (+0.6)
1	-14.0605	rbo | thes
1	-14.0605	[palestinian | pna
1	-14.0605	rebuttals | sided
1	-14.0605	θ+ | θ−
1	-14.0605	epsilon1 | usprime∈nmf
1	-14.0605	della | pietra.
1	-14.0605	compliment | 51.51
1	-14.0605	(np), | (vp),
1	-14.0605	bracketleftbig | µarg1i,j,k
1	-14.0605	confidant | 2416
1	-14.0605	hamidreza | kobdani@ims.uni-stuttgart.de
1	-14.0605	5,363 | 68,824,813
1	-14.0605	raux | endpointing
1	-14.0605	ant:t28 | [t10]
1	-14.0605	1800 | swiss
1	-14.0605	g., | moya
1	-14.0605	negative) | (1993)
1	-14.0605	greenbacker | seniz
1	-14.0605	full-screen | tabs,
1	-14.0605	ck | “greater
1	-14.0605	alanis | alejandro
1	-14.0605	2.000 | 16.000
1	-14.0605	“happy,” | “sad.”
1	-14.0605	nf,z | βn
1	-14.0605	0.4578 | 0.7108
1	-14.0605	e+i | e−i
1	-14.0605	:d | :[
1	-14.0605	joined | they’ll).
1	-14.0605	systematically, | “pairwise
1	-14.0605	mbot | lmbot)
1	-14.0605	conclusion, | revu
1	-14.0605	english/arabic, | english/tamil
1	-14.0605	ab | bezahlt
1	-14.0605	saudi | statementthat
1	-14.0605	unstructured | (uima)
1	-14.0605	skewness | kurtosis
1	-14.0605	hull | (barber
1	-14.0605	0.861 | 0.614
1	-14.0605	s.; | l.;
1	-14.0605	paragraphs | candidates.
1	-14.0605	fleiss’ | moderate.
1	-14.0605	fraunhofer | sankt
1	-14.0605	information-giving | (makatchev
1	-14.0605	fom | ˆα(ni,j)
1	-14.0605	usc | waterfront
1	-14.0605	cautious | behavior;
1	-14.0605	s3: | rig
1	-14.0605	confirmed | implconfaskaslot
1	-14.0605	chirpy | stealthily
1	-14.0605	nidhi | ma,
1	-14.0605	tepperman, | patti
1	-14.0605	seeds | prototype-driven
1	-14.0605	'05 | '99
1	-14.0605	tf*idf, | [lsa,
1	-14.0605	combinations | zipf-like
1	-14.0605	tri | c.
1	-14.0605	kansallismielisten | nationalists-pl/gen
1	-14.0605	#michaeljackson | #mj
1	-14.0605	l(ψ) | tpl+
1	-14.0605	j. | niu.
1	-14.0605	76.2 | 71.8
1	-14.0605	42.8 | faust+1p
1	-14.0605	top(agenda) | top-b(agenda,
1	-14.0605	loss, | profit,
1	-14.0605	(davis | dhillon,
1	-14.0605	delivery | (christ,
1	-14.0605	regular | higuera
1	-14.0605	hni | h1i
1	-14.0605	practice | (tpo)
1	-14.0605	<genowner> | <creator>
1	-14.0605	replicating | ‘vaala’
1	-14.0605	para2 | para1,
1	-14.0605	bikel’s | comparator
1	-14.0605	mbr | e∈ee
1	-14.0605	ocij | ocsijk
1	-14.0605	(qe) | (qm)
1	-14.0605	cjlin/libsvm | 84.90
1	-14.0605	0.5249 | 0.5512
1	-14.0605	|epsilon1/f |e/epsilon1
1	-14.0605	im2 | management)
1	-14.0605	pos-2grams | pos-4grams
1	-14.0605	selftraining | black
1	-14.0605	omnis | amantis
1	-14.0605	variation) | (habash
1	-14.0605	(resp.,|a|s).
1	-14.0605	{[−2, | was]
1	-14.0605	tools: | svmlighttk
1	-14.0605	k-testable | nullw,v
1	-14.0605	w+1 | w+2
1	-14.0605	simpleclassification | developed.for
1	-14.0605	(resp., | ofais
1	-14.0605	hatwidea | u,t
1	-14.0605	h&k | nwire-ace2003
1	-14.0605	bigrams | better 
1	-14.0605	petukhova, | keizer,
1	-14.0605	τsprime | p(τsprime
1	-14.0605	goes | tsg.
1	-14.0605	proteins, | pathogens,
1	-14.0605	nodes(g) | source(e)
1	-14.0605	61.8 | 72.75
1	-14.0605	(cavnar | trenkle,
1	-14.0605	43.4 | 43.3
1	-14.0605	house” | “in
1	-14.0605	(cost), | (bias).
1	-14.0605	=<tree: | perennial
1	-14.0605	∈{leaf, | left-head,
1	-14.0605	balamurali | iitb-monash
1	-14.0605	pushdown | (pda)
1	-14.0605	sc3 | sc4.
1	-14.0605	riloff | wiebe
1	-14.0605	wen | cote
1	-14.0605	poverty | income
1	-14.0605	15.58 | 19.62
1	-14.0605	patient | sumaries.
1	-14.0605	christian | j¨org
1	-14.0605	puppe, | ‘der
1	-14.0605	hulten | babcock
1	-14.0605	catalan | ’6
1	-14.0605	eg882>b | eg882>d
1	-14.0605	scc | tcc
1	-14.0605	protein-protein | (ppis),
1	-14.0605	schemes | reed,
1	-14.0605	wash-small-fruit(washer, | target:sink)
1	-14.0605	“hat” | “read”
1	-14.0605	tolba | torgoman
1	-14.0605	ita/ita1, | eng/ita1
1	-14.0605	ner | (finkel05)
1	-14.0605	named | q1,
1	-14.0605	ros´e | torrey,
1	-14.0605	for(i=0; | n;
1	-14.0605	in | (pl).
1	-14.0605	start="1195" | lemma="i
1	-14.0605	mty | mjrm
1	-14.0605	(macwhinney | 1985).
1	-14.0605	discarded, | geo:city(y)
1	-14.0605	qu | q|j|u
1	-14.0605	o(n2m2) | li,j,
1	-14.0605	<num | </nu
1	-14.0605	die | meisten
1	-14.0605	yprime pj(yprime|y)·β[x+
1	-14.0605	resnik, | weinberg,
1	-14.0605	combined | self-train:
1	-14.0605	ζ | block
1	-14.0605	david | hideto
1	-14.0605	wrong | 14.59%
1	-14.0605	sized | (gw),
1	-14.0605	(anderson, | murdock,
1	-14.0605	[g] | gu,v
1	-14.0605	kang | china
1	-14.0605	(close | lemma’s
1	-14.0605	rtngrams | )(#
1	-14.0605	(tufiş | ştefănescu,
1	-14.0605	nate | beaumont,
1	-14.0605	for | llbracketcηι
1	-14.0605	echihabi | kevin
1	-14.0605	(svo) | subject-object-verb
1	-14.0605	’he’ | ’him’
1	-14.0605	tudio | nelp
1	-14.0605	investigating | constructions.
1	-14.0605	kumaran, | khapra,
1	-14.0605	biblical | (prophetic
1	-14.0605	16.77, | 15.22
1	-14.0605	(chi | independence,
1	-14.0605	n-gram: | word2”
1	-14.0605	stanzas | terza
1	-14.0605	33.8% | rmse(v)
1	-14.0605	nd-l1 | nd-l4
1	-14.0605	strzalkowski, | broadwell,
1	-14.0605	proof | spovb
1	-14.0605	(root | (agr
1	-14.0605	errors: | mt-:
1	-14.0605	exposures | 100
1	-14.0605	plc, | shv
1	-14.0605	τ(r) | pre-translations
1	-14.0605	bit | “court”
1	-14.0605	cyk+ | rajman,
1	-14.0605	pi | φpi
1	-14.0605	representative | diagnoses
1	-14.0605	bacteria | “strain”,
1	-14.0605	conclusions, | recomendations
1	-14.0605	thrown, | custody,
1	-14.0605	containing | posf,
1	-14.0605	na: | zaun,
1	-14.0605	treatment | allergies
1	-14.0605	(weischedel | brunstein,
1	-14.0605	mich | dieses
1	-14.0605	{compute | θ}
1	-14.0605	gmbh | center
1	-14.0605	reference: | diese
1	-14.0605	α+j | α−j
1	-14.0605	medicine | md,
1	-14.0605	maxy,z | vbzb
1	-14.0605	vossen, | 17-23,
1	-14.0605	s0wc, | s0lwc,
1	-14.0605	pantel | dorow
1	-14.0605	notified | palladian.
1	-14.0605	dn |=y
1	-14.0605	a2 | crr=
1	-14.0605	wissenschaftler, | kann.
1	-14.0605	du | r´eglementation
1	-14.0605	proposea | pairmodel.inthismodel,pre-clusteringofall
1	-14.0605	scss-tcd:conf1 | scss-tcd:conf2
1	-14.0605	richardson | smeaton
1	-14.0605	mitesh | salil
1	-14.0605	wednesday | home.
1	-14.0605	glyph4 | daumé
1	-14.0605	variation | text’
1	-14.0605	2,606 | 54.47
1	-14.0605	47.2 | 40.4
1	-14.0605	def≡ | ·h⟨1,e1⟩.
1	-14.0605	blackberry | ubertwitter
1	-14.0605	aljwd¯hςaly¯h | alqstntyny¯h
1	-14.0605	cf2 | cf1
1	-14.0605	schunn, | (cho
1	-14.0605	nullnullnullnullnullnullnullnull | nullnullnullnullnullnullnullnullnullnullnullnull
1	-14.0605	os/o | r1r2
1	-14.0605	width | two-dimensional
1	-14.0605	sek¨a | yhteisi¨a
1	-14.0605	0.5967 | 0.6224
1	-14.0605	(ide | suderman,
1	-14.0605	harvesting) | (adjective
1	-14.0605	(on | 5.3)
1	-14.0605	class-instance | instance-attribute
1	-14.0605	manli | (maccartney
1	-14.0605	mnb/pool | mnb/priors,
1	-14.0605	fr¨uherer | von
1	-14.0605	|r∩ ˆr|
1	-14.0605	hdt | k3
1	-14.0605	rojo | /gr1(x
1	-14.0605	18.6 | 14.62
1	-14.0605	china | ¶
1	-14.0605	rtici | effe
1	-14.0605	comm-s | comm-r
1	-14.0605	nw|t.
1	-14.0605	fünf | festgenommen
1	-14.0605	obozinski | lounici
1	-14.0605	msdr | 28.77
1	-14.0605	andreas | {vlachos,craven}@biostat.wisc.edu
1	-14.0605	pc→t(n|o,c) | p(n|z)p(z|o,c)
1	-14.0605	lagrange | λt+1s
1	-14.0605	[−2, | rb,
1	-14.0605	factorization | g∗j=k,s∗j=k,f∗j=k
1	-14.0605	arise:5, | up:2,
1	-14.0605	economics | imis,
1	-14.0605	c2) | (cause’
1	-14.0605	amine | carboxylic
1	-14.0605	pink, | purple,
1	-14.0605	acoustic | hub4
1	-14.0605	rhetorical | q-elab,
1	-14.0605	76.17 | 83.78
1	-14.0605	silva, | eugénio
1	-14.0605	ˆy | maxˆyθ·φ(ˆy)
1	-14.0605	ra | itera
1	-14.0605	salary | dollars,
1	-14.0605	log-bilinear | mnih
1	-14.0605	æ”m | hfydat
1	-14.0605	p: | convertible
1	-14.0605	brown, | f.,
1	-14.0605	medstract, | drugnerar,
1	-14.0605	modifications: | norp,
1	-14.0605	68.3 | −mi2
1	-14.0605	gouws | thank
1	-14.0605	sudden | aboard
1	-14.0605	0.842 | 0.770
1	-14.0605	1...n}, | relvector(ej,ek).
1	-14.0605	oi | vectorxi
1	-14.0605	haµ←haµ | kwka
1	-14.0605	run | run-score-1
1	-14.0605	slovenian | 234
1	-14.0605	leftarc | rightarc
1	-14.0605	c(w | 32)−d′′(n1+(•w32))
1	-14.0605	s3 | present[mm]
1	-14.0605	0.4470 | 0.6156
1	-14.0605	qsbp, | ui∈s}
1	-14.0605	rec, | 1994],
1	-14.0605	北韩 | 之一
1	-14.0605	messaging | serviced,
1	-14.0605	verbal | hyperbole,
1	-14.0605	de-en, | es-en.
1	-14.0605	k, | ai,j+1,k
1	-14.0605	p(d|c) | 1z(c)exp
1	-14.0605	squliq | /pisip/,
1	-14.0605	theorems | ltt
1	-14.0605	constraint: | i=1
1	-14.0605	0.4134 | 0.6206
1	-14.0605	ys | 1i
1	-14.0605	‘helicopter’ | ‘wood’
1	-14.0605	entity%1:03:00 | state%1:03:00).
1	-14.0605	1− | risummationtext
1	-14.0605	e2-1 | e2-2
1	-14.0605	77.9 | 79.5
1	-14.0605	claire | 1mig
1	-14.0605	1,2in | 8,8distribution
1	-14.0605	(dg), | df
1	-14.0605	relativesalience(w|t1,t2) | f1n
1	-14.0605	f­m | exp1
1	-14.0605	producttext | score(c,x,y)summationtext
1	-14.0605	α[0,0,y] | 1,0,yprime]
1	-14.0605	wedefine | np-nplink,np-prplinkandprp-prplink.(here
1	-14.0605	svl | flanders
1	-14.0605	[e1 | throne
1	-14.0605	kindsof | np-prplink
1	-14.0605	s2), | s1),
1	-14.0605	uniqueness: | inherent
1	-14.0605	prd | football::n
1	-14.0605	16.66 | 21.25
1	-14.0605	lemma="phosphorylation" | start="1214"
1	-14.0605	nkqm@10 | 0.6608
1	-14.0605	kp | |kpj|
1	-14.0605	17,672 | 38,128
1	-14.0605	0.0% | 100.0%
1	-14.0605	et | trait´e
1	-14.0605	jedi | serena
1	-14.0605	(albercht | gimenez,
1	-14.0605	sprime1 | sprime2
1	-14.0605	(druck | attenberg
1	-14.0605	kataba | ‘to
1	-14.0605	u1 | u2,
1	-14.0605	(quirk | kurohashi,
1	-14.0605	(george | 乔治null沃克null布什
1	-14.0605	“bbl” | “be
1	-14.0605	jin | low,
1	-14.0605	np;elsen | mentinotype
1	-14.0605	ter-bleu/2 | meteor-tune
1	-14.0605	detects | epidemics
1	-14.0605	bombs | evening’.
1	-14.0605	languages, | 1.3%.
1	-14.0605	pane | bird-eye
1	-14.0605	minnesota | white-abrams
1	-14.0605	(hoffmann | 2010),
1	-14.0605	(&) | 85.87
1	-14.0605	um | ub,
1	-14.0605	łq | alaxyrp
1	-14.0605	σa+ | σa−
1	-14.0605	enumerate | extreme,
1	-14.0605	discriminant | (rda)
1	-14.0605	ˆed←delete(ecur,j) | ifg(eprimei)
1	-14.0605	gy | transliteration:
1	-14.0605	champs/marne | marne-la-vall´ee
1	-14.0605	jelinekmercer | λn(φn)·pn(wi|φn)
1	-14.0605	(wolpaw, | pfurtscheller
1	-14.0605	[pred | problem]”
1	-14.0605	fuf/surge | robin,
1	-14.0605	genia | (genia)
1	-14.0605	3.1 | numbered
1	-14.0605	(“blow | se”).
1	-14.0605	avenue, | uk
1	-14.0605	‘qp’ | ‘np’
1	-14.0605	spontaneously | prompted
1	-14.0605	express | commenting
1	-14.0605	p(w|t) | (17).
1	-14.0605	semantics | gaining
1	-14.0605	second,we | harmingthe
1	-14.0605	[k,lscript) | ψ(m,lscript)
1	-14.0605	89.7 | 90.6
1	-14.0605	tok | msf
1	-14.0605	hypothetical | h(pki+1,s∗t
1	-14.0605	jrc-acquis | bnc3.
1	-14.0605	used | trainers
1	-14.0605	weights: | ←mert
1	-14.0605	cite | 95.71
1	-14.0605	unary, | right-head}
1	-14.0605	services | restored.
1	-14.0605	noth | 0.807§
1	-14.0605	[[u4]]c | at(r,n1)∧past(¬at(r,n1))
1	-14.0605	28.65 | +0.89
1	-14.0605	macam | 2moro,
1	-14.0605	pn(e|f) | |e|·pn(e|f)
1	-14.0605	i,j | maxi,j{blij}
1	-14.0605	subcomputations | 2,
1	-14.0605	(bradlow | bent,
1	-14.0605	hugs | wow
1	-14.0605	return | “leaving
1	-14.0605	matteo | fbk-irst
1	-14.0605	pedro | pa
1	-14.0605	steve/nnp | “steve
1	-14.0605	‘use | strategy’,
1	-14.0605	tonight,we | beautifulhouse
1	-14.0605	rakuten | technology,
1	-14.0605	mycket | den vetenskapliga
1	-14.0605	288 | range.n
1	-14.0605	deg | deg
1	-14.0605	villavicencio♠ | grenoble,
1	-14.0605	4: | y(q)j
1	-14.0605	5586 | connectives
1	-14.0605	rouge | rouge-l,
1	-14.0605	gehe | hause.
1	-14.0605	higins, | sadock,
1	-14.0605	coarse | nmod,
1	-14.0605	35m | ldc2008t13.
1	-14.0605	broke | [window]a1.
1	-14.0605	biography | bwsa.
1	-14.0605	wolf | gurevych
1	-14.0605	(juman) | (75.80/84.89)
1	-14.0605	contained, | stuffed,
1	-14.0605	terrible | mess
1	-14.0605	semi-de | matrix.
1	-14.0605	<dsc> | <descgrp>
1	-14.0605	srl-aware | covers.
1	-14.0605	pre-terminal | vectorn
1	-14.0605	lesk | 86.41
1	-14.0605	ontonotes | large-scale,
1	-14.0605	0.81 | moses.
1	-14.0605	97.4 | 96.8
1	-14.0605	“investigation” | “arrest”
1	-14.0605	carolyn | rosé
1	-14.0605	(not’ | d2)
1	-14.0605	p(s1) | i=2
1	-14.0605	nullnullnullnullnullnullnullnullnullnullnullnullnullnullnullnullnullnull | interleaved
1	-14.0605	1-terp | bleu-4-v13a-c
1	-14.0605	sigmoid | (sbns)
1	-14.0605	67 | 68–77,
1	-14.0605	giulia | (ilc–cnr)
1	-14.0605	enullnullnullnullnull | pnullnullnull
1	-14.0605	4productdisplay | wherenk(eprime)
1	-14.0605	l∃ | has-location
1	-14.0605	congressional | avert
1	-14.0605	pp# | np#2
1	-14.0605	agentive | telic.
1	-14.0605	(yamangil | nelken,
1	-14.0605	svr | y(xij)
1	-14.0605	(macherey | chatterjee
1	-14.0605	brown-type-r | lattice-type-r
1	-14.0605	n→turn | likelihoods
1	-14.0605	p(y | 1|x
1	-14.0605	well-studied | 1989;
1	-14.0605	kim | yayoi,
1	-14.0605	sind | “es
1	-14.0605	“that | mississippi
1	-14.0605	reservoir | (s−k)
1	-14.0605	stéphane | meystre
1	-14.0605	(laberge | samuels,
1	-14.0605	eu | amida
1	-14.0605	extractions | woepos’s,
1	-14.0605	kawahara, | shikano,
1	-14.0605	pi(u) | pi(v)
1	-14.0605	metaphone | (philips,
1	-14.0605	♠ | ♥
1	-14.0605	0.276 | acc.
1	-14.0605	micinn | multimodal
1	-14.0605	helikopter | lading
1	-14.0605	setups | bleur4n4
1	-14.0605	lobster, | spaghetti,
1	-14.0605	89.4 | 85600
1	-14.0605	f1(%) | mentions)
1	-14.0605	fighter | “allied
1	-14.0605	50.92 | 28.21
1	-14.0605	⎩ | ⎨
1	-14.0605	p(e|f) | prime,e)
1	-14.0605	institut | sprachverarbeitung
1	-14.0605	kuzman | discussions
1	-14.0605	iis-0953149 | psc-cuny
1	-14.0605	adjective, | infinitive,
1	-14.0605	distinguish | propositions:
1	-14.0605	rerank | non-local)
1	-14.0605	hash | srilm-h,
1	-14.0605	chemical | two-component-system
1	-14.0605	air | tidal
1	-14.0605	“interactions | c”
1	-14.0605	negatively | latency
1	-14.0605	0.563 | wordends
1	-14.0605	tagging: | option.
1	-14.0605	thefirstandlastwordsofthegivennpin | aredefinitenp;elsen
1	-14.0605	(n.*|j.*)
1	-14.0605	grpip | grpilr
1	-14.0605	roget | positives
1	-14.0605	absolutely | annotate
1	-14.0605	nullpos | null50
1	-14.0605	(prepositions | (numerals),
1	-14.0605	ω1 | ω2
1	-14.0605	agirre, | alfonseca,
1	-14.0605	nadya=ko | yad
1	-14.0605	he, | gosforth,
1	-14.0605	multicatalytic | cofactor
1	-14.0605	address: | way,
1	-14.0605	{jianxing, | chuats}@comp.nus.edu.sg
1	-14.0605	fusion | paths,
1	-14.0605	67.1 | 57.5
1	-14.0605	fortaleza | fortaleza,
1	-14.0605	cp-nonel | ip-nonel
1	-14.0605	plan-elab | (2e)
1	-14.0605	(las) | (uas)
1	-14.0605	eof | ubbl
1	-14.0605	24.64 | 24.21
1	-14.0605	0-9 | “-”;
1	-14.0605	æ | (tak)
1	-14.0605	18-22 | 38-42
1	-14.0605	48.2 | 47.4
1	-14.0605	-measure | 93.01%.
1	-14.0605	consistent, | consistent
1	-14.0605	u. | valencia,
1	-14.0605	“now | crude
1	-14.0605	u,t | ,t
1	-14.0605	wr | br87
1	-14.0605	obama | condemned
1	-14.0605	comprises | indexing;
1	-14.0605	thesefeatures | 42.66%
1	-14.0605	30-35 | 35-40
1	-14.0605	downtown | lighting
1	-14.0605	deceptive | spam—fictitious
1	-14.0605	p(w|t).
1	-14.0605	‘null’ | ‘gen’
1	-14.0605	86.25 | nf-negative
1	-14.0605	v, | θ(w,v)
1	-14.0605	|vd(fn)|
1	-14.0605	)|straight(
1	-14.0605	aimed | hprd50
1	-14.0605	[k,lscript) | α(k,lscript)
1	-14.0605	ingthegoldstandardmentionboundaries,wemean weusethegoldstandardnounphraseboundaries.) | aftergettingthe
1	-14.0605	lower-order | (katz,
1	-14.0605	21 | knowledge-based,
1	-14.0605	informatica | dip.
1	-14.0605	se | s˚a
1	-14.0605	0.91 | 0.450.5
1	-14.0605	(haveliwala, | teleport
1	-14.0605	bernstein | ratner,
1	-14.0605	2,909 | 1,546
1	-14.0605	(al-sughaiyer | al-kharashi,
1	-14.0605	singular | plural;
1	-14.0605	72 | 74–82,
1	-14.0605	ax | ay,
1	-14.0605	trad | trad+ind
1	-14.0605	like | amadeus
1	-14.0605	boyer, | e.y.
1	-14.0605	conj | -dep→
1	-14.0605	s51s48 | s49s48s48s37
1	-14.0605	“stretchy | modeling
1	-14.0605	universidade | aplicada
1	-14.0605	genetic | multiobjective
1	-14.0605	min(|c(e)∩c(e prime)|,|l(e)∩l(eprime)|)
1	-14.0605	mbr | (arg
1	-14.0605	lmax, | term
1	-14.0605	rocked | mindoro
1	-14.0605	hotel | hilton
1	-14.0605	wikitopics | fifteen
1	-14.0605	prec | wds
1	-14.0605	dinu | substitution;
1	-14.0605	wal | schip
1	-14.0605	la | deber
1	-14.0605	onstar, | vue’s
1	-14.0605	fa8750-09-c-3705 | iis0429360.
1	-14.0605	10,576 | 864
1	-14.0605	carpenter, | woley,
1	-14.0605	(lu | zhai,
1	-14.0605	früh:morning | morgen:morning
1	-14.0605	err | no-pre
1	-14.0605	hgood: | hbad:
1	-14.0605	cotb | cotc
1	-14.0605	pre-order | phylogeny:
1	-14.0605	x: | (w*)
1	-14.0605	worlds | non
1	-14.0605	christos | granroth-wilding,
1	-14.0605	soviet | doklady,
1	-14.0605	1984 | (erjavec,
1	-14.0605	vms | 151%
1	-14.0605	su, | x∈x(su)
1	-14.0605	jin | (ptm)
1	-14.0605	data-set | 1709
1	-14.0605	lf | msa,
1	-14.0605	asking, | interactions”
1	-14.0605	celct | livememories
1	-14.0605	yj, | θjk
1	-14.0605	hasspouse | gpeleader
1	-14.0605	venturi | moruzzi,
1	-14.0605	<trans | d="1"
1	-14.0605	women | joy–sadness
1	-14.0605	rewritten | atxy
1	-14.0605	prague | (padt)
1	-14.0605	non-existence | extension.
1	-14.0605	bats | ”won
1	-14.0605	journalists | tortured
1	-14.0605	discharge | obese
1	-14.0605	waltham, | 02452
1	-14.0605	<proptext | prohibition
1	-14.0605	hamnosys | signwriting
1	-14.0605	human1 | human2
1	-14.0605	prn | conjp
1	-14.0605	concordu | predx
1	-14.0605	(hickl | bensley,
1	-14.0605	forze | agenti.
1	-14.0605	nullij | nulli
1	-14.0605	)= p(θ|x
1	-14.0605	snow, | jurafsky,
1	-14.0605	evokes | him”
1	-14.0605	xx-enter | itg.
1	-14.0605	sociolinguistic | nyquist
1	-14.0605	july/sa | subway/d
1	-14.0605	open | keynes
1	-14.0605	4productdisplay | parenleftbiggm
1	-14.0605	r): | tijij
1	-14.0605	52.12 | 46.63
1	-14.0605	vi | vj:
1	-14.0605	genderagreement | verbagreement
1	-14.0605	shankar | wolfgang
1	-14.0605	kpath, | kbow,
1	-14.0605	epsrc | ep/g051070/1
1	-14.0605	+lem | +lem+na
1	-14.0605	field | brien
1	-14.0605	cluster-level | none-xb
1	-14.0605	z,z−j,f−j,β) | ·,z
1	-14.0605	trial | conviction
1	-14.0605	31.5 | 86.0
1	-14.0605	redakteur | kalten
1	-14.0605	steedman, | normal-form
1	-14.0605	iter | carattere
1	-14.0605	discrepancies | gold-standard
1	-14.0605	berkeley, | 94720
1	-14.0605	en-es | en-gr
1	-14.0605	paraphrases | jailed,
1	-14.0605	african | hispanics,
1	-14.0605	(mainguy, | villemonte
1	-14.0605	0.690 | 0.949
1	-14.0605	authoremail | authoraddress
1	-14.0605	afii62772afii62821afii62817ا | tl:
1	-14.0605	thierry | umr8094 cnrs
1	-14.0605	good* | market*
1	-14.0605	coffee, | soup,
1	-14.0605	litman | lrdc
1	-14.0605	chordates | eutherians
1	-14.0605	madrid, | anselmo@lsi.uned.es
1	-14.0605	multimodal | (mmil)
1	-14.0605	loga092de | base092de
1	-14.0605	|pi|
1	-14.0605	schunn | generously
1	-14.0605	lp-mert | 1dmert
1	-14.0605	´etudes | normale
1	-14.0605	(gi) | polarity.4
1	-14.0605	87.3 | nflgame
1	-14.0605	leave | blank,
1	-14.0605	thank | maintainers
1	-14.0605	continuity | −∩−
1	-14.0605	wl | lew
1	-14.0605	bleu-4 | (papineni
1	-14.0605	gerundive | be-verb
1	-14.0605	narrative | knowledgefor
1	-14.0605	지난 | jinan
1	-14.0605	sta | tsy
1	-14.0605	logios | pseudocorpus
1	-14.0605	400,000, | events.
1	-14.0605	texting, | social
1	-14.0605	reevoo.com | gsmarena.com.
1	-14.0605	president, | ecsc
1	-14.0605	william | wpc02009@pomona.edu
1	-14.0605	infon | fun,
1	-14.0605	london/sa | a/s
1	-14.0605	per-document | per-topic
1	-14.0605	(ttsa-a-te, | 5
1	-14.0605	di | dj)
1	-14.0605	agirre, | aitor
1	-14.0605	abtθjhxdðrzsšsdt | ý
1	-14.0605	evidences | ifg(eprimes)
1	-14.0605	øk | gayi
1	-14.0605	alphabet | σi,e
1	-14.0605	(b), | (e),
1	-14.0605	sigir | workshop.
1	-14.0605	privacy | westin
1	-14.0605	3,4the | 9,10continue
1	-14.0605	α↓i | ui1a
1	-14.0605	(lang, | tomita,
1	-14.0605	wine | vino:
1	-14.0605	mihalcea, | baccianella
1	-14.0605	 on |  meat.
1	-14.0605	sopn | aggregations
1	-14.0605	(tune) | newssyscombtest
1	-14.0605	departmentof | macquarieuniversity
1	-14.0605	wie | wahrscheinlichkeit,
1	-14.0605	ipus | hold-preceding
1	-14.0605	1k | 27.43
1	-14.0605	januar | obligatorisch
1	-14.0605	53.05% | evaluationresultson
1	-14.0605	abstractness(implicit | repair(yes
1	-14.0605	amianto | cloreto
1	-14.0605	delmas | pou
1	-14.0605	eva | vecchi
1	-14.0605	italianfigurecaption | italiansummary
1	-14.0605	bales* | middling*
1	-14.0605	production | null1nullnullnullnullnullnullr(g)
1	-14.0605	irland | weil
1	-14.0605	telegraph | guadalajara,
1	-14.0605	seinem | spannenden
1	-14.0605	“a/b | “bind”
1	-14.0605	i(pos(x) | pos(y))
1	-14.0605	(ben-avi | winter,
1	-14.0605	missing="yes"> | offset="-1">
1	-14.0605	involvement | society.
1	-14.0605	amaro, | chaves,
1	-14.0605	opinion | ntcir-7.
1	-14.0605	bg | n-gram+syntactic
1	-14.0605	approach | niehues
1	-14.0605	〈source〉(c | ...
1	-14.0605	truthful | 400
1	-14.0605	altaf | tx
1	-14.0605	p(t|c).
1	-14.0605	`aob¯amˇa | yˇuyˇı
1	-14.0605	ye | baldwin’s
1	-14.0605	ofv | whichv
1	-14.0605	u* | jj+
1	-14.0605	’’ | −rrb−
1	-14.0605	pasqua, | pepsi,
1	-14.0605	p(θ,b|x
1	-14.0605	noordman | blijzer
1	-14.0605	obama | 4975
1	-14.0605	f-score(%) | 75.33
1	-14.0605	2000 | descriptionof
1	-14.0605	(cohen | hunter,
1	-14.0605	taiwan | jiangsu
1	-14.0605	16.51 | 17.28
1	-14.0605	‘fantastic’ | gerçek
1	-14.0605	[into | bedroom]
1	-14.0605	bojar | tamchyna
1	-14.0605	clothes | vapor.’
1	-14.0605	outer | space”
1	-14.0605	sie | soldbedingungen
1	-14.0605	collobert | weston
1	-14.0605	manchester | (horridge
1	-14.0605	[quedó | hueco
1	-14.0605	say | authors)
1	-14.0605	,〉,int | ,imp
1	-14.0605	(either | automatically)
1	-14.0605	hartu | kontu
1	-14.0605	ea | (eh)
1	-14.0605	jrip | unigram,unicue
1	-14.0605	guangyou | beijing
1	-14.0605	social | endless
1	-14.0605	ut | p-missing
1	-14.0605	beijing | (peking
1	-14.0605	women, | men)
1	-14.0605	argm-dir | argm-prp
1	-14.0605	etter | btv_
1	-14.0605	(rumelhart | mcclelland,
1	-14.0605	ν− | ν+
1	-14.0605	university | lixxx345@umn.edu
1	-14.0605	θv,w | ρr
1	-14.0605	(beam-predict) | chart
1	-14.0605	p(t,w|w∗) | pc(i)
1	-14.0605	hovy | hovy}@isi.edu
1	-14.0605	sofie | peer1,2
1	-14.0605	indefinite, | appositive,
1	-14.0605	kipper, | anna
1	-14.0605	implementations: | (lsi).
1	-14.0605	leaked, | poured,
1	-14.0605	p(xi,j|yi,j;θ)
1	-14.0605	trans4 | trans3.
1	-14.0605	predicatingresultsfor | pairs,weuseclosest-firstclusteringtolinkthemand
1	-14.0605	patiala, | 9-11,
1	-14.0605	auch | schlechte
1	-14.0605	ks | wn+wk
1	-14.0605	1892 | gazette*
1	-14.0605	o-dep | {mr},
1	-14.0605	markup | <tm
1	-14.0605	yt2 | 2(logs1)
1	-14.0605	sectionheader | italiantablecaption
1	-14.0605	inist | mig
1	-14.0605	woz-07 | 68.3%
1	-14.0605	“(” | “)”.
1	-14.0605	gmds | 0.327
1	-14.0605	v1 | l(2)
1	-14.0605	(5) | mineditedcost>0
1	-14.0605	bren | .allies,
1	-14.0605	decision-related | (drdas)
1	-14.0605	l02 | l04
1	-14.0605	∗this | stellenbosch
1	-14.0605	dirichlet | (ferguson,
1	-14.0605	porto | porto,
1	-14.0605	p(f|e) | l(f|e)
1	-14.0605	maju-odeon | „an
1	-14.0605	69.11% | 75.59%
1	-14.0605	lid=''l1'' | relatedtotime=''t1''
1	-14.0605	87.6% | 89.4%
1	-14.0605	p(cm,n |
1	-14.0605	1/1 | citizens
1	-14.0605	familiarity | involvement
1	-14.0605	hyperplane | halfspace,
1	-14.0605	.70 | .74
1	-14.0605	n: | place’,
1	-14.0605	gmds: | (wan
1	-14.0605	cause-effect, | instrument-agency,
1	-14.0605	daelemans, | zavrel,
1	-14.0605	gmbh | saarbr¨ucken,
1	-14.0605	0.7054 | 0.6141
1	-14.0605	orange | being
1	-14.0605	review | exclamatory
1	-14.0605	randomly-selected | composed
1	-14.0605	cfv=3,h=∞ | cfv=3,h=2
1	-14.0605	telecommunications | vienna,
1	-14.0605	l2c2, | lyon,
1	-14.0605	0.024 | 0.038
1	-14.0605	nkqm@5 | 0.6688
1	-14.0605	(s,t), | φdice(su,s,t)
1	-14.0605	ans | crusade
1	-14.0605	herzog | wird
1	-14.0605	hand<adv> | inf{
1	-14.0605	〈num〉sg) | collection∧
1	-14.0605	x1 | 〈[1,4],e〉
1	-14.0605	limitations | orientation)
1	-14.0605	decision-related | rouge-f1
1	-14.0605	good* | low*
1	-14.0605	age, | founder,
1	-14.0605	<person | <company
1	-14.0605	anja | bn2
1	-14.0605	thrgiza | thrupdate
1	-14.0605	“for” | “against”
1	-14.0605	max | zock
1	-14.0605	history, | alignment(proportion)
1	-14.0605	→ | {pr(cn
1	-14.0605	sv1 | sv2
1	-14.0605	duluth, | 55812
1	-14.0605	0.5809 | 0.5835
1	-14.0605	graeme | gh@cs.toronto.edu
1	-14.0605	donthave_incorr | pos_out
1	-14.0605	vl | null=
1	-14.0605	(rubenstein | goodenough,
1	-14.0605	(term | wiki200;
1	-14.0605	(merged | ptkrel:
1	-14.0605	word-expert | (daelemans
1	-14.0605	cnt(w,t,ci) | max(cnt(w,t,ri),cnt(w,t,ci))
1	-14.0605	cij, | 〈si,cij〉,
1	-14.0605	dsm-iv-tr | psychiatric
1	-14.0605	refsum | 0.75694
1	-14.0605	past: | [[stmt(β,α,at(β,λ),{})]]c
1	-14.0605	coreferent | disreferent
1	-14.0605	=10null(pizza’(null), | nullnullnullnull[null(null,
1	-14.0605	aline+base | m2m+base
1	-14.0605	))·˜p(wi|bon−1(wi−11
1	-14.0605	),,( | αθθ
1	-14.0605	o: | por[by]
1	-14.0605	(2002) | internal,
1	-14.0605	interinvlist | invlist
1	-14.0605	authorities | stop
1	-14.0605	potency | visualizations.
1	-14.0605	diderot, | f-75205,
1	-14.0605	count | (z,w)
1	-14.0605	fun | playing
1	-14.0605	䤕 | vv
1	-14.0605	talks | r7
1	-14.0605	early-stopping | (a1h),
1	-14.0605	corporation | nsc99-2911-i-002-001,
1	-14.0605	wsj-only | wsj-only
1	-14.0605	yujie | technology
1	-14.0605	kendall’s | metricsmatr’s
1	-14.0605	thenimplementa | whethertwopre-clustersrefertothesameentity.
1	-14.0605	mckeown | linda
1	-14.0605	(ptk) | trees.
1	-14.0605	0.1337 | 0.1885
1	-14.0605	’found’, | ’specific’,
1	-14.0605	21.3 | 180
1	-14.0605	ψ | andy,
1	-14.0605	dies | tasche.
1	-14.0605	uce | litat
1	-14.0605	x | 〈e1e2,f2f1〉.
1	-14.0605	umr | nantes
1	-14.0605	coord↑↓, | sub↑↓,
1	-14.0605	republic | (jiangsu,
1	-14.0605	kumamoto | osaka
1	-14.0605	wordsim | combined-to-text+
1	-14.0605	scoremsd | contextmsd
1	-14.0605	(mt | phuket,
1	-14.0605	c8 | c5,
1	-14.0605	s52 | s102s32
1	-14.0605	microsoft | 94041
1	-14.0605	as, | kneserney),
1	-14.0605	s4 | “mcep
1	-14.0605	i;t | )z
1	-14.0605	 ),|in
1	-14.0605	j | xms
1	-14.0605	markov | (mcl)
1	-14.0605	⊕f(d(v1,e),...,d(v|e|,e))
1	-14.0605	hiroshi | masakazu
1	-14.0605	0.15 | (-ngram)
1	-14.0605	(verbs), | det
1	-14.0605	kitchen | rust
1	-14.0605	superstar: | slimy
1	-14.0605	saction | intelligenc
1	-14.0605	30,000 | (percentages
1	-14.0605	speling | spech
1	-14.0605	token | constituent’s
1	-14.0605	itg-based | conquer
1	-14.0605	charniak, | blaheta,
1	-14.0605	ccm | distituent
1	-14.0605	g|α0,p0 | dp(α0,p0)
1	-14.0605	pseudo-documents | (adjectives
1	-14.0605	sprime | pg(s)
1	-14.0605	ˆy∈r | xlatticetopβ
1	-14.0605	namesor | whichcan
1	-14.0605	的 | 经济
1	-14.0605	(stein | vendrame
1	-14.0605	|g|.
1	-14.0605	retirement | 18,
1	-14.0605	trial | between-participant
1	-14.0605	perpetrator: | instrument:
1	-14.0605	extrapolation | needed)
1	-14.0605	influence | gravity’.
1	-14.0605	nacn | nun
1	-14.0605	t0 | λu
1	-14.0605	(2c) | (2d)
1	-14.0605	orthogonal | f,g
1	-14.0605	cd) | vp)
1	-14.0605	suzuki, | dickinson,
1	-14.0605	tongue | poly
1	-14.0605	angel | spitkovsky,
1	-14.0605	and, | realistically,
1	-14.0605	immediate | utterances).
1	-14.0605	403 | c´edex
1	-14.0605	82.64 | 82.42
1	-14.0605	25.6 | 28.9
1	-14.0605	synsetcj | synseta
1	-14.0605	(1)⇐⇒ | (2)⇐⇒
1	-14.0605	government | cosmat
1	-14.0605	passenger | frontand
1	-14.0605	varieties, | syndication
1	-14.0605	dialogic | 63%.
1	-14.0605	inst(hovy | instinwordnet
1	-14.0605	“indian | govt.
1	-14.0605	ˆetre | pense
1	-14.0605	degres | fredom
1	-14.0605	drug | (39)
1	-14.0605	markinf | markpart
1	-14.0605	topics | 0.80,
1	-14.0605	>141 | <title
1	-14.0605	italianabstract | englishabstract
1	-14.0605	triage | ie-based
1	-14.0605	limi→∞ | (0,0,...,0),...,
1	-14.0605	unit, | kessler,
1	-14.0605	2; | parallel;
1	-14.0605	k | wk...w1,
1	-14.0605	topwords | topwordsdocs
1	-14.0605	:class | :subject
1	-14.0605	lowell | 01854
1	-14.0605	hpsg | run.”
1	-14.0605	cardiovascular | mediastinal
1	-14.0605	month=abs; | time=abs)
1	-14.0605	le | hanno
1	-14.0605	amani, | amani
1	-14.0605	lp | 80:9%
1	-14.0605	e4 | e4:
1	-14.0605	chambre | ville
1	-14.0605	digits | hyphens,
1	-14.0605	children] | remembered.
1	-14.0605	definition | (compatibility
1	-14.0605	2-11-16 | jdkim@dbcls.rois.ac.jp
1	-14.0605	svm | snta1
1	-14.0605	acld | user.
1	-14.0605	‘g’ | ‘h’
1	-14.0605	tf | 0.181
1	-14.0605	63.6 | touchdownpartialcount
1	-14.0605	ax | bx
1	-14.0605	uproar | cudahy,
1	-14.0605	t29 | t28,
1	-14.0605	ofxgiven | p(x|y;θ)
1	-14.0605	traced, | 1700-1750,
1	-14.0605	steedman, 2007), | normalform
1	-14.0605	2011a) | (song
1	-14.0605	light | ph
1	-14.0605	five-man | yemen
1	-14.0605	votes | hi(seq)
1	-14.0605	i;h;jihx2;x1i;hx1;x3i | j;j;jc1ihx1;x3i;hx3;wji
1	-14.0605	null | wds
1	-14.0605	k=1 | 2∑
1	-14.0605	equality, | superlative,
1	-14.0605	cn, | dai∈cm
1	-14.0605	20012000 | 189,840
1	-14.0605	when | mentionboundariesareused,theresultsare53.41%
1	-14.0605	natalie | gwendolyn
1	-14.0605	gme | ntsc
1	-14.0605	predicate-argument | rester,
1	-14.0605	bahl | 1983).
1	-14.0605	silver | (ssc)
1	-14.0605	r(x, | domain(r)
1	-14.0605	company: | price:
1	-14.0605	“i | summer”:
1	-14.0605	fre | ocupied
1	-14.0605	psychology | sexuality
1	-14.0605	115026 | 12073
1	-14.0605	% | powered
1	-14.0605	youth | practitioners,
1	-14.0605	ωold, | wk,m,wd,n
1	-14.0605	stable, | 6.2).
1	-14.0605	pa | ive
1	-14.0605	personality | cultures.
1	-14.0605	c/tanger | spain
1	-14.0605	lovet, | reder,
1	-14.0605	(strube | (jarmasz,
1	-14.0605	u, | wj-s
1	-14.0605	8.85% | 252%
1	-14.0605	85800 | 86200
1	-14.0605	annotate | man-month).
1	-14.0605	angry | he
1	-14.0605	“gm | onstar
1	-14.0605	ae | 88.08
1	-14.0605	pezzulo | alfio
1	-14.0605	2' | 2<
1	-14.0605	rwk | x1,x2,pre(x3)
1	-14.0605	(2003) | biographic
1	-14.0605	computer | oriental
1	-14.0605	(null1 | (null2
1	-14.0605	leeds | ls2
1	-14.0605	28% | crf,
1	-14.0605	biological | spoiiif
1	-14.0605	1.001 | 15387
1	-14.0605	#word | orig
1	-14.0605	iis0546554 | iis-0910611.
1	-14.0605	(i,j,r)+ | bp,q=1
1	-14.0605	online-b | udein
1	-14.0605	ˆy)))||φ(x,y)−φ(x, ˆy)||2
1	-14.0605	score(ci,−→w) | builtpps.remove(bestarc.child)
1	-14.0605	tprime≤t≤|e|, | bsprime,t,u,vprime,
1	-14.0605	70.7% | 63.0%
1	-14.0605	branching | hb(seq)
1	-14.0605	vectorr1 | vectorr2,
1	-14.0605	throwing | crowding
1	-14.0605	86 | 64
1	-14.0605	'80 | '03
1	-14.0605	logp(sk|wi;r,ψ,bc).
1	-14.0605	before, | after.
1	-14.0605	(jarmasz | szpakowicz,
1	-14.0605	purple | 1.3
1	-14.0605	represented | lang
1	-14.0605	null | interprets
1	-14.0605	naval | warfare
1	-14.0605	isi | 252,759
1	-14.0605	macquarie | (msol)
1	-14.0605	multiply | c1i
1	-14.0605	58.3 | 72.2
1	-14.0605	sprimesprimeprimesprimeprimeprime | (∃sprimeprime,sprimeprimeprime)
1	-14.0605	bc | .000
1	-14.0605	200k | 23.67**
1	-14.0605	raket | ‘fire’
1	-14.0605	bestarc | ciepsilon1candidatearcs
1	-14.0605	aredemonstrative | datafilepublishedbybergsmaandlin(2006)
1	-14.0605	c=correct, | f=fscore
1	-14.0605	98195, | hlin@ee.washington.edu
1	-14.0605	θl+gq | 5.
1	-14.0605	deleted | prep(mn),
1	-14.0605	1k | |nl,iq
1	-14.0605	franz | och1
1	-14.0605	1g | 9-months
1	-14.0605	binary+patterncoarse+pattern | mention)
1	-14.0605	scene | [torralba
1	-14.0605	clinton, | crowley,
1	-14.0605	st´ephane | fabrice
1	-14.0605	(verb | harvesting)
1	-14.0605	backhand | two-handed
1	-14.0605	k|wi−1,...,wi−n+1)
1	-14.0605	josé | louçã
1	-14.0605	e3 | e4,
1	-14.0605	npj; | npk−1.
1	-14.0605	(352) | (31)
1	-14.0605	strongest | black.
1	-14.0605	e,a | ¯e,¯a
1	-14.0605	mentionsarethesamestring;elsen | thementionsagreeinnumber;elsen
1	-14.0605	hugo | vain
1	-14.0605	inverse | consulted,
1	-14.0605	hond | heeft
1	-14.0605	successful | 777
1	-14.0605	1,500 | 1,935
1	-14.0605	typed | conquer(country,city)
1	-14.0605	it | os/o
1	-14.0605	78% | dosmac
1	-14.0605	type(explorative/hesitant/medium) | type(destination
1	-14.0605	stipulates | anonymous)
1	-14.0605	arguments | propagation.
1	-14.0605	(λ,µ) | δt,bindsummationtextp,q
1	-14.0605	0.4685 | 0.3784
1	-14.0605	48.7 | 56.9%
1	-14.0605	thrown | arrested,
1	-14.0605	20*1 | vbx
1	-14.0605	typographical | incompletely
1	-14.0605	dies | lovers
1	-14.0605	arteriolar | nephrosclerosis;
1	-14.0605	arity | 45.14%
1	-14.0605	lazy | prefaulting
1	-14.0605	ma | xi1
1	-14.0605	abbott, | bowmani,
1	-14.0605	(en), | (de),
1	-14.0605	log-probability | log-probability
1	-14.0605	signs | caves
1	-14.0605	sorry, | wow,
1	-14.0605	p(rp|r)
1	-14.0605	0.4608 | 0.6248
1	-14.0605	tcrf | trr
1	-14.0605	tolerates | target-language-specific
1	-14.0605	s-patterns | npatterns
1	-14.0605	root-morph | “morph”
1	-14.0605	mainegaron | laaiisenson
1	-14.0605	ø | ςmal
1	-14.0605	xdiamsa | xdia
1	-14.0605	translates | persia
1	-14.0605	deeply | language-generation
1	-14.0605	next | wkk−n+2
1	-14.0605	p(cd(wi)|h,c1:d−1).
1	-14.0605	(64%) | (36%).
1	-14.0605	saldo | (borin
1	-14.0605	wolf | t-shirt
1	-14.0605	earth | (emd),
1	-14.0605	hyponyms, | holonyms,
1	-14.0605	e∪{i | ∃inegationslash=
1	-14.0605	suny | buffalo,
1	-14.0605	grosz | attentional
1	-14.0605	hum | corr
1	-14.0605	superior | ˜oes,
1	-14.0605	inverted | (il)
1	-14.0605	print-oriented | prose,
1	-14.0605	0u��t | [4,6)
1	-14.0605	reflexes | perlocutionary
1	-14.0605	bin-2 | 33.18
1	-14.0605	62.15 | 17.77
1	-14.0605	action: | 1-a-iv,
1	-14.0605	observed | we(q,e),
1	-14.0605	s49s53 | s116s111s112s32s107
1	-14.0605	(ap) | 1r
1	-14.0605	one | 1893
1	-14.0605	hbi | outperformed
1	-14.0605	litkowski | hargraves
1	-14.0605	object-bd | localization-city
1	-14.0605	71.8 | 77.2
1	-14.0605	concordances | anklaga
1	-14.0605	0.6088 | 0.5355
1	-14.0605	system/human | interchangeability:
1	-14.0605	recovery | dubey,
1	-14.0605	(lo | 2011a,b)
1	-14.0605	93.55 | 12.5k
1	-14.0605	tanmoy | saikh,
1	-14.0605	(klementiev | hermjakob
1	-14.0605	y(q) | q:
1	-14.0605	reranking | syntbip*
1	-14.0605	#words | 75.12
1	-14.0605	medtronic" | firstname:
1	-14.0605	(ferreira, | (brasoveanu,
1	-14.0605	‘clean’ | ‘raw’
1	-14.0605	classificationmodelcombinedwith | nounphraseis
1	-14.0605	parameters | dev+test
1	-14.0605	meetings | utterances)
1	-14.0605	fb2(rs2 | rt2)
1	-14.0605	slippery | raining
1	-14.0605	und | drei
1	-14.0605	(about | hundred
1	-14.0605	khush | kaam
1	-14.0605	surpass | praecurrere
1	-14.0605	p[pi] | pi∈e∗
1	-14.0605	theory | relativity
1	-14.0605	2500, | montr´eal
1	-14.0605	kristy | carolina
1	-14.0605	47th | (acl-ijcnlp’2009),
1	-14.0605	subsidiary | retailer,
1	-14.0605	lsa | screwball
1	-14.0605	olution, | traditinalpair-wise
1	-14.0605	km04: | saj10a:
1	-14.0605	giant | slick.
1	-14.0605	f1: | f2:
1	-14.0605	w4”,x) | pred(l,“v
1	-14.0605	42 | 43–47,
1	-14.0605	nanchen, | abstract
1	-14.0605	aux, | to][2,
1	-14.0605	sussex | 9qj,
1	-14.0605	wheeled | class-instances.
1	-14.0605	is|was”.
1	-14.0605	worst-case | average-case
1	-14.0605	(bllip) | providence,
1	-14.0605	nullsummationtextp:v2(p)=v | nullsummationtextp:v1(p)=v
1	-14.0605	context-free | mark-jan
1	-14.0605	{jj} | {nn}
1	-14.0605	0.591 | 30000
1	-14.0605	interlayer | insulation
1	-14.0605	moments | portraying
1	-14.0605	8am | 10am”)
1	-14.0605	‘in | ‘from
1	-14.0605	/ | 99.05
1	-14.0605	p-value | pn(x≥x)
1	-14.0605	“cananea” | s4,
1	-14.0605	u5(x,y,di,p) | =
1	-14.0605	89.6 | 86000
1	-14.0605	turns) | subjects.
1	-14.0605	140 | pr-as
1	-14.0605	venkataraman | fleck
1	-14.0605	summarizedin | 0.270
1	-14.0605	contacts | ligand
1	-14.0605	|ak|summationdisplay | |al|summationdisplay
1	-14.0605	p(r1 | p(r2
1	-14.0605	carbonell | goldstein
1	-14.0605	0..i,pi,j,wj..|w|
1	-14.0605	(>5counts) | 75.43
1	-14.0605	89.11 | 87.07
1	-14.0605	subsequence | (lodhi
1	-14.0605	data | punctuation-based
1	-14.0605	attempt | (pad´o
1	-14.0605	ceca, | devront
1	-14.0605	72.7% | 45.5%
1	-14.0605	pointing | torso
1	-14.0605	ssa | 0.6179
1	-14.0605	usr | π10
1	-14.0605	”w” | n∈nwords
1	-14.0605	gazetteers | geonames
1	-14.0605	renters, | americans,
1	-14.0605	{{ | ?translation
1	-14.0605	sarcasm, | fact/feeling
1	-14.0605	either | kinds:
1	-14.0605	shb | (σ,k|β,a)
1	-14.0605	1|k).
1	-14.0605	holger | christophe
1	-14.0605	rottmann | reorderings.
1	-14.0605	“ctb” | 9.4m
1	-14.0605	multi-view | (mvm),
1	-14.0605	‘ | «/
1	-14.0605	lig | 5217
1	-14.0605	cutting | papers,
1	-14.0605	a-b | b-c
1	-14.0605	finley | badecker,
1	-14.0605	composition | penn,
1	-14.0605	≈arg | αn·
1	-14.0605	33.4 | 40.9
1	-14.0605	<=-5 | >=5
1	-14.0605	p(b), | ,t(b)1
1	-14.0605	propcontent | ab´en´eficiaires
1	-14.0605	sentence-spanning | sequences
1	-14.0605	i∈k | c=k+1
1	-14.0605	change | members:
1	-14.0605	anotate | ambitious.
1	-14.0605	(wps), | (pro),
1	-14.0605	durango | journal,
1	-14.0605	× | btngs,
1	-14.0605	chap-6 | eval-2
1	-14.0605	bow, | latter,
1	-14.0605	anigal | pottigal
1	-14.0605	i< | i++)
1	-14.0605	piat | solche
1	-14.0605	israel | (grants
1	-14.0605	fraser | burnaby,
1	-14.0605	pφ(x|y).
1	-14.0605	zhao, | chunyu
1	-14.0605	clustering: | iteratively
1	-14.0605	s(e,a,b) | ei,t=1
1	-14.0605	pos_out_corr | pos_out_incorr
1	-14.0605	productive | n”,
1	-14.0605	nt1 | nt2
1	-14.0605	mterater | fluency-based
1	-14.0605	teragrid | tg-dbs110003.
1	-14.0605	formants | resonances
1	-14.0605	(gi−gu) | (hj−hv)
1	-14.0605	x2, | x2)
1	-14.0605	bought | opel”,“gm
1	-14.0605	―q‖ | ―هق
1	-14.0605	loyalists | republicans
1	-14.0605	m. | (c2)
1	-14.0605	ve, | (y|x)
1	-14.0605	if|occ(t1)|= | if|occ(t)|>
1	-14.0605	mwes | (jackendoff,
1	-14.0605	minimizing | parenleftbigparenleftbig
1	-14.0605	comunity | years
1	-14.0605	strongly | r(28)
1	-14.0605	pθo(ot |ht)
1	-14.0605	〈type,subsetsqequal〉 | approp
1	-14.0605	(2009; | morphologicalvariants
1	-14.0605	rae | method)
1	-14.0605	jr., | meatpacking
1	-14.0605	power | handbok
1	-14.0605	s111 | s116s101
1	-14.0605	pollard | sag.
1	-14.0605	post-processing | bioenex
1	-14.0605	φprimei(y) | φprimej(n)
1	-14.0605	pt0 | pt2
1	-14.0605	fogg | bagheera
1	-14.0605	(hripcsak | rothschild,
1	-14.0605	onomatopoeic | mimetic
1	-14.0605	hendrix) | db:4
1	-14.0605	low), | explicit),
1	-14.0605	castro-bleda | prat,
1	-14.0605	yuseong-gu, | daejeon,
1	-14.0605	hall, | kravalova,
1	-14.0605	(sabidussi, | freeman,
1	-14.0605	h | hprime
1	-14.0605	blonder | laboratories
1	-14.0605	ace05 | (ace05,
1	-14.0605	12693 | 6194
1	-14.0605	(jp), | (fr),
1	-14.0605	51.2 | 52.1
1	-14.0605	152 | 153
1	-14.0605	“accept” | “acknowledge”
1	-14.0605	≥ | g(i)ak(z1,z2,...,z|n|)
1	-14.0605	npj | npj,
1	-14.0605	secondary | [2,5],
1	-14.0605	saar | q22:
1	-14.0605	uni-grams: | bi-grams:
1	-14.0605	rater, | (xj)
1	-14.0605	dale | reiter’s
1	-14.0605	replicator | (schuster
1	-14.0605	v(t,u) | c(t,u)c(v,u)∗log
1	-14.0605	p(y | βixi
1	-14.0605	explorations | thesaurus
1	-14.0605	adjacent | (m,n)
1	-14.0605	mitac | toll-free
1	-14.0605	takamura | takashi,
1	-14.0605	broke | (1a,b),
1	-14.0605	tempeval-2010 | (kolya
1	-14.0605	#h1n1 | #musicmonday
1	-14.0605	|sk|
1	-14.0605	coreference: | (muc
1	-14.0605	pe˜nas | 28040
1	-14.0605	rep | 2par.
1	-14.0605	stein | ormel
1	-14.0605	0.1622 | thresholds.
1	-14.0605	response: | φik
1	-14.0605	e.g., | (⊑
1	-14.0605	001 | sup´erieure,
1	-14.0605	reads | citations,
1	-14.0605	mbdp-1 | 0.696
1	-14.0605	manuscript | illustrations,
1	-14.0605	slovak | germans
1	-14.0605	bits, | n−1summationdisplay
1	-14.0605	gundel, | zacharski,
1	-14.0605	p,p | psss,
1	-14.0605	are/s | identifying/ga
1	-14.0605	plurals | demonstratives
1	-14.0605	holloway | egham,
1	-14.0605	0.403 | 0.662
1	-14.0605	cognitives | bron,
1	-14.0605	u | vl◦vr
1	-14.0605	company’s | allocation]
1	-14.0605	fixed-gap | flexible-gap
1	-14.0605	72085 | firstname.lastname@lium.univ-lemans.fr
1	-14.0605	<event | kybot="phosphorylation
1	-14.0605	ij | v;
1	-14.0605	iesb | iono
1	-14.0605	zusammenstoß | bagatellisieren.
1	-14.0605	973,902 | 10,597
1	-14.0605	0.7477 | 0.7916,
1	-14.0605	modelβ | +···+
1	-14.0605	arabic-english | data,and
1	-14.0605	evaluation | (lp/up)
1	-14.0605	b-2000 | belgium
1	-14.0605	l, | θ.
1	-14.0605	np-complete, | straightforwardreductionfromtherecognitionproblem
1	-14.0605	disruptive | ongoing
1	-14.0605	88.59% | 86.81%
1	-14.0605	als | ady
1	-14.0605	(norm) | 0.180
1	-14.0605	general | farr´us
1	-14.0605	(open | (closed
1	-14.0605	mumbai, | 400076
1	-14.0605	ˆt | lc∈t
1	-14.0605	refinements | degradation.
1	-14.0605	‘time’, | ‘region’.
1	-14.0605	decorates | clusters,
1	-14.0605	software, | (eternals).
1	-14.0605	substitution | (ssrs)
1	-14.0605	"medical | #defibrillators,
1	-14.0605	allusion | 76.64
1	-14.0605	song-title(smile) | suchst
1	-14.0605	licenses | dative
1	-14.0605	spatial | bruteforce
1	-14.0605	log2(k | 1)-bit
1	-14.0605	ast | beam,
1	-14.0605	4.60 | 4.72
1	-14.0605	rapidly | 20–30
1	-14.0605	<role | rtype="theme"
1	-14.0605	ramisch♠ | ♠institute
1	-14.0605	indefinite | bemerkungen’
1	-14.0605	dcs | semresp
1	-14.0605	chinese | 在
1	-14.0605	98.1 | ole
1	-14.0605	combinedwith | an
1	-14.0605	ontonotesconstitutesone | integrated
1	-14.0605	s55s56 | out–of–domain
1	-14.0605	(y|x),
1	-14.0605	subordinate | super-ordinate
1	-14.0605	nn|ae13 | nn|dd08
1	-14.0605	(ce1, | (ce2,
1	-14.0605	lc | outperform
1	-14.0605	und | teurer
1	-14.0605	(r-1), | (r-2),
1	-14.0605	qin | stephan.vogel}@cs.cmu.edu
1	-14.0605	s57s48s37 | s114s97
1	-14.0605	npk | arg0arg0,
1	-14.0605	s3wc, | s1uwc,
1	-14.0605	(f¨ul¨op | vogler,
1	-14.0605	[c1 | struggling
1	-14.0605	bigrams | 62.02
1	-14.0605	m5 | m9
1	-14.0605	interrogative | wor¨uber
1	-14.0605	(t, | ct)
1	-14.0605	dznullvnieks, | knulldu
1	-14.0605	sistemi | firenze
1	-14.0605	cfv=∞,h=1 | cfv=4,h=∞
1	-14.0605	chat | fantasy
1	-14.0605	‡human | den@cs.umd.edu
1	-14.0605	+0.20 | +0.39
1	-14.0605	yellow, | orange),
1	-14.0605	25.59 | 23.73
1	-14.0605	0.256 | 0.081
1	-14.0605	λ4, | [i|λ2],
1	-14.0605	courtneynapoles1 | benjaminvan
1	-14.0605	k-prefix | k-prefix.
1	-14.0605	srilm’s | berkeleylm’s
1	-14.0605	k,b,c | i(goal)
1	-14.0605	beň | cerň
1	-14.0605	faith | ˜p(wi|boprime)
1	-14.0605	translational, | clinical
1	-14.0605	kate | kwiatkowski
1	-14.0605	decision | (summary)
1	-14.0605	alywm | mfxxtyn
1	-14.0605	(pas¸ca, | kozareva
1	-14.0605	40% | 194
1	-14.0605	hgfc | 3.2.3.
1	-14.0605	dynamics | sigmund,
1	-14.0605	=20null(child’(null), | nullnullnullnull[nullnullnullnullnull(nullnullnullnull1nullnullnull2)
1	-14.0605	m23 | m24
1	-14.0605	representation: | (vbp
1	-14.0605	shuller | al.,(2010)
1	-14.0605	and | production...
1	-14.0605	agenda, | insert(expand(candidate,
1	-14.0605	47.0 | 60.7
1	-14.0605	l(θ) | (b|φ(x,y))
1	-14.0605	sw | sw)
1	-14.0605	17,328 | d3.
1	-14.0605	prefix | (determiner)
1	-14.0605	w | ”
1	-14.0605	eyal | bar-ilan
1	-14.0605	fj1 | p(fj1
1	-14.0605	 |xi|productdisplay
1	-14.0605	person | (org)
1	-14.0605	uni0642uni0631uni0632 | fear”.
1	-14.0605	post-translational | (witze
1	-14.0605	diab | nyc,
1	-14.0605	pc | 2.4ghz
1	-14.0605	<lb | </p>
1	-14.0605	coordination | (ferrara,
1	-14.0605	obi | dduc
1	-14.0605	center | (crulp)
1	-14.0605	src | untuk
1	-14.0605	c-svm | package.
1	-14.0605	riloff | 84112
1	-14.0605	scoret(k) | +β#(c,w)
1	-14.0605	parse1 | gold2
1	-14.0605	monologue | repositories
1	-14.0605	interpolate | αp(w|θq)
1	-14.0605	adjunction | targetstockprice
1	-14.0605	n&gagreement | thementionshave
1	-14.0605	ibm1 | ibm2-inv
1	-14.0605	vol | (s¨oderwall,
1	-14.0605	| | nonfin
1	-14.0605	•for | −choose
1	-14.0605	synonym | “earphone”
1	-14.0605	ter−4bleu | newstest2009).
1	-14.0605	colour | sleep?
1	-14.0605	geld | abgebrochen
1	-14.0605	eleven | combined:
1	-14.0605	anger | disgust
1	-14.0605	gtrec | gtrec,
1	-14.0605	msa’ | hjwman
1	-14.0605	chap-5 | eval-1
1	-14.0605	mf: | obj,
1	-14.0605	difficulties | liars
1	-14.0605	flowsthrough, | instate,
1	-14.0605	avgtriangle | run-score-2
1	-14.0605	green, | grey).
1	-14.0605	die | ausl¨ander
1	-14.0605	source-hypothesis | hypothesis-source
1	-14.0605	85.79 | 86.01
1	-14.0605	r−→r|t→h) | p(l
1	-14.0605	“left” | “straight”,
1	-14.0605	andrea | dragomir
1	-14.0605	200k | 2.1m
1	-14.0605	ofcial | iarpa,
1	-14.0605	that” | that”.
1	-14.0605	jung-angseon-eul | seung-yongchaleul
1	-14.0605	phrase-emission | p(e,f|σ)
1	-14.0605	steve | ignacio
1	-14.0605	honda | wako,
1	-14.0605	d | t1,...,tk,
1	-14.0605	restricted-many-to-many | τ-one-to-one
1	-14.0605	(da), | (el),
1	-14.0605	graph-based | class-instance
1	-14.0605	mt05 | medar
1	-14.0605	i8 | i9
1	-14.0605	grammatically | “dt”
1	-14.0605	p(fk|yj)
1	-14.0605	unary(null, | nullnull←binary(˜null,
1	-14.0605	no | 0.25.
1	-14.0605	(das | antweiler
1	-14.0605	[γ1, | reindexing
1	-14.0605	15176593 | 15176589
1	-14.0605	mishne, | soboroff.
1	-14.0605	s2 | government.
1	-14.0605	frames | intensional
1	-14.0605	(sh) | (hs)
1	-14.0605	“gene” | “pathway”
1	-14.0605	log-likelihood | λnvn,
1	-14.0605	theaveragescoresofthesethreemetricsare | participatingsystems.
1	-14.0605	 | s1o2
1	-14.0605	test11 | hybrid-2011
1	-14.0605	figurative | (flp)
1	-14.0605	preliminary | notations
1	-14.0605	,n | {p(k)·
1	-14.0605	detection,we | determinewhethereach
1	-14.0605	177, | barcelona
1	-14.0605	[distrib | root]
1	-14.0605	“65 | 05
1	-14.0605	east | 100190,
1	-14.0605	(arnold | dauchet,
1	-14.0605	fs(zs) | summationtextr∈rs
1	-14.0605	ag | bielefeld,
1	-14.0605	<|y|)]
1	-14.0605	3: | 47.13
1	-14.0605	ˆpnull | defnull
1	-14.0605	(test | 44.16
1	-14.0605	thesis | fct
1	-14.0605	p(a) | ,t(b)2
1	-14.0605	74.4 | −lr
1	-14.0605	ocr | (märgner
1	-14.0605	el | pecera.
1	-14.0605	0.704 | 0.913
1	-14.0605	slope | 0.000678
1	-14.0605	genomes | metagenomes:
1	-14.0605	fan | shuqi
1	-14.0605	(ll) | phi
1	-14.0605	in-houseexperiments,the | coreferenceresolutioncan
1	-14.0605	wikiproject | irish
1	-14.0605	(p<.01 | χ2)
1	-14.0605	ppl | 151.1
1	-14.0605	vanlehn, | mit.
1	-14.0605	हमार | पधार
1	-14.0605	criticism | criticism
1	-14.0605	0.528 | 0.380
1	-14.0605	% | i-tweet
1	-14.0605	mlc | hile
1	-14.0605	(x-y) | bir
1	-14.0605	(sable | (meier,
1	-14.0605	(vm) | (m-1)
1	-14.0605	training/testing | testing/training
1	-14.0605	english↔german | english↔french.
1	-14.0605	lun-wei | hsin-his
1	-14.0605	ψ | ψ+
1	-14.0605	postprocessing | filter
1	-14.0605	82% | dx
1	-14.0605	fik | ‘granddaughters’
1	-14.0605	s; | >:
1	-14.0605	chengqing | and
1	-14.0605	quieren | hueco
1	-14.0605	1-2 | 11
1	-14.0605	a→ | →〈aa〉|〈ba〉|〈ca〉|〈ac〉|〈bc〉|〈cc〉
1	-14.0605	boldi | vigna
1	-14.0605	0.746 | 0.865
1	-14.0605	हष | (harsh)
1	-14.0605	parse | (accuracy
1	-14.0605	lscript1,∞ | (turlach
1	-14.0605	wenn | zoff
1	-14.0605	k_ho | eex
1	-14.0605	aria | aria42,
1	-14.0605	characterized | proteins.
1	-14.0605	dfki | saarbrücken,
1	-14.0605	+sib | +cov+sib
1	-14.0605	servers | w−1
1	-14.0605	rebif | aricept
1	-14.0605	“tor | 00133
1	-14.0605	nullরশম | ত
1	-14.0605	china, | 90920004.
1	-14.0605	p(x | ni(x)
1	-14.0605	cruys | moir´on
1	-14.0605	(s,t); | (s,t).
1	-14.0605	a˙ | tel
1	-14.0605	statistics-based | documentbased
1	-14.0605	“’s” | v+.
1	-14.0605	gholamreza | reza@monash.edu
1	-14.0605	turesaregivenbelow: | wordfeaturesthey
1	-14.0605	0.994 | (wc)
1	-14.0605	give(e1,x1,x2):0 | getting(e0,x1,x2):18
1	-14.0605	ptms | dna.
1	-14.0605	contextwitha | thefeaturesetdescribingci
1	-14.0605	0.82, | 0.58,
1	-14.0605	andjingjiang‡ | ‡schoolofinformationsystems,singaporemanagementuniversity,
1	-14.0605	deniz | university
1	-14.0605	window-based | dinu
1	-14.0605	lightturnstileleftn/(s|np):λ
1	-14.0605	ono | auer
1	-14.0605	2p | faust+2n
1	-14.0605	meronyms, | “see
1	-14.0605	expanded | headphones
1	-14.0605	π∗ij(s) | q∗ij
1	-14.0605	(subtype | “u.s.
1	-14.0605	pattern[j] | “vb”)
1	-14.0605	eibe | update;
1	-14.0605	(blei | moreno,
1	-14.0605	0.1191 | 0.2446
1	-14.0605	"a | $62.00
1	-14.0605	(nullnull1nullnullnullnullnullnullnulln) | nullnulli
1	-14.0605	[s | a]
1	-14.0605	raytheon | 02138
1	-14.0605	[β,γ] | pmproja([b,c])
1	-14.0605	enriched | (ettm)
1	-14.0605	s1wc, | s0cs1cs2c,
1	-14.0605	qii | qjj.
1	-14.0605	bibliography | englishkeywords
1	-14.0605	redundant, | piece
1	-14.0605	+c, | position,
1	-14.0605	levelt | kelter
1	-14.0605	deception | at-chance
1	-14.0605	sdev | sdev
1	-14.0605	normalization | b3,b2,b1∈σ:
1	-14.0605	a∈possibleactions(s0) | s1←updatestate(s0,
1	-14.0605	s55s48s37 | s99s99s117
1	-14.0605	w=0.01 | s=0.5
1	-14.0605	in–domain: | sentences)
1	-14.0605	dict | visu
1	-14.0605	.90 | .90
1	-14.0605	bach | 2009a,b)
1	-14.0605	ti2i2 | tij+1ij+1
1	-14.0605	particle | nominal;
1	-14.0605	eu-fp7 | (ict-248116)
1	-14.0605	projects | prometeo
1	-14.0605	la2 | ra2
1	-14.0605	i | it].
1	-14.0605	c.; | y.;
1	-14.0605	2001a) | rouge
1	-14.0605	 | c2s2o2
1	-14.0605	d’aides | au-del`a
1	-14.0605	:11233, | 65dec
1	-14.0605	overall | 71.12
1	-14.0605	finite | nonfinite
1	-14.0605	“70 | (70
1	-14.0605	newarc1 | newarc2
1	-14.0605	k-local | nullw
1	-14.0605	boring; | pretentious;
1	-14.0605	[i|λ4], | [λ3|i],
1	-14.0605	ntt | laboratories
1	-14.0605	changes | per-token
1	-14.0605	userusage | minimum-cost
1	-14.0605	fieldgoalpartialcount | 96.5
1	-14.0605	mobile | traped
1	-14.0605	72.6 | 84.6
1	-14.0605	bostonturnstileleftnp:bos | bostonturnstileleftn\n
1	-14.0605	shinsuke | tatsuya
1	-14.0605	segmentrtgramsn | #)()1(#
1	-14.0605	covariance | trick:
1	-14.0605	.64 | .46
1	-14.0605	p(wi|wi−1i−n+1)≈p(wi|φ(wi−1i−n+1))
1	-14.0605	(via | “built-in”
1	-14.0605	a130 | a132
1	-14.0605	 | q1xt1
1	-14.0605	k∑ | rj∑
1	-14.0605	f1s | 0.1935
1	-14.0605	phases: | phase.
1	-14.0605	(dialogue | reading)
1	-14.0605	un | ldc-released
1	-14.0605	universit´e | bd
1	-14.0605	silicon | 6.24
1	-14.0605	|x|
1	-14.0605	np[nb]/n | s/(s\np)
1	-14.0605	airways | chesley
1	-14.0605	couti,t | λi,t
1	-14.0605	lillian, | shivakumar.
1	-14.0605	(cohen | tsarfaty,
1	-14.0605	104 | −50
1	-14.0605	pays | (answer:
1	-14.0605	ˆes←ecur | g(ecur)≥g(e)
1	-14.0605	twitter, | micro-blogging
1	-14.0605	recognizer | step(α)
1	-14.0605	t,τ | t,t
1	-14.0605	ij | )+
1	-14.0605	jazzy | 21.69%
1	-14.0605	’@ | ¸a«
1	-14.0605	haiian | beijing
1	-14.0605	{pr(k)·pr(cn | approxorequal
1	-14.0605	ffe | nda
1	-14.0605	hdp | piq
1	-14.0605	sibling | “pp
1	-14.0605	character | ‘people’
1	-14.0605	buckwalter | (bama)
1	-14.0605	(mcdonald, | carreras,
1	-14.0605	biology, | genetics,
1	-14.0605	screenshot | visualization
1	-14.0605	the | broke.
1	-14.0605	α[x,0,y] | β[x,1,y]
1	-14.0605	(ahn | 2004)
1	-14.0605	pre-test | }.
1	-14.0605	(honesty–white, | joy–yellow,
1	-14.0605	cirulo | ci1(x
1	-14.0605	european | barreiro
1	-14.0605	o(|v|2)
1	-14.0605	categories | lenneberg,
1	-14.0605	kg882>d | acyg3(%
1	-14.0605	∆, | r(e1,e2)
1	-14.0605	lemma | hr-error
1	-14.0605	(nigam | ghani,
1	-14.0605	man” | bitten
1	-14.0605	electrical | ottawa
1	-14.0605	classificationmodelsfor | modelto
1	-14.0605	(farrar | langendoen,
1	-14.0605	news-dev2009a | news-dev2009b
1	-14.0605	<<adj1, | <adj2,
1	-14.0605	1,619 | 13,517
1	-14.0605	hat | pizza”,
1	-14.0605	78% | games
1	-14.0605	22.18 | 55.45
1	-14.0605	dobj> | stat3
1	-14.0605	sentiment” | sex,
1	-14.0605	vnull1(v,y) | vnull2(v,y)
1	-14.0605	gayaa | thaa
1	-14.0605	0.483 | 0.463
1	-14.0605	gnu | (glpk)
1	-14.0605	viktim | vini
1	-14.0605	asif | patna
1	-14.0605	fructose | profundus
1	-14.0605	ˆe←e | ˆes←eprimes
1	-14.0605	room | lamp”
1	-14.0605	sarcastic, | r(16)
1	-14.0605	data | n=|ts|;
1	-14.0605	we’d | martell
1	-14.0605	yusuke | yusuke@nii.ac.jp
1	-14.0605	finies | effac´e
1	-14.0605	(m¨uller | kasper,
1	-14.0605	(tsao | wible,
1	-14.0605	92.9 | 67.3
1	-14.0605	{source, | opinion}
1	-14.0605	top-down | history
1	-14.0605	ila | vach
1	-14.0605	heuristicrules | performedto
1	-14.0605	mt05 | 39.39
1	-14.0605	(rprime | mprimen)
1	-14.0605	jes´us | tecnol´ogico
1	-14.0605	par | french-english
1	-14.0605	pd,l(fs,as|es).
1	-14.0605	that-complementizer | that-mentioning.
1	-14.0605	||m -combination
1	-14.0605	operator | hiero-alike
1	-14.0605	iraq) | age-wise.
1	-14.0605	entrer | vigueur
1	-14.0605	sujith | {sravi,knight}@isi.edu
1	-14.0605	creole) | (english)
1	-14.0605	pprime0 (e|x)
1	-14.0605	antonymous | paraphrases/entailment
1	-14.0605	exm-best | 15.72
1	-14.0605	(hachey | grover,
1	-14.0605	mt1 | mt2
1	-14.0605	d´esilets | hermet
1	-14.0605	fp7 | (classic,
1	-14.0605	party,lose | nullpresident,lose,allynull
1	-14.0605	not | otherbackgroundknowledgesourcesas
1	-14.0605	descriptive | transactional
1	-14.0605	0.26 | onlinea
1	-14.0605	0.291 | 0.408
1	-14.0605	count: | ¯lscript:
1	-14.0605	physical | arg1s
1	-14.0605	“b-1 | warthog”
1	-14.0605	4.01 | 4.09
1	-14.0605	(w/ | non-terminals)
1	-14.0605	restrictions: | [+int
1	-14.0605	retraining | rescoring
1	-14.0605	student: | ack
1	-14.0605	implementing | crf)
1	-14.0605	mouton | gruyter.
1	-14.0605	ncls | 0.838§
1	-14.0605	147.1 | 140.7
1	-14.0605	white | bicycles)
1	-14.0605	ester | patricio
1	-14.0605	fr | nous
1	-14.0605	p−2,−1 | p−1,+1
1	-14.0605	(left) | (right):
1	-14.0605	poll | ntrie
1	-14.0605	hansard | he-en
1	-14.0605	i;h1;kihx2;x1i;hx1;x3i | k;h2;jihx1;x3i;hx3;x4i
1	-14.0605	recent | zukerman,
1	-14.0605	content-container, | message-topic,
1	-14.0605	map | p20
1	-14.0605	10·lm | 10·sumlm
1	-14.0605	bl-wb | dv-wb
1	-14.0605	p(e2|f)
1	-14.0605	(c=1000) | two-sided
1	-14.0605	measures, | wordnet::similarity
1	-14.0605	51.9 | 51.9
1	-14.0605	suburb | london).
1	-14.0605	singular | “dp→
1	-14.0605	compression | p(w|ˆd)
1	-14.0605	succeeded | prince.
1	-14.0605	γ: | γi.
1	-14.0605	mati | etv
1	-14.0605	0.6886 | 0.6215
1	-14.0605	grefenstette | sadrzadeh,
1	-14.0605	g1875ℎg1857g1870g1857 | g34201,
1	-14.0605	way,suite | rey,ca90292
1	-14.0605	s\np, | >ts/(s\np)
1	-14.0605	scott, | caine,
1	-14.0605	(bilous | krauss,
1	-14.0605	chew | eysenbach
1	-14.0605	to) | (nns
1	-14.0605	(ristad | yianilos,
1	-14.0605	fixing | v0v
1	-14.0605	vulgar | ‘conspicuously
1	-14.0605	brett | gelfand
1	-14.0605	research–almaden | 95120,
1	-14.0605	gold1 | gold3
1	-14.0605	piek, | monachini,
1	-14.0605	haizhou | aiti
1	-14.0605	daoud | hertfordshire
1	-14.0605	bilmes | bilmes@ee.washington.edu
1	-14.0605	0.686 | 0.715
1	-14.0605	pn(e|f)·s(e,eprime)
1	-14.0605	2,051 | 2,525
1	-14.0605	arrives | tristan
1	-14.0605	σ | beta(δ0,δ1)
1	-14.0605	25.27 | 33.35
1	-14.0605	αij, | ρkj(1)]+
1	-14.0605	ming | zhongyang
1	-14.0605	+subj | ilp
1	-14.0605	character] | file].
1	-14.0605	uf | (ast-cova)
1	-14.0605	markjohnson | mark.johnson@mq.edu.au
1	-14.0605	commonlyused | whethera
1	-14.0605	paiwan | saisiat
1	-14.0605	conjecture | non-stylistic
1	-14.0605	49.56 | arabic
1	-14.0605	unifor | ppgia
1	-14.0605	saliency | recursive
1	-14.0605	/w/and | bop
1	-14.0605	a0 | $
1	-14.0605	(chapman | 2001),
1	-14.0605	average) | (+0.46
1	-14.0605	motivating | “flights
1	-14.0605	evaluativity | edislike
1	-14.0605	91.50 | 91.24
1	-14.0605	08] | ~15
1	-14.0605	zarrieß | kuhn
1	-14.0605	ojos | blancos
1	-14.0605	qur’an, | 'blue')
1	-14.0605	85.01 | 84.98
1	-14.0605	route | map;
1	-14.0605	p(ti|si))p(s) | later”)
1	-14.0605	institut | uclouvain
1	-14.0605	gonz´alez-rubio | inform´atica
1	-14.0605	farmers | resisted
1	-14.0605	let’s | something
1	-14.0605	correspondences | resource.
1	-14.0605	“please | me
1	-14.0605	0.605 | 0.500
1	-14.0605	guihua | cong,
1	-14.0605	b. | (fl´orez
1	-14.0605	secon | (runtime
1	-14.0605	dcu | “≥others”systemscore
1	-14.0605	hanks, | lazo.
1	-14.0605	istituto | zampolli”
1	-14.0605	hash | buckets.
1	-14.0605	tk(t1,t2) | n2∈nt2
1	-14.0605	vprime≤v≤|f|, bs,tprime,uprime,v,
1	-14.0605	85.50 | 92.76
1	-14.0605	0.825 | 0.876
1	-14.0605	decry | denounce
1	-14.0605	co-targets | “yoga”
1	-14.0605	relates | psychology,
1	-14.0605	deterministically | γ/nsen
1	-14.0605	algebraically, | parenleftbig−→w
1	-14.0605	2-gram | 214.3
1	-14.0605	traf2 | cd40
1	-14.0605	district | ningbo
1	-14.0605	le | sergue¨ı
1	-14.0605	length(b) | compatible(bk,mi)
1	-14.0605	invest: | “year”,
1	-14.0605	briche | ...
1	-14.0605	-述べ | -様
1	-14.0605	t4me | programme.
1	-14.0605	dr | score.
1	-14.0605	gprime1 | gprime2.
1	-14.0605	73.7 | 75.0
1	-14.0605	thementionshave | yif | thefeaturesetdescribingtherelationshipbetweenci
1	-14.0605	ib;k | ib;j
1	-14.0605	md | jjs
1	-14.0605	degradation | watermarking.
1	-14.0605	euro, | giga
1	-14.0605	sentence: | interleukin
1	-14.0605	rescores | noisychannel
1	-14.0605	solver, | solver.
1	-14.0605	subtype | “air”
1	-14.0605	gmds | ∗roi:
1	-14.0605	‘sleep’ | ‘parent’.
1	-14.0605	assigned, | successful,
1	-14.0605	pred(l0, | x3),
1	-14.0605	shortest | v(s)
1	-14.0605	(0/0=0) | 0.649
1	-14.0605	stringmatch | bothpronouns
1	-14.0605	markup | (sgml)
1	-14.0605	(7) | (98)
1	-14.0605	negated, | family-related
1	-14.0605	repeated | acoustically
1	-14.0605	north | march
1	-14.0605	table4: | goldmentionboundaries
1	-14.0605	fellbaum, | elkateb
1	-14.0605	angle | (<>)
1	-14.0605	xiaohua | harbin,
1	-14.0605	monsieur | devra
1	-14.0605	c(w2 | n1+(•w
1	-14.0605	(m) | 8.41
1	-14.0605	jay | ashok
1	-14.0605	forwn∈w(d) | +αz)
1	-14.0605	261.0 | 258.1
1	-14.0605	verde | /bl1(x
1	-14.0605	moses+del | 30k
1	-14.0605	saar | 5.1,
1	-14.0605	mstep(counts); | counts.clear();
1	-14.0605	songs | dances
1	-14.0605	james | doolittle,
1	-14.0605	patterngrammarbank) | (userusage,mineditedsug)
1	-14.0605	tm-scs | xabionlp
1	-14.0605	p(w|z).
1	-14.0605	variational | q(ψ,θp,θb,θa,φ)
1	-14.0605	grano | miel
1	-14.0605	ugly | slums
1	-14.0605	s80 | s48s46s48s54
1	-14.0605	ries, | cocaro,
1	-14.0605	ball | midfield,
1	-14.0605	mantle | lithosphere.
1	-14.0605	p(w|s)
1	-14.0605	‘der | f¨alle,
1	-14.0605	512 | nonterminal
1	-14.0605	vehicle | five-star
1	-14.0605	worth | foodstuff
1	-14.0605	texans | 1901.
1	-14.0605	follo | (b|φ).
1	-14.0605	55.2 | 37.9
1	-14.0605	anc | cri
1	-14.0605	∩er2|=|eprime1∩eprime2|, |er1
1	-14.0605	“for | all/most
1	-14.0605	their≥others | tied
1	-14.0605	7725-cc1 | ’produce’,
1	-14.0605	(gerrish | blei,
1	-14.0605	fellowship, | worldly
1	-14.0605	bılal=ne | bilal.m.sg=erg
1	-14.0605	99.13 | 99.30
1	-14.0605	waiting(true/false), | no),
1	-14.0605	problem), | goaltest(problem,
1	-14.0605	unlabeled | arc-labeler
1	-14.0605	victor | nikos
1	-14.0605	ψ(pprime) | (pprimeprime,i)
1	-14.0605	eat”, | “to”←“eat”.
1	-14.0605	muc | 0.5192
1	-14.0605	parent:pos | parent:case;
1	-14.0605	eror | synonyms.
1	-14.0605	(brown | ratner,
1	-14.0605	(siegel | castellan,
1	-14.0605	offenbar | rakete
1	-14.0605	extended | (kipper,
1	-14.0605	compression | deletion-only
1	-14.0605	#michael | #swine
1	-14.0605	t2: | f2n
1	-14.0605	wiley | sons,
1	-14.0605	cucchiarini, | strik,
1	-14.0605	(goal | springhouse)
1	-14.0605	ws: | (b|φ(x,y))
1	-14.0605	kahn | (mehay
1	-14.0605	dipti | sharma
1	-14.0605	stemming | (gpn),
1	-14.0605	(ud) | (sd)
1	-14.0605	chemin | fer
1	-14.0605	linche | linchen
1	-14.0605	lee, | vaithyanathan,
1	-14.0605	400 | traink
1	-14.0605	null(null)) | nullnullnull[(nullnull[null(null)nullnullnull(null)]
1	-14.0605	(p,o) | a∗←
1	-14.0605	camera | bright
1	-14.0605	maximize | =summationtextv(βv
1	-14.0605	1hrl | malibu,
1	-14.0605	-0.42 | 1.55
1	-14.0605	hoping | attackers.
1	-14.0605	-1.2 | -0.4
1	-14.0605	strapparava | valitutti
1	-14.0605	combining | entity-mention
1	-14.0605	houssini`ere, | 92208
1	-14.0605	black | near-synonym
1	-14.0605	shouldperformfeatureselectiontoavoidsomeuselessfeatures | scores.
1	-14.0605	id=" | ingt
1	-14.0605	chicken | salmon
1	-14.0605	fazer | tomar
1	-14.0605	s49s48 | s109
1	-14.0605	illc-uva | emnlp-wmt
1	-14.0605	0.343 | 0.326
1	-14.0605	alberto | fondazione
1	-14.0605	80k | 40.53
1	-14.0605	nf-negative | pr-positive
1	-14.0605	+wn | +at
1	-14.0605	milestone | game2
1	-14.0605	cros-chain | chains)
1	-14.0605	expression | frozen.
1	-14.0605	s2o1 | s2or
1	-14.0605	decision-related | decision(s)
1	-14.0605	herman | howald
1	-14.0605	q4xt4 | dnx
1	-14.0605	predx | 35.18
1	-14.0605	conll | however:
1	-14.0605	stripping, | 14(3):
1	-14.0605	avg | stdev
1	-14.0605	ruiqiang | keiji
1	-14.0605	russe | lavrov
1	-14.0605	f-val | 6k
1	-14.0605	meta-thesaurus | dictionnaire
1	-14.0605	c-feel-it: | annotator:
1	-14.0605	0.848 | cont.
1	-14.0605	shutdown | haggling
1	-14.0605	incorrect | correctable,
1	-14.0605	coders | instructed
1	-14.0605	lower-case | non-alpha-num
1	-14.0605	vuur | ‘rocket’
1	-14.0605	truth-theoretic | compositional
1	-14.0605	tape | 18,493
1	-14.0605	p(f|e1)
1	-14.0605	100m, | (1b)
1	-14.0605	“human | protein”
1	-14.0605	datelines, | formatting,
1	-14.0605	long-form | e3:
1	-14.0605	balanced | 3-docs
1	-14.0605	[dt|prp$]
1	-14.0605	e. | wallach
1	-14.0605	(relational | system)
1	-14.0605	(2)), | i-indexed
1	-14.0605	dev | rm-full
1	-14.0605	ost | qua
1	-14.0605	consequently | moreover
1	-14.0605	separatesingletonmentions | mentiondetectionstep
1	-14.0605	brought | bear,
1	-14.0605	anticipation | cross-gender
1	-14.0605	46.31 | 49.57
1	-14.0605	veg | veggie
1	-14.0605	m=1λm(φm)·pm(wi|φm)summationtextn m=1λm(φm)
1	-14.0605	(ll), | (fb),
1	-14.0605	:-) | :-(
1	-14.0605	b@ | ›
1	-14.0605	2-17) | o(|ecur|·|σ|·cg),
1	-14.0605	wk | ,wk
1	-14.0605	(forall | n1)(not’
1	-14.0605	(64m | 8.9958
1	-14.0605	ˆσ | chunks(ˆσ,σ∗)
1	-14.0605	maidan | revolution
1	-14.0605	3.74 | 3.55
1	-14.0605	beer, | salad,
1	-14.0605	white | distractors.
1	-14.0605	l∪ψ | t∪pi
1	-14.0605	woz-08 | asr-08-train.
1	-14.0605	premodifier, | formulaic,
1	-14.0605	397 | 1425
1	-14.0605	echo | hears
1	-14.0605	qualified | found,
1	-14.0605	manuscript | europe.
1	-14.0605	29.9 | rm-c
1	-14.0605	notedthateachpronounmustbe | nominalmentions,
1	-14.0605	“nach | eis”
1	-14.0605	image | [2,4]
1	-14.0605	square6 | northwest
1	-14.0605	∝ | [wid∈sj,xid
1	-14.0605	),( | ),,|(lo
1	-14.0605	˜null, | cancombine(˜null,
1	-14.0605	(henceforth | np)
1	-14.0605	wij | zij
1	-14.0605	conceived | conceiver’s
1	-14.0605	v1, | [s↓,s↓]
1	-14.0605	2007a; | imamura
1	-14.0605	dlm0 | dlm2
1	-14.0605	disambiguate | (batter,
1	-14.0605	zhiyuan | xinxiong
1	-14.0605	covers | xinhua
1	-14.0605	hashtags | #happy,
1	-14.0605	searching | wide
1	-14.0605	re-score | (sproat
1	-14.0605	˜p(wi|wi−11 | ρn
1	-14.0605	{vino|||wine} | in1
1	-14.0605	terse | contain.
1	-14.0605	num | ‘five’
1	-14.0605	(international | excellence)
1	-14.0605	department | interior/national
1	-14.0605	castao, | ingria,
1	-14.0605	genie | guesser,
1	-14.0605	claim | wrong.
1	-14.0605	nearest | (automobile
1	-14.0605	clsp | baltimore
1	-14.0605	alie | lour
1	-14.0605	neutral | (npov)
1	-14.0605	spot | ordinary*
1	-14.0605	mihalcea, | zanzotto
1	-14.0605	w∗ | (ftf
1	-14.0605	(white, | pink,
1	-14.0605	kavalan | centralami
1	-14.0605	whnp[i168] | np[→i169]
1	-14.0605	94305, | jurafsky@stanford.edu
1	-14.0605	[pushes | [turns
1	-14.0605	28.1 | t-fr
1	-14.0605	dk(q) | w(r,dk(q),h)
1	-14.0605	national | league
1	-14.0605	90.6% | decisons.
1	-14.0605	(gec) | commercially
1	-14.0605	mcdonald, | attardi
1	-14.0605	organization) | employees,
1	-14.0605	enter-s1.1 | enter-s1.2
1	-14.0605	leeds, | 9jt
1	-14.0605	#swineflu | #music
1	-14.0605	massacre” | “malaysia
1	-14.0605	formerdoes | labelany
1	-14.0605	’this’, | ’,’,
1	-14.0605	extents, | cohen’
1	-14.0605	tuples | 10,672
1	-14.0605	le | l'
1	-14.0605	david | mimno@cs.princeton.edu
1	-14.0605	iphone | cellband”,
1	-14.0605	city | plains
1	-14.0605	solo | ben.
1	-14.0605	‘bilal | nadya.’
1	-14.0605	self-training: | wsj-train
1	-14.0605	(fung, | langlais,
1	-14.0605	cfv=∞,h=∞ | cfv=∞,h=2
1	-14.0605	cuadrado | triangulo
1	-14.0605	kaczynski | 10-count
1	-14.0605	competence | one´s
1	-14.0605	_bom | </titl
1	-14.0605	reality | raye,
1	-14.0605	40k | 38.49
1	-14.0605	∆p | min{l(cn1),l(cn2)}.
1	-14.0605	inference, | form.
1	-14.0605	0.187 | 0.197
1	-14.0605	going | bronchoalveolar
1	-14.0605	nv | elhuyar
1	-14.0605	(graeser, | whiten,
1	-14.0605	reiter, | horacek,
1	-14.0605	r1, | inverses
1	-14.0605	“this | outfit
1	-14.0605	pearson’s | 0.102.
1	-14.0605	ship | guns,
1	-14.0605	 | r^�om�{
1	-14.0605	(task | operator:
1	-14.0605	places | organisations.
1	-14.0605	forbidding | nulliu,
1	-14.0605	(quirk | 1974)
1	-14.0605	has | 680k
1	-14.0605	(evens | michael,
1	-14.0605	53 | .06)
1	-14.0605	543 | 467
1	-14.0605	nus-tesla-f | sheffield-rose
1	-14.0605	phd | lisbon.
1	-14.0605	fully-aligned | publicly
1	-14.0605	knn, | suggesting
1	-14.0605	deixar | (put
1	-14.0605	(susan | vinokourov
1	-14.0605	(wilmington) | 15,
1	-14.0605	„for | of‟,
1	-14.0605	form: | <j
1	-14.0605	{np,}* {and|or}
1	-14.0605	adequately | iwia
1	-14.0605	spread | unverified
1	-14.0605	central | subset5.
1	-14.0605	35.04 | 32.28(+.84)
1	-14.0605	c∈c p(c|z)
1	-14.0605	x,y | sprime,tprime
1	-14.0605	“l” | “l”,
1	-14.0605	der | mutter
1	-14.0605	14.8% | 31.1%
1	-14.0605	truth | t(ψ)
1	-14.0605	(pyramid), | (rouge,
1	-14.0605	( | ,where
1	-14.0605	(i) | 86.10
1	-14.0605	re1(x | azul
1	-14.0605	(university | tokyo)
1	-14.0605	))(# ||exp(
1	-14.0605	chaining | likelihood.
1	-14.0605	stylistic | dialect.
1	-14.0605	sumbody, | pleeeaas,
1	-14.0605	catib | ¨@
1	-14.0605	jeanne | brett,
1	-14.0605	(meystre | haug,
1	-14.0605	85.5% | 59,979
1	-14.0605	xiaofeng yang, | chew
1	-14.0605	f(%) | cov(%)
1	-14.0605	gamma(α,β) | mog
1	-14.0605	0.661 | 0.701
1	-14.0605	+2.8 | ’6
1	-14.0605	15.1 | 22.4
1	-14.0605	japan | saitama,
1	-14.0605	),()1( | ×−−+
1	-14.0605	advb | persecution
1	-14.0605	node:en:afun | parent:number
1	-14.0605	policy: | ->
1	-14.0605	patents | abdominal
1	-14.0605	12x | 15x
1	-14.0605	circlemultiplydisplay {j|1≤j≤n,jnegationslash=i}
1	-14.0605	adjoining | (schabes,
1	-14.0605	t/punct | t/dur
1	-14.0605	a=appos, | d=dnp,
1	-14.0605	para2 | para1.
1	-14.0605	follower” | doors).
1	-14.0605	exhaustive | (bar-hillel
1	-14.0605	(sh) | m.s.,
1	-14.0605	ρ(wi|wi−1i−n+1) | γ(wi−1i−n+1)·˜p(wi|wi−1i−n+2)
1	-14.0605	null←initchart() | nullnull←binary(null,
1	-14.0605	nn | assignment.
1	-14.0605	heuristicrules | generatinghighlyprecisepre-clusters.for
1	-14.0605	relative, | 2.0%
1	-14.0605	experimentalresults | datasetforconll-2011sharedtaskintable3,table
1	-14.0605	56.9 | 50.0
1	-14.0605	0.376 | 4.20%
1	-14.0605	upspeak | downspeak,
1	-14.0605	consolidated | chairman
1	-14.0605	25.75 | 19.77
1	-14.0605	nullnull[null(null)nullwalk’(null)]null | nullnull[nullnull(null)nullwalk’(null)])nullnullnull[nullnull(null)nullnull(null)]
1	-14.0605	(noeman | madkour,
1	-14.0605	 | 0*1
1	-14.0605	:= | output:optimal
1	-14.0605	hyperedges | hyperedges,
1	-14.0605	∀ | l2
1	-14.0605	attached | relation%1:03:00
1	-14.0605	“싸고 | (cheaper
1	-14.0605	macbride | micciulla.
1	-14.0605	‘etwas, | sieht’
1	-14.0605	reg+ | noreg+,
1	-14.0605	mt | ·vectorxt,gt)−(vectorwt−1
1	-14.0605	smart | tall.
1	-14.0605	inal | xiste
1	-14.0605	(eisenstein | xing,
1	-14.0605	jucse-1 | jucse-3
1	-14.0605	sebastian | computerlinguistik
1	-14.0605	有 | 少数
1	-14.0605	 | .,
1	-14.0605	(x,y) | l(ˆy1,y)
1	-14.0605	(zettlemoyer | 5.3.
1	-14.0605	‘splice’, | ‘freq.’
1	-14.0605	97.1 | 98.3
1	-14.0605	inter-human-agreement | image-centered
1	-14.0605	augmentative | comunication,
1	-14.0605	“how | key”,
1	-14.0605	(: | (,
1	-14.0605	mte | olia
1	-14.0605	(6 | 4p
1	-14.0605	sheffield | “personal”
1	-14.0605	clc-hdp-lg | 1000c-40l,
1	-14.0605	nullnull1nullnull | nullnull2nullnull
1	-14.0605	rk(ρ,p) | l(p)
1	-14.0605	prince | rahman
1	-14.0605	in-vocabulary | oovs.
1	-14.0605	(rapp, | yee,
1	-14.0605	#human-references | per-system
1	-14.0605	[slot1,slot2], | policy:
1	-14.0605	87.75 | decr
1	-14.0605	chinese | cqzong}@nlpr.ia.ac.cn
1	-14.0605	78.1 | 77.7
1	-14.0605	null)]])nullnullnull[nullnull1(null)nullnull1(null)] | null)]])nullnullnull[nullnull2(null)nullnull2(null)]
1	-14.0605	player’s | 2.3
1	-14.0605	positional | mos-beg
1	-14.0605	mid | hut
1	-14.0605	rammed | oncoming
1	-14.0605	dissimilation | (suzuki,
1	-14.0605	¨uber | davon
1	-14.0605	(unused) | '02
1	-14.0605	60.5 | 76.9
1	-14.0605	0.932 | 0.926
1	-14.0605	‘vaalaa’ | morpheme
1	-14.0605	diab | (tufis¸
1	-14.0605	fish | chips,
1	-14.0605	lucia | stafford
1	-14.0605	paper | (prating)
1	-14.0605	alexander | temple
1	-14.0605	s(e,eprime) | lmax−l(e,eprime).
1	-14.0605	−1.0 | −0.50
1	-14.0605	bennet | dundee)
1	-14.0605	productive | (scarborough,
1	-14.0605	ec | “consolider
1	-14.0605	cat | mammal
1	-14.0605	θt | θx.
1	-14.0605	“red | “boston
1	-14.0605	brands | 90.33%
1	-14.0605	snippet | filtering.
1	-14.0605	234 | ]'
1	-14.0605	learninga | ambiguousdata
1	-14.0605	“bank | australia”
1	-14.0605	mello’s | yugoslavia.
1	-14.0605	pos(h) | pos(a)
1	-14.0605	store | size1
1	-14.0605	clusters: | ({june}).
1	-14.0605	automaton | 〈q,σ,δ,q0,ω,λ〉
1	-14.0605	s2: | electricity.
1	-14.0605	indicate | (∗)
1	-14.0605	three/d | and/s
1	-14.0605	my | 0.0140
1	-14.0605	rule, | (lhs)
1	-14.0605	(eu | simbad
1	-14.0605	z(c) | dprime∈ω(c)
1	-14.0605	appeared, | co¨occurred
1	-14.0605	serial | sv-lvcs:
1	-14.0605	dupont-rosenfeld | mechanism
1	-14.0605	socks | lap.
1	-14.0605	chung-chi | shih-ting
1	-14.0605	54.0 | 25.8
1	-14.0605	transliteration: | alahly
1	-14.0605	a1, | predicate1
1	-14.0605	motor | dritten
1	-14.0605	blac | infor
1	-14.0605	tuning, | 03,
1	-14.0605	cand⊕bk | ={}then
1	-14.0605	clc-base, | 100c40l,
1	-14.0605	osition | linea
1	-14.0605	adj | n↓
1	-14.0605	(german), | (hindi).
1	-14.0605	mean-squared | (mse)
1	-14.0605	(q1) | (q2)
1	-14.0605	caused | turmoil
1	-14.0605	“x1:世界杯” | “在”
1	-14.0605	cprime | cons(c,cprime)
1	-14.0605	σ | nulla,b,cnull
1	-14.0605	b+l | 3-gram
1	-14.0605	metric | asymmetric.
1	-14.0605	{productindi} | {dealerindi}
1	-14.0605	0.02, | 0.06.
1	-14.0605	explains | exciting
1	-14.0605	roark | hollingshead’s
1	-14.0605	modified | controlled
1	-14.0605	subasic | huettner.
1	-14.0605	yoav | gurion
1	-14.0605	todiamondmath | thediamondmath
1	-14.0605	8831 | 28413
1	-14.0605	order-1 | 84.35
1	-14.0605	nullnullnullnullnull(nullnullnull1nullnull2) | nullnullnull1[(nullnull[null1(null)nullnullnull1(null)]
1	-14.0605	wo | “beg
1	-14.0605	intra-cluster | inter-cluster
1	-14.0605	‘descend | fall
1	-14.0605	(inesc-id | piddac
1	-14.0605	comparatively | zhuoyue
1	-14.0605	+g+cs | +np
1	-14.0605	withgoldmentionboundariesusingunifiedclassification | 53.73%
1	-14.0605	pisa, | alessandro.lenci@ling.unipi.it
1	-14.0605	blossy | retuned
1	-14.0605	counterpart | ‘phrase
1	-14.0605	96.00% | 94.47%
1	-14.0605	nf | 85.65
1	-14.0605	high-performance | solver.
1	-14.0605	[i | eat
1	-14.0605	y-bo-da | y.)
1	-14.0605	add | shared:
1	-14.0605	nurses | bt-nurse
1	-14.0605	water | dripping
1	-14.0605	dog | tail,
1	-14.0605	energy | electrons)
1	-14.0605	require: | w′syn
1	-14.0605	snake | deadly
1	-14.0605	woche | landwirte
1	-14.0605	obgleich), | gleich).
1	-14.0605	hter | 0.3423
1	-14.0605	moun | blese
1	-14.0605	"怒られる | ra
1	-14.0605	contaminated | bachman
1	-14.0605	islam: | doers
1	-14.0605	gprime1 | gprime2
1	-14.0605	94.5% | 92.6%
1	-14.0605	ward | tsukahara
1	-14.0605	11% | re-scoring.
1	-14.0605	mt06 | 05,
1	-14.0605	62.53 | 63.45
1	-14.0605	train | europarl-v4
1	-14.0605	novelty | prerequisites
1	-14.0605	video | subtitles)
1	-14.0605	hp | n+hn
1	-14.0605	(vp | already)
1	-14.0605	v2 | arg1.
1	-14.0605	84.6% | 83.6%
1	-14.0605	<eventpast> | <eventpastp>
1	-14.0605	n1 | n1’
1	-14.0605	stemming | stop-word
1	-14.0605	pyyheliinat | kaapissa.
1	-14.0605	gapped | pt(oxw|sy
1	-14.0605	sub | circledot
1	-14.0605	(a) | ?c
1	-14.0605	nullnullnullnullnull | (null))).
1	-14.0605	precipitation | precipitation%1:23:00
1	-14.0605	agence | presse
1	-14.0605	s3h | ϕ(x)
1	-14.0605	to | vpz,y
1	-14.0605	"wuensch" | lastname:
1	-14.0605	grctlsa | lctlsa
1	-14.0605	s49s48s48s48s48 | s86
1	-14.0605	20, | 70,
1	-14.0605	classificationbased | coreferenceresolutionfeaturesto
1	-14.0605	1p | faust+2p
1	-14.0605	sclg882mi | jstg882ig
1	-14.0605	s∗= | s:|s|≤l
1	-14.0605	satta | padua,
1	-14.0605	gin | armagnac
1	-14.0605	per | mij,
1	-14.0605	angola, | portugal,
1	-14.0605	x | ��t
1	-14.0605	questionnaire | (revu-nl),
1	-14.0605	transform | y(q)j
1	-14.0605	colleagues | gth-upm.
1	-14.0605	malaria | mosquitoes
1	-14.0605	pocketsphinx | (huggins-daines
1	-14.0605	cp(vp), | ∀vpz,y.
1	-14.0605	lieberman | geotagged
1	-14.0605	peptide | crystal
1	-14.0605	1790, | died.
1	-14.0605	“acts | costimulatory
1	-14.0605	walks | ramage,
1	-14.0605	bigrams, | fourgrams
1	-14.0605	crowdflower, | tasks
1	-14.0605	implicit | explicit.
1	-14.0605	20.91 | 35.1%
1	-14.0605	walker | amsler,
1	-14.0605	(whittaker | stenton,
1	-14.0605	d,l | pd,l(fs,as|es)
1	-14.0605	ron | angela
1	-14.0605	p(w | p(zd
1	-14.0605	potuit | “una
1	-14.0605	sentencetightening | tightening
1	-14.0605	ros´e | 15213
1	-14.0605	terminology | easily:
1	-14.0605	69.5 | 65.8
1	-14.0605	device | 12-month
1	-14.0605	2081 | (37%)
1	-14.0605	“al” | “at”
1	-14.0605	imdb62 | 62,000
1	-14.0605	[other | prt,
1	-14.0605	总理 | 会
1	-14.0605	61.6% | lda2
1	-14.0605	h(d|w | w,k)
1	-14.0605	3.7g | 7.5g
1	-14.0605	1971; | 1994)
1	-14.0605	recognition), | (machine
1	-14.0605	ak | left-hand
1	-14.0605	accuracy(%) | (2,000
1	-14.0605	(shannon | weaver,
1	-14.0605	neuen | neuem
1	-14.0605	simon | elaine
1	-14.0605	lengthpenalty | −80.8,complval
1	-14.0605	wnf | p(wn|wn−11
1	-14.0605	people’s | (taiwan,
1	-14.0605	covariance | ψ(x,y)m
1	-14.0605	ℓ(qi)∑ | −z
1	-14.0605	jst | [ye10]
1	-14.0605	(5>2>1 | 5>4>3)
1	-14.0605	(shi | mihalcea,
1	-14.0605	(ms), | (mp),
1	-14.0605	<param2> | <param3>
1	-14.0605	(sg) | respondable
1	-14.0605	;3%"44)4822) | '(*+)
1	-14.0605	icia091ro | uaic092ro
1	-14.0605	bren ies, | ,.all
1	-14.0605	die | innere
1	-14.0605	r` | negau
1	-14.0605	french/english | through
1	-14.0605	t→h | tnotarrowrighth
1	-14.0605	74.0 | 79.0
1	-14.0605	i∈s | sim2(si,sk)
1	-14.0605	vino | hasta
1	-14.0605	∗school | ‡instituto
1	-14.0605	tapanainen | järvinen’s
1	-14.0605	none | learnthigh
1	-14.0605	empirically-derived | empirically-derived
1	-14.0605	explanations | spots
1	-14.0605	crafts | vessels
1	-14.0605	manshadi, | swift,
1	-14.0605	(mcnamee, | (pablo-sanchez
1	-14.0605	hitotsubashi, | 101-8430,
1	-14.0605	aural | immersion
1	-14.0605	formalisation | applicable
1	-14.0605	rk | rk−1,
1	-14.0605	0.2573 | 0.2243
1	-14.0605	sorbonne | umr-i
1	-14.0605	θi|λ−1 | (c)|
1	-14.0605	0.7251 | 0.2105
1	-14.0605	anotherwellknown | whichhasonlyonereferenceinthetext.
1	-14.0605	exposed | h−m,···
1	-14.0605	meta-post | word-based
1	-14.0605	kasper | krieger
1	-14.0605	proposition | “coaches
1	-14.0605	“bag | codewords”
1	-14.0605	miller, | zamanian,
1	-14.0605	nzi|di | zprime(nwi|zprime+β)
1	-14.0605	jets” | killer”
1	-14.0605	0.85714/ | enlv
1	-14.0605	stsg | 0.086
1	-14.0605	1/λmax, | λmax
1	-14.0605	site(entity) | site(entity)+
1	-14.0605	valley | affluent
1	-14.0605	p(zj|w1)
1	-14.0605	mcfgs | strongly
1	-14.0605	≤|j|≤
1	-14.0605	200,000 | 50,000100,000declining
1	-14.0605	charged | with”
1	-14.0605	nlp=natural | snowball=snowball
1	-14.0605	retraining | (once
1	-14.0605	“trivial” | 2-role
1	-14.0605	assistant | a-propos
1	-14.0605	(brown | levinson,
1	-14.0605	reserver | patient[1]
1	-14.0605	tin2009-14675-c03) | (faust
1	-14.0605	nuschime | uc3m
1	-14.0605	pt(zi | 1|t
1	-14.0605	estar | ficar
1	-14.0605	(sandor | (garcia,
1	-14.0605	54.56 | 55.07
1	-14.0605	airline | aircraft
1	-14.0605	a∈{aprime⊆v :∀j∈f,|aprime∩pej
1	-14.0605	logp(d|q), | p(d|q)
1	-14.0605	l1crf | 04)
1	-14.0605	patgram | cost<mineditedcost
1	-14.0605	houda | limsi-cnrs
1	-14.0605	universit¨at | http://www.ukp.tu-darmstadt.de
1	-14.0605	add->album(‘mtv | filled
1	-14.0605	x) | y),(vp
1	-14.0605	parents | makeover
1	-14.0605	balat | kurdish-armenian
1	-14.0605	en | fr,
1	-14.0605	c1i | c2j
1	-14.0605	whittaker | raj
1	-14.0605	indian | telugu,
1	-14.0605	greet | (“hi
1	-14.0605	everyone | suffocating
1	-14.0605	exp | inapra
1	-14.0605	written | treasure
1	-14.0605	non-linear | equations,
1	-14.0605	(1 | l),
1	-14.0605	negev | be’er
1	-14.0605	biomedical | (wattarujeekrit
1	-14.0605	o(n3|g|)
1	-14.0605	takuya | †department
1	-14.0605	(frey | dueck,
1	-14.0605	kproductdisplay | pj(ak|ak−1)·
1	-14.0605	govern | (strapparava
1	-14.0605	text::nsp | phi
1	-14.0605	ford∈ddo | samplezi∝(nz|di
1	-14.0605	macaon | interfaced
1	-14.0605	λ-hybrid | λ-productions,
1	-14.0605	concordant | pairstotal
1	-14.0605	triples, | (bi,uj,uk),
1	-14.0605	0.3-point | (69%
1	-14.0605	0.3680 | 0.4168
1	-14.0605	black, | sungbok
1	-14.0605	β≡b | /∈s,x
1	-14.0605	p(¯a|¯q) | p(¯q|¯a),
1	-14.0605	90.9 | 57.1
1	-14.0605	elizabeth | lester
1	-14.0605	[cognizer | him]
1	-14.0605	attracting | investments,
1	-14.0605	rt(wi) | j:wj→wi
1	-14.0605	combined | 0.6651
1	-14.0605	taskease: | mmpresent.:
1	-14.0605	mt02 | bin-2
1	-14.0605	sight | suprising
1	-14.0605	(using | ruleset)
1	-14.0605	zheng-jun | meng
1	-14.0605	massimo | vergata”
1	-14.0605	vladimir | †umiacs
1	-14.0605	climate | calories
1	-14.0605	amount | 76.44
1	-14.0605	hundreds | athlete,
1	-14.0605	(de), | (it),
1	-14.0605	instantiating | [talk]z
1	-14.0605	nullnullnull[nullnull(nullnullnull)nullate’(null, | nullnullnull2[(nullnull[null2(null)nullnullnull2(null)]
1	-14.0605	caused, | followed,
1	-14.0605	p(e2,f|e1) | p(e2|f,e1)p(f|e1)
1	-14.0605	order | 214.6
1	-14.0605	biotechnology | (ncbi)
1	-14.0605	[null], | nullnullnullpi(1)nullnullnullnullnullnullpi(i).
1	-14.0605	y | colocar
1	-14.0605	gb | raid
1	-14.0605	inu | inl.
1	-14.0605	administrator | cathedral
1	-14.0605	edwards, | fitch,
1	-14.0605	bloggers | 30’s
1	-14.0605	featuresdescribingci | datafilepublishedbybergsma
1	-14.0605	conceptnet | singh,
1	-14.0605	turn | left.
1	-14.0605	upward | (slope
1	-14.0605	(bojar | tamchyna,
1	-14.0605	nullnull1null2[ | null1(null))
1	-14.0605	modifying | h-shaped
1	-14.0605	θdq | p(w|θdq)
1	-14.0605	plurals, | (msms)
1	-14.0605	residue | methylated
1	-14.0605	peace | 7330
1	-14.0605	0.10.2 | 0.350.4
1	-14.0605	0.93 | (lemma,
1	-14.0605	مـ | (ma,
1	-14.0605	ringlstetter, | žorga,
1	-14.0605	nullnull | nullnullnullnull.
1	-14.0605	({may}) | boprimeprime
1	-14.0605	acquired | -*2
1	-14.0605	mário | oliveira.
1	-14.0605	conditioning | equation
1	-14.0605	p(zq|z∗(i)q
1	-14.0605	gn(eprime) | ng∈nk(eprime)
1	-14.0605	marchetti, | monica,
1	-14.0605	patient | bi¯albuh˘
1	-14.0605	structured, | st)
1	-14.0605	dictionary | 43,021
1	-14.0605	understudy | gisting
1	-14.0605	letzte | wollten
1	-14.0605	w2 | meat
1	-14.0605	(insert, | replace)
1	-14.0605	0.545 | 0.399
1	-14.0605	“<” | “>”
1	-14.0605	domain(ri) | range(rj)
1	-14.0605	choosen | correlation
1	-14.0605	alpage, | inria,
1	-14.0605	r-1 | #sent
1	-14.0605	〈e1,f1〉 | 〈e2,f2〉
1	-14.0605	plan | hogar
1	-14.0605	s⊆v | i∈s
1	-14.0605	principles | (volume
1	-14.0605	government | endorsement
1	-14.0605	itself | ishizuka,
1	-14.0605	vpcomb | vplex
1	-14.0605	</timex3> | eventinstanceid=''e1''
1	-14.0605	plane | air
1	-14.0605	≡ | ⋗
1	-14.0605	superposition | l1’s
1	-14.0605	01234 | ✉
1	-14.0605	lagrangian | parenleftbigp
1	-14.0605	(i,t)+ | ai,j,r=1
1	-14.0605	(improving | coherence),
1	-14.0605	dlm1 | dlm3
1	-14.0605	semcor | monowords
1	-14.0605	gs | 0.1193
1	-14.0605	admissible | exact)
1	-14.0605	feelings, | wiebe,
1	-14.0605	[i,j]h | unsealed
1	-14.0605	classificationbasedpre-clusterpair | differentmodelsto
1	-14.0605	neighborhood | n(s,t)
1	-14.0605	period, | semicolon,
1	-14.0605	1443 | 2399
1	-14.0605	glyph31glyph50glyph38glyph38glyph48glyph42glyph45glyph40glyph1glyph42glyph45glyph1glyph32glyph6glyph9glyph5glyph9glyph33 | glyph28glyph51glyph50glyph47glyph51glyph50glyph15
1	-14.0605	everything | v:eats
1	-14.0605	afosr | cmu.
1	-14.0605	(brody | elhadad
1	-14.0605	red | p17�s
1	-14.0605	paris-sud | limsi-cnrs
1	-14.0605	participatedin | geniaevent
1	-14.0605	dominant | submissive
1	-14.0605	en-cs | 0.178
1	-14.0605	propositions: | steve_young
1	-14.0605	cashier | filmmaker
1	-14.0605	null5 | null7
1	-14.0605	(berlin | kay,
1	-14.0605	bath | famyly
1	-14.0605	worst | 69.03
1	-14.0605	constitute | ce)
1	-14.0605	অংnullক | _বnull
1	-14.0605	typed | parses.
1	-14.0605	liwc+bigrams+svm | liwcsvm.
1	-14.0605	predicates | mohanan,
1	-14.0605	-0.40 | -0.24
1	-14.0605	marc | roser
1	-14.0605	p(syn|s) | p(syn|n)
1	-14.0605	dmoz | non-regional
1	-14.0605	relative, | damit
1	-14.0605	acc | negativep
1	-14.0605	felice | “antonio
1	-14.0605	technology, | moe,
1	-14.0605	intra-class | s3hf
1	-14.0605	ca | <lastname>@ict.usc.edu
1	-14.0605	wolverhampton | wolverhampton,
1	-14.0605	β1 | β2,
1	-14.0605	orig | etel
1	-14.0605	(functional, | separable)
1	-14.0605	weer | ‘poll’
1	-14.0605	“kehrten” | “returned”
1	-14.0605	collection|series
1	-14.0605	moreover | tuzhilin,
1	-14.0605	crestan, | borkovsky,
1	-14.0605	piece | 56.81
1	-14.0605	bost, | 0.859
1	-14.0605	strict | n=5,m=50
1	-14.0605	fest | dtl
1	-14.0605	0.06 | +number
1	-14.0605	wird | sanktionen
1	-14.0605	“company” | “unit”,
1	-14.0605	endorses | refutes
1	-14.0605	(jain | pantel,
1	-14.0605	reach | sprimel,
1	-14.0605	0.628 | 0.794
1	-14.0605	human-targeted | (hter)
1	-14.0605	taggers | “african
1	-14.0605	strict-unsup | 43%
1	-14.0605	1990; | 1993;
1	-14.0605	heeman, | filimonov
1	-14.0605	atxxta | bty
1	-14.0605	(dagan | itai,
1	-14.0605	··· | conserves
1	-14.0605	1025 | 1026
1	-14.0605	cornell | cardie@cs.cornell.edu
1	-14.0605	(sense | qu’il
1	-14.0605	bills | wage
1	-14.0605	cure | claim
1	-14.0605	(vsm) | 1975).
1	-14.0605	down | annotator.
1	-14.0605	duluth-2 | submission-pmi
1	-14.0605	peiling | ‘represent’
1	-14.0605	usion | intr
1	-14.0605	book(x3):20 | book(x2):20
1	-14.0605	postpositions), | (particles),
1	-14.0605	l(w) | u∑
1	-14.0605	bestbind(c) | cbindi,p,q
1	-14.0605	t-nl | t-fr
1	-14.0605	3-5 | keihanna
1	-14.0605	[the | bonds]
1	-14.0605	hm1_wc, | hm12_wc
1	-14.0605	luretech | hooks
1	-14.0605	study | (mohammad,
1	-14.0605	neighbours | n=20).
1	-14.0605	[arg3 | pieces].
1	-14.0605	[2,4] | [3,5].
1	-14.0605	hits-at | page-af
1	-14.0605	dependency | 〈wp,wa,r〉,
1	-14.0605	20000 | 60000
1	-14.0605	n,d | recursions
1	-14.0605	eg | j.
1	-14.0605	kulture | digit
1	-14.0605	] | [by
1	-14.0605	summarization | greenbacker
1	-14.0605	(brockett | dolan,
1	-14.0605	(buyko | bergler,
1	-14.0605	nieuwenhuis | anwv
1	-14.0605	crop* | corn*
1	-14.0605	“mac” | “windows”
1	-14.0605	inversely, | (strongly)
1	-14.0605	vocal | vibration
1	-14.0605	s12, | s18.
1	-14.0605	(beck | sauerland,
1	-14.0605	tesla-m | 31.0%
1	-14.0605	dang, | trang,
1	-14.0605	saarland | d-66123,
1	-14.0605	w1w2… | (2≤n≤18)
1	-14.0605	34.9 | 36.8
1	-14.0605	galley | redmond,
1	-14.0605	record | bhattacharya
1	-14.0605	(tonelli | delmonte,
1	-14.0605	jucse-2 | scss-tcd:conf3
1	-14.0605	parenleftbigg | parenrightbigg2
1	-14.0605	jerusalem | protest
1	-14.0605	p7 | p8
1	-14.0605	exp5 | all:
1	-14.0605	san | 95125
1	-14.0605	abc | washington
1	-14.0605	university | 142,
1	-14.0605	i∈k | k−1summationdisplay
1	-14.0605	(c0) | (c1)
1	-14.0605	naval | muri
1	-14.0605	22.18% | 73.26%.
1	-14.0605	α[x,1,y] | β[n,1,y]
1	-14.0605	p(zq|r)p(r|q).
1	-14.0605	(left) | autonomously
1	-14.0605	∀b∈σ | p(ra2|b3,b2,b1)
1	-14.0605	yn | ypi(n)
1	-14.0605	lower | tagging).
1	-14.0605	rome, | amateur
1	-14.0605	34.41 | 39.27
1	-14.0605	danger | softness
1	-14.0605	prob(n2 |
1	-14.0605	estela | borja
1	-14.0605	(matusov, | (sim
1	-14.0605	s53 | s53s48s37
1	-14.0605	subsequences | (of
1	-14.0605	t,·) +β|v|
1	-14.0605	afii62833afii62808 | afii62782
1	-14.0605	“regulation” | regulation”
1	-14.0605	p1s7: | "joanne"
1	-14.0605	3.62 | 4.07
1	-14.0605	0.264 | 0.477
1	-14.0605	(whatever | e.g.,
1	-14.0605	fish | near-synonym
1	-14.0605	relativizers | who...
1	-14.0605	vafin | vvpp
1	-14.0605	src | afii62825afii62832afii62820afii62782
1	-14.0605	polymerasecomplex | proteinfamily
1	-14.0605	(buller | burgoon,
1	-14.0605	cafarella, | broadhead,
1	-14.0605	covariance | magnitude
1	-14.0605	mangoville | odeee
1	-14.0605	k-t-b | related’
1	-14.0605	els | hoste1,2,3
1	-14.0605	logic1 | roots.
1	-14.0605	(reisinger | pa¸sca,
1	-14.0605	[party, | party]
1	-14.0605	(as), | (cu),
1	-14.0605	science, | arts,
1	-14.0605	mimno | nj
1	-14.0605	marks) | catch-all
1	-14.0605	0.357 | 2.74%
1	-14.0605	p(a | s,t−1
1	-14.0605	p(xi,j|yi,j;θ).
1	-14.0605	(next | publication
1	-14.0605	lexical+prod-rule[both]*(n=all) | lexical+reranking(n=all)
1	-14.0605	∧∃y[y | ∧∀z[z·producttexty
1	-14.0605	archival | (ead)
1	-14.0605	score(q,d) | t∈d
1	-14.0605	1,ϕ(b | )
1	-14.0605	jwj | rules:
1	-14.0605	σ∗ | p∗prefix
1	-14.0605	0.6144 | 0.3664
1	-14.0605	avsnitt | bok
1	-14.0605	4,9 | 2,phos.
1	-14.0605	d=1 | n=1
1	-14.0605	combinatorial | (valences)
1	-14.0605	z) | conclusion.
1	-14.0605	ja | -te
1	-14.0605	question | motion.
1	-14.0605	ceafm | 0.5843
1	-14.0605	gaussian | (grf)
1	-14.0605	smoothness | (large
1	-14.0605	concept-culture-wise | “blue”
1	-14.0605	double-dop | p=0.
1	-14.0605	partialparsingalgorithm | initcandidatearcs(w0
1	-14.0605	intensive | (portet
1	-14.0605	train | 1.7m
1	-14.0605	0.65 | min.
1	-14.0605	cc | mkm
1	-14.0605	parenleftbiggmprime | ce(ng)·pn(e|f)
1	-14.0605	categorial | (ccg),
1	-14.0605	wenn | m¨ussen
1	-14.0605	(df) | buttons
1	-14.0605	(smolensky, | (saul
1	-14.0605	“chrysler: | $800
1	-14.0605	shudnt | togtherrr
1	-14.0605	french | (sagot
1	-14.0605	s1lwc, | s0ws1c,
1	-14.0605	engineering, | {balamurali,adityaj,pb}@cse.iitb.ac.in
1	-14.0605	nguyen, | tsuji.
1	-14.0605	mosteller | wallace
1	-14.0605	derivation, | +summationtexte
1	-14.0605	tokenization, | stripping,
1	-14.0605	vino | aceite
1	-14.0605	u10a | u2r]
1	-14.0605	〈arg〉(d∧ | 〈prop〉(p∧
1	-14.0605	suffixes | h@+
1	-14.0605	fscore | 41.66
1	-14.0605	{(q,f,q | )},
1	-14.0605	f-91 | orsay
1	-14.0605	amitava | amitava.santu@gmail.com
1	-14.0605	ettm | hybhsum.
1	-14.0605	“dog” | “focal”
1	-14.0605	classes | morphology]
1	-14.0605	bd | poxi
1	-14.0605	sms | 17,192
1	-14.0605	γ2] | [δ1,
1	-14.0605	transducers | foma
1	-14.0605	politecnico | roma,
1	-14.0605	en-es | 0.2591
1	-14.0605	unobtrusive | superposed
1	-14.0605	advice | bowling
1	-14.0605	中国 | 发展
1	-14.0605	furniture | manipulable
1	-14.0605	maxent | prod-rule[stanford](n=1000)
1	-14.0605	normalisation | concept-dependent
1	-14.0605	höglund, | t.,
1	-14.0605	mgs | stabler
1	-14.0605	f) | tm(e|
1	-14.0605	matteo | speranza,
1	-14.0605	“all” | “athletes”
1	-14.0605	4.00% | 10.00%
1	-14.0605	pmi-yahooboss | inv.edge-counts
1	-14.0605	s55s52 | s56s48
1	-14.0605	bank, | river,
1	-14.0605	men’s | 3,499
1	-14.0605	p(e|f), | 2004):
1	-14.0605	(23) | vali.
1	-14.0605	θ | 0≤i≤j≤|w|pr
1	-14.0605	duty, | responsibility,
1	-14.0605	“moderate” | “chair”
1	-14.0605	16.79 | 18.31
1	-14.0605	rithm | rmal
1	-14.0605	[the | 3,000
1	-14.0605	jj|ga16|n | nn|di10
1	-14.0605	order) | ł.
1	-14.0605	machine-generated | promising
1	-14.0605	macbook | nwz
1	-14.0605	room] | right]
1	-14.0605	21.4 | 22.2
1	-14.0605	(vp | (aux
1	-14.0605	univ | sorbonne
1	-14.0605	estimator | estimator
1	-14.0605	varna | varna,
1	-14.0605	ii) | donate
1	-14.0605	ba | ($k)
1	-14.0605	test10 | hybrid-2010
1	-14.0605	6.3 | p(gt|αt)
1	-14.0605	phrasings | 12-month
1	-14.0605	(harel | koren,
1	-14.0605	newstest2009 | eparl+nc
1	-14.0605	reward | $0.10
1	-14.0605	words | 6004
1	-14.0605	(rieser | janarthanam
1	-14.0605	1|b1,...,bn,n)
1	-14.0605	warne | deadly
1	-14.0605	smaller) | employing
1	-14.0605	differentlinktypestendto | coreferentialor
1	-14.0605	e∈e | eprime
1	-14.0605	dozens | youths
1	-14.0605	vilar, | name.surname@dfki.de
1	-14.0605	〈2,3,4〉, | 〈6〉,
1	-14.0605	78.9 | 85.2
1	-14.0605	lamb | duck
1	-14.0605	"mdt" | devices)
1	-14.0605	592 | {apbelis,myazdani,ananchen,pgarner}@idiap.ch
1	-14.0605	alfredo | emms
1	-14.0605	299 | t/causal
1	-14.0605	(hence | distribution).
1	-14.0605	wil | gregory
1	-14.0605	1929-1930 | today*
1	-14.0605	(in | [after
1	-14.0605	prem | yashwant
1	-14.0605	chakraborty, | tapabrata
1	-14.0605	sites, | sites),
1	-14.0605	carvalho, | sarmento,
1	-14.0605	pwrel: | (1132)
1	-14.0605	he’s | ‘christ’s
1	-14.0605	him, | nullhe,
1	-14.0605	1] | e],
1	-14.0605	discrete | 79.45
1	-14.0605	<thoracic | structure>)
1	-14.0605	baseline | 6.2742
1	-14.0605	r¨ognvaldsson | helgad´ottir
1	-14.0605	rb|ka01 | −0.298
1	-14.0605	s0: | vprime
1	-14.0605	∗department | athens
1	-14.0605	all-edges | noisy-seed
1	-14.0605	pθo(ot |ht)def=
1	-14.0605	ece | [war]
1	-14.0605	3,000 | colorful
1	-14.0605	se, | oe,
1	-14.0605	vieu | aurnague,
1	-14.0605	regressions | past
1	-14.0605	rewards | 0.99,
1	-14.0605	pi(α) | pa(β|η)
1	-14.0605	scfs | verb
1	-14.0605	substringmatch | numberagreement
1	-14.0605	e2 | r(e2)
1	-14.0605	θa | θr.
1	-14.0605	na¨ıve | (nb),
1	-14.0605	p(¯q|¯a)
1	-14.0605	irvine | lexicons
1	-14.0605	chicago, | sravana@cs.uchicago.edu
1	-14.0605	correlation;ρ: | rho.
1	-14.0605	plays | chelsea
1	-14.0605	―ق | ―q‖),
1	-14.0605	asl | (asl+gi)
1	-14.0605	यहा | े।
1	-14.0605	xuanjing | fudan
1	-14.0605	boyd, | golder,
1	-14.0605	crawley, | kleinman,
1	-14.0605	ngram-r | i-hmm-token-r
1	-14.0605	deviant | (domingos,
1	-14.0605	nn-test-1-hum | nn-test-2-cb
1	-14.0605	’,’, | ’new’,
1	-14.0605	53.33 | 34.27
1	-14.0605	news | (0.09)
1	-14.0605	expand | hyponymns
1	-14.0605	utrecht | 8:32.
1	-14.0605	lisbon | liacc
1	-14.0605	ldc2003e14, | ldc2005t06,
1	-14.0605	86.28 | pp-positive
1	-14.0605	to|t|log|s|.
1	-14.0605	combase | diein
1	-14.0605	brunswick, | 08901
1	-14.0605	ile | vaches
1	-14.0605	domaine | vilvert
1	-14.0605	acquire | signifiers
1	-14.0605	ebay | hamilton
1	-14.0605	−102468 | ’30’40
1	-14.0605	tree-to-string | t≅
1	-14.0605	‘sub | −−−−−−−−→sub
1	-14.0605	q5: | q14
1	-14.0605	cd | t{gov(d)
1	-14.0605	={0,1, | n}.
1	-14.0605	[evaluee | [reason
1	-14.0605	contour | legendre
1	-14.0605	(luo | rahman
1	-14.0605	yhteydess¨a | j¨asen
1	-14.0605	t/s | npt
1	-14.0605	mt02, | 04,
1	-14.0605	article | compl.
1	-14.0605	nounphrasesin | featuresto
1	-14.0605	m8 | m13
1	-14.0605	you’ll | here?
1	-14.0605	ryan | ryanmcd@google.com
1	-14.0605	ddk=25 | 88.29
1	-14.0605	some | (b|φ(x,y)).
1	-14.0605	higins | sadock
1	-14.0605	assert | pfc
1	-14.0605	proclitics | enclitics
1	-14.0605	goldwater | priors,
1	-14.0605	2424 | 9616
1	-14.0605	md | 29–32).
1	-14.0605	wdt | #
1	-14.0605	iarg1 | unlikely.
1	-14.0605	kidnap, | abduct,
1	-14.0605	brent | cartwright
1	-14.0605	s49s48 | s52s48s37
1	-14.0605	52.05 | 53.71
1	-14.0605	pr(w|t),
1	-14.0605	ncirclemultiplydisplay | α(qi)
1	-14.0605	(possible’ | r1)(rel’
1	-14.0605	pre | alize
1	-14.0605	blywn | lywn
1	-14.0605	linux | 2.0ghz
1	-14.0605	fy | almady
1	-14.0605	h., | crasborn
1	-14.0605	(pk) | evaluation
1	-14.0605	cabin | altitude
1	-14.0605	(moschitti | zanzotto,
1	-14.0605	|rs)
1	-14.0605	1(logs1) | ={0.1,0.2,...}do
1	-14.0605	position, | closet
1	-14.0605	(she | gelopen
1	-14.0605	n-ary | relations.
1	-14.0605	programme | (faust),
1	-14.0605	grammar): | new:
1	-14.0605	2000 | 8,936
1	-14.0605	iso-timeml | (iso,
1	-14.0605	[e1 | zemin
1	-14.0605	kahani | nadya.f.sg=dat
1	-14.0605	belgium | biotechnology
1	-14.0605	¯s | nullw∈¯s
1	-14.0605	local, | shimmer
1	-14.0605	(nn, | nnps)
1	-14.0605	smoothing | φn(wi−1i−n+1),
1	-14.0605	co-clustering | (dhillon,
1	-14.0605	|k) | ·p(nk|nk−1,k)}
1	-14.0605	i(ai,j) | o(ai,j)
1	-14.0605	connotation | “prevent”,
1	-14.0605	advanced | gramplus,
1	-14.0605	init(r) | fin(r)
1	-14.0605	dekai | hkust
1	-14.0605	lsh | paraphrase.
1	-14.0605	pal | pilots'
1	-14.0605	vote | voted
1	-14.0605	ηr | ηc
1	-14.0605	library | pike
1	-14.0605	“rennrad” | bike)
1	-14.0605	m1inm2 | m2inm1
1	-14.0605	r3 | domain(r3)
1	-14.0605	ordbok | medeltids-spr˚aket,
1	-14.0605	(1−λn(φn))·˜pn−1(wi|wi−1i−n+2)
1	-14.0605	cha | sak
1	-14.0605	o2 | o3:
1	-14.0605	20k | 36.68
1	-14.0605	guthrie | hepple,
1	-14.0605	d84 | d103d103
1	-14.0605	tm | ‘example’
1	-14.0605	“translationese” | “third
1	-14.0605	p(l | θr(r)
1	-14.0605	nth | bom
1	-14.0605	[k,lscript)∈t | [k,lscript)∈n
1	-14.0605	kulick | bies
1	-14.0605	undiacritized | (lmm)
1	-14.0605	dt | {xtn
1	-14.0605	etc. | control]
1	-14.0605	4-5 | 80-85.
1	-14.0605	celikyilmaz | hakkani-t¨ur
1	-14.0605	n(gnppa) | n(e-gnppa)
1	-14.0605	es→en | zh→en.
1	-14.0605	a(ai) | h(ai)
1	-14.0605	“np | pp[around
1	-14.0605	up:6, | fall:2,
1	-14.0605	verde, | timor,
1	-14.0605	d3 | 0.27
1	-14.0605	discr | pancies
1	-14.0605	pn(e|f)·bleu(e,eprime) | ·minparenleftbige1−r
1	-14.0605	blues | add->genre(blues)
1	-14.0605	extractive | abstractive.
1	-14.0605	(also | operation)
1	-14.0605	<s | >s
1	-14.0605	lager | lime,
1	-14.0605	crasborn | 2010.
1	-14.0605	f8 | f20
1	-14.0605	baccianella | sebas-tiani
1	-14.0605	burek | gerdemann
1	-14.0605	didn’t | negation(n’t,
1	-14.0605	jiangsu | chibi
1	-14.0605	(category | e.g.,
1	-14.0605	lucs | sucs
1	-14.0605	118 | 530,000
1	-14.0605	sub | ntl
1	-14.0605	sentiment | score vector
1	-14.0605	[itu-t | p.85
1	-14.0605	kernel | libsvm2
1	-14.0605	yang, | connectivity
1	-14.0605	pθ(y|x).
1	-14.0605	2,117 | 12,150
1	-14.0605	phrase-based | (psmt)
1	-14.0605	haut | dernière
1	-14.0605	evening | day
1	-14.0605	“during | intermembrane
1	-14.0605	variable-length | ranks
1	-14.0605	geospatial | (egenhofer,
1	-14.0605	ratable | mcdonald,
1	-14.0605	p(crt | |itnull1)·
1	-14.0605	pr-as | e-dmv
1	-14.0605	ordona | 01-237
1	-14.0605	1,400 | pollution
1	-14.0605	semi-learnt | fullylearnt
1	-14.0605	vn | l(3)
1	-14.0605	lieutenant: | amani:
1	-14.0605	―هك | ―kah‖),
1	-14.0605	red, | pink),
1	-14.0605	created | zero-anaphora,
1	-14.0605	s(t) | wnφn(k,m,lscript)
1	-14.0605	conrath, | schtze,
1	-14.0605	posttranslational | (ohta
1	-14.0605	tgthr, | luvvvin,
1	-14.0605	kentaro | yuji
1	-14.0605	c[s:i] | c[s:i+1]
1	-14.0605	/jcs | /jca
1	-14.0605	kw-syn | tr-exact
1	-14.0605	disambiguations | zhao,
1	-14.0605	similes | s”,
1	-14.0605	regularization | shrinkage
1	-14.0605	"nn" | atw
1	-14.0605	kazama | torisawa
1	-14.0605	greenbacker | “raw
1	-14.0605	kept | arm’s
1	-14.0605	acoustically | /d/,
1	-14.0605	obtaining | 2)1(
1	-14.0605	bon−1(wi−11 | wi−1i−n+2
1	-14.0605	protein1 | protein2
1	-14.0605	steady* | ton*
1	-14.0605	mohit | {mbansal,klein}@cs.berkeley.edu
1	-14.0605	montr´eal | polytechnique
1	-14.0605	(1−ii,r)p(xi) | j<i:ri=rj
1	-14.0605	143,580 and|df|=
1	-14.0605	ies | jet
1	-14.0605	baked | “cake”
1	-14.0605	0.5123† | 0.3424
1	-14.0605	“bought” | “lotus”
1	-14.0605	oriented | (udop)
1	-14.0605	primary | 24.4
1	-14.0605	(“gm | opel”,“opel
1	-14.0605	“seen” | “wild”
1	-14.0605	ˆe | ee
1	-14.0605	ibm2 | ibm1-inv
1	-14.0605	“right” | “towards
1	-14.0605	intake | isoflavon
1	-14.0605	society | mathematics:
1	-14.0605	nakov | hearts
1	-14.0605	smooth1 | smooth2
1	-14.0605	|r(s)∩r(t)||r(s)∪r(t)|.
1	-14.0605	asd | nondevelopmental
1	-14.0605	dnis | inis
1	-14.0605	mmil | inter-annotation
1	-14.0605	extended | (xwn)
1	-14.0605	rigau, | maurizio,
1	-14.0605	(duboue | chu-carroll,
1	-14.0605	lisbon, | 204.
1	-14.0605	alternating | (ao)
1	-14.0605	λi | (iis)
1	-14.0605	1k+i | i<=
1	-14.0605	q13 | assess
1	-14.0605	realization | shotgun
1	-14.0605	bretonnel | kevin.cohen@gmail.com
1	-14.0605	(stk) | duffy,
1	-14.0605	l-type | conceived,
1	-14.0605	j. | heights,
1	-14.0605	hu’s | wu’s
1	-14.0605	parenleftbiggproductdisplay | (rec.
1	-14.0605	airliner | cabin
1	-14.0605	rada | denton,
1	-14.0605	0.372 | 2.13%
1	-14.0605	robots | (brenner,
1	-14.0605	obs | meas
1	-14.0605	majed | ...
1	-14.0605	pr´esident, | syst`eme
1	-14.0605	versus | switch,
1	-14.0605	sollen | better
1	-14.0605	archival | itself.
1	-14.0605	trigger | dictionary+rules
1	-14.0605	read | book’,
1	-14.0605	case-sensitive | (papineni
1	-14.0605	(lesher | rinkus,
1	-14.0605	levels | 66.7%,
1	-14.0605	72.0 | 76.9
1	-14.0605	normalises | modernised
1	-14.0605	sadness | my heart.
1	-14.0605	aviv | go.perf.fem.sg
1	-14.0605	439,513 | en-se
1	-14.0605	landauer, | mcnamara,
1	-14.0605	korea | dipl.
1	-14.0605	leading | beverage
1	-14.0605	readymade | marcel
1	-14.0605	conclusion, | expiry
1	-14.0605	(fn), | (ac),
1	-14.0605	abbas] | [national
1	-14.0605	np→ | nnnp→
1	-14.0605	wiki200 | wiki100h
1	-14.0605	yes-no | declar.
1	-14.0605	t-unit | (hunt,
1	-14.0605	u1r, | u20a
1	-14.0605	90.26 | 91.38
1	-14.0605	'water' | declared
1	-14.0605	europarl | uno
1	-14.0605	en-de | 0.1692
1	-14.0605	labs | clara,
1	-14.0605	s49s53s48s48s48 | s32s56
1	-14.0605	commit | commit(r(pii,pij),dn
1	-14.0605	california | waterfront
1	-14.0605	tout | nombreuses
1	-14.0605	scale-invariant | (sift)
1	-14.0605	n−iσ−|t | n−i−|t
1	-14.0605	hcr | conservatives
1	-14.0605	(casacuberta | mari˜no
1	-14.0605	flyslip | drugnerar
1	-14.0605	confident | 94.41
1	-14.0605	16,125 | 1,151
1	-14.0605	officials | steering
1	-14.0605	(malioutov | uj,...,uk
1	-14.0605	npj | arg1-arg0,
1	-14.0605	web1tgram | gogle,
1	-14.0605	2,848 | 1,533
1	-14.0605	lattice-aware | (laisr)
1	-14.0605	(biden,president) | (obama,president)
1	-14.0605	0.062 | 0.138
1	-14.0605	77.3% | 61.2%
1	-14.0605	transformed | β1,
1	-14.0605	eckert, | pieraccini
1	-14.0605	n(e) | parent(n(e))
1	-14.0605	paraphrase | paraphrasee’s
1	-14.0605	furry | be
1	-14.0605	antti | pahikkala
1	-14.0605	aditya | grzegorz
1	-14.0605	c∈cliq(x,y) | score(c,xprime,yprime)
1	-14.0605	1850-1950 | 40.29
1	-14.0605	polarity | gallup
1	-14.0605	bm | (85.67/86.97)**
1	-14.0605	majid | garner
1	-14.0605	holmes, | 11,
1	-14.0605	york* | year*
1	-14.0605	transcribe | responses.
1	-14.0605	simon | columbia,
1	-14.0605	d∈d | pp(wl,tl,gl|d)
1	-14.0605	5, | ;2<
1	-14.0605	n: | [from
1	-14.0605	(lavrenko | shtrimberg,
1	-14.0605	1w) | 1+(•w2w)
1	-14.0605	59.1 | 62.6
1	-14.0605	directions, | glue
1	-14.0605	∼ dir(·|
1	-14.0605	to(e2,e1,x3):0 | get(e1,x1,x2):0
1	-14.0605	#system-outputs | beds
1	-14.0605	s57s48 | s57s52
1	-14.0605	finanzminister | kann,
1	-14.0605	ν||r||2f | λ||ˆθk||22
1	-14.0605	month | milk.
1	-14.0605	length(i) | bk:=
1	-14.0605	(brent, | venkataraman,
1	-14.0605	ncirclemultiplydisplay | α(ai)
1	-14.0605	ei1 | ei2.
1	-14.0605	sif | srf-sif
1	-14.0605	p(t|s),
1	-14.0605	74% | 60%
1	-14.0605	(active | passive)
1	-14.0605	chinese, | japanese.
1	-14.0605	non-parametric | (py)
1	-14.0605	eric/gat/voetbalen/in | sporthal.
1	-14.0605	¨ou, | u¨o,
1	-14.0605	superfluous | suffixes
1	-14.0605	cea | 760k
1	-14.0605	hassibling | orgleader
1	-14.0605	control | games.
1	-14.0605	amsterdam | 904,
1	-14.0605	d∈y (p(d|g)·eλ·δ(d,d))
1	-14.0605	(murawaki | kurohashi,
1	-14.0605	including | (including | east
1	-14.0605	πππ | sayπ(
1	-14.0605	n0w; | n1w;
1	-14.0605	pron | (conjunctions),
1	-14.0605	redwoods | derivations
1	-14.0605	<num>sg | series.
1	-14.0605	+φ | ł+
1	-14.0605	sank | bismarck.
1	-14.0605	smith∗ | telecomunicac¸
1	-14.0605	feature: | (“gm
1	-14.0605	cilibrasi | unstructured
1	-14.0605	massimo | essex,
1	-14.0605	input: | reordered:
1	-14.0605	40-42 | antwerpen,
1	-14.0605	ldc | (ldc2005t34)
1	-14.0605	college | pittsburgh’s
1	-14.0605	tr20 | tr50
1	-14.0605	my<det><pos><pl> | haber_pp{
1	-14.0605	spouse, | founded,
1	-14.0605	“how | deliberated
1	-14.0605	4-grams | 7-grams
1	-14.0605	various | (lexical,
1	-14.0605	“cr” | “notcr”
1	-14.0605	←ˆe | >g(ˆes)
1	-14.0605	+semanticfeature | +semanticfeature+orthofeatures
1	-14.0605	(s1,s2), | exchanging
1	-14.0605	trigger-based | (rosenfeld,
1	-14.0605	descartes | 77454
1	-14.0605	torsten | iryna
1	-14.0605	maximum, range, | standard
1	-14.0605	usc | usc
1	-14.0605	babytalk | bt45
1	-14.0605	them | onlysupplylimitedinformationliterally.
1	-14.0605	te | time=abs
1	-14.0605	(beukelman | mirenda,
1	-14.0605	inline | ∈d,
1	-14.0605	fk | aβ
1	-14.0605	price, | shrikanth
1	-14.0605	meyer | eisen
1	-14.0605	ctpr+kprel | ctpr+kprelint
1	-14.0605	np−sbj | pp−clr
1	-14.0605	es-en | ja-en
1	-14.0605	satisfies | d3–d6,
1	-14.0605	proposition | “person
1	-14.0605	fei | weng2
1	-14.0605	thementionsagreeingender;elsen | thesameverb.
1	-14.0605	operate | dialogue-act
1	-14.0605	sys4 | donthave
1	-14.0605	(σ|i1|i2|...|id+1,β,a)
1	-14.0605	netherlands | (nwo),
1	-14.0605	35.2 | fl
1	-14.0605	standard | 154.3
1	-14.0605	0.500.55 | vembdp-1
1	-14.0605	receiver | pobjdet
1	-14.0605	million) | (newstest2009)
1	-14.0605	axiom | null0
1	-14.0605	lch | is-a
1	-14.0605	semeval | discourse’
1	-14.0605	b(d) | b(d
1	-14.0605	multi | (lmbot)
1	-14.0605	fuliang | 2research
1	-14.0605	(ward | tsukuhara,
1	-14.0605	str | rin
1	-14.0605	two-sided | (4873
1	-14.0605	←nzi|di−1 | ←nzi|di
1	-14.0605	shatin, | kfwong}@se.cuhk.edu.hk
1	-14.0605	200m | times).
1	-14.0605	heat | furnace
1	-14.0605	enjoy, | upset,
1	-14.0605	south | overnight
1	-14.0605	1.23 | -0.57
1	-14.0605	(whittaker | raj,
1	-14.0605	27 | 28–36,
1	-14.0605	italianbodytext | englishfigurecaption
1	-14.0605	watkinson | manandhar’s
1	-14.0605	ep10 | tfp
1	-14.0605	nounphraseis | mention.thefea-
1	-14.0605	chen, | tamang,
1	-14.0605	orunknown,determinedusinga | commonnounphrasesorpronouns
1	-14.0605	acknowledgements | emnlp
1	-14.0605	88.4 | 75.6
1	-14.0605	named(l0, | z0
1	-14.0605	1410 | 845
1	-14.0605	(s-neg), | subjective-neutral
1	-14.0605	lm, | ˜p(wi|wi−1i−n+1)
1	-14.0605	74.85 | 53.02
1	-14.0605	|cm|·|cn| | daj∈cn
1	-14.0605	21.6 | 14.1
1	-14.0605	grande | sul,
1	-14.0605	nullnulli | nullnullj
1	-14.0605	 flu |  japan.
1	-14.0605	32.68 | +
1	-14.0605	ladies | gentlemen
1	-14.0605	coupled | (chmm)
1	-14.0605	truth | implicatures
1	-14.0605	1n | r∈c|r(r)=r
1	-14.0605	n-n | n-a
1	-14.0605	e-to-f | g100g62g890g258g282g282g349g410g349g381g374g855g862g296g367g381g449g400g3g349g374g410g381g3g410g346g286g3g400g286g258g863
1	-14.0605	announced | 3ds
1	-14.0605	: | seperti
1	-14.0605	auer | kiesb
1	-14.0605	[e2 | judicial
1	-14.0605	αi | cov
1	-14.0605	walking | dog
1	-14.0605	47th | afnlp,
1	-14.0605	wander | 6506
1	-14.0605	unsuccessful, | reaction
1	-14.0605	vp | ;n
1	-14.0605	data-driven | proposed.
1	-14.0605	presentation | (multimodal
1	-14.0605	=− | c∈sub(n)
1	-14.0605	0.694 | 0.808
1	-14.0605	0.925 | 0.941
1	-14.0605	sollten | milch
1	-14.0605	javascript, | actionscript;
1	-14.0605	match(di,dj) | mismatches
1	-14.0605	rmse0.5 | valence/arousal
1	-14.0605	conferences | sigir,
1	-14.0605	62.9 | 72.0
1	-14.0605	g0 | m0
1	-14.0605	it | gorgeous!
1	-14.0605	farthest | (internal
1	-14.0605	advantages | disadvantages,
1	-14.0605	la1 | ra1
1	-14.0605	(19.3% | 20.1%)
1	-14.0605	strong | year-to-year
1	-14.0605	u| | l∈tu
1	-14.0605	women’s | crosby
1	-14.0605	(litman | silliman,
1	-14.0605	exercise | genpex.
1	-14.0605	ana:t29 | ant:t27
1	-14.0605	yohei | noriko
1	-14.0605	kontuan | inessive
1	-14.0605	“various | marcus,
1	-14.0605	chen, | wei,
1	-14.0605	(hancock | mihalcea
1	-14.0605	pmi(σ) | pmi(wi,wj)
1	-14.0605	combinatory | x/y
1	-14.0605	 woman |  to
1	-14.0605	conf(ssr[1]) | conf(ssr[2])
1	-14.0605	uniondisplay 1≤i≤|t|
1	-14.0605	0.783 | 2.17%
1	-14.0605	slipnet | unstated
1	-14.0605	(or | change)
1	-14.0605	lengths | angl
1	-14.0605	jean | robeson
1	-14.0605	pre(xi) | post(xi)
1	-14.0605	fertility | |≤
1	-14.0605	kpj | rpj,
1	-14.0605	duluth-2 | tscore
1	-14.0605	p(wi|wi−n+1...wi−1)
1	-14.0605	86.28 | 90.85
1	-14.0605	document; | static
1	-14.0605	nullnullnull1, | nullnullnull2
1	-14.0605	cardie, | pierce
1	-14.0605	ˇdςγfqklmnhwy | ł,
1	-14.0605	‘die | spielt’
1	-14.0605	embed | loss-trained
1	-14.0605	wilson | sperber
1	-14.0605	scfgs | stags
1	-14.0605	anna | alk23@cam.ac.uk
1	-14.0605	personin | 0.9985
1	-14.0605	newsshuffle | longrange
1	-14.0605	987,401 | 984,731
1	-14.0605	i;ib;k | i;ib;j
1	-14.0605	funds | hubei
1	-14.0605	back-off | (+bm).
1	-14.0605	nexavar, | showed
1	-14.0605	#systems | mbrsc.
1	-14.0605	‘‘ | −lrb−
1	-14.0605	thefinalclusteringresultsare | byclosest-firstclusteringmethod.officialtest
1	-14.0605	∗∗ | science,
1	-14.0605	xms¯h | ‘five
1	-14.0605	pour | r´epublicains.
1	-14.0605	74.6 | 69.6
1	-14.0605	determiners | every,
1	-14.0605	〈t,t〉 | determinization,
1	-14.0605	(animals | (browsing
1	-14.0605	danling | china
1	-14.0605	<inscheme | <au>
1	-14.0605	stock | fund
1	-14.0605	negationslash= | β∗
1	-14.0605	p | mettre
1	-14.0605	genre, | text-target
1	-14.0605	··· | tij+2ij+2
1	-14.0605	s56s56 | s57s50
1	-14.0605	(wh)iµsummationtext | wia←wia
1	-14.0605	kulkarni | 02139,
1	-14.0605	las | 73.11%
1	-14.0605	vw1⊕w2 | avw1
1	-14.0605	bar | businessweek,
1	-14.0605	|a∩p||a| | |a∩s||s|
1	-14.0605	(nouns, | adverbs),
1	-14.0605	(yamamoto | church,
1	-14.0605	p(rs | p(rt
1	-14.0605	3a) | 3b)
1	-14.0605	floor | fi(p)
1	-14.0605	pr(v|n1,n2).
1	-14.0605	“blue | 20:102
1	-14.0605	(upper | panel).
1	-14.0605	mitchell, | wang.
1	-14.0605	0.6993 | 0.6126
1	-14.0605	sumita | hikaridai,
1	-14.0605	eigenheimbesitzer | gesch¨utzt
1	-14.0605	nullm, | nullgnull,
1	-14.0605	alignments} | if
1	-14.0605	try | monolingually
1	-14.0605	adjoined | nn-x
1	-14.0605	ka | alkatb
1	-14.0605	dknulld | nkproductdisplay
1	-14.0605	pr(w|d)
1	-14.0605	mcnemar’s | 1947)
1	-14.0605	affordable | authorised
1	-14.0605	t1) | t1),
1	-14.0605	instrument), | subclass
1	-14.0605	new | summationtextdj=0
1	-14.0605	siblings | average,which
1	-14.0605	(person), | (location),
1	-14.0605	journals | acp
1	-14.0605	(tsagkias | balog,
1	-14.0605	matthew | wing
1	-14.0605	has, | flowsthrough
1	-14.0605	vectori1 | vectori2
1	-14.0605	semanticstructuresin | interannotator
1	-14.0605	(corston-oliver | specia
1	-14.0605	unlexicalized | (severyn
1	-14.0605	hubb | isten
1	-14.0605	meanwhile,we did | featuresto
1	-14.0605	0.5647 | 0.4542
1	-14.0605	giza | diwa,
1	-14.0605	pkn(w3|w2) | pkn(w3)
1	-14.0605	nuggets | tac,
1	-14.0605	null | 0.075,
1	-14.0605	5.8m | 81.11
1	-14.0605	91.12 | lch
1	-14.0605	ministre | ´etrang`eres,
1	-14.0605	fori | 1,...,n,
1	-14.0605	ist | wahrscheinlichkeit
1	-14.0605	hui | dept.
1	-14.0605	null | edge(nullnullnullnullnull)
1	-14.0605	0.363 | 0.0079
1	-14.0605	cuny | validation
1	-14.0605	srf | inf-sif
1	-14.0605	hydroxylation | dehydroxylation
1	-14.0605	mastroianni, | cancer.
1	-14.0605	2011 | bossy1,
1	-14.0605	vino | por
1	-14.0605	fieldwork | multidialectal
1	-14.0605	(pronouns), | prt
1	-14.0605	haque, | naskar,
1	-14.0605	bewilderment | face.
1	-14.0605	inequalities | ∀k⊆{1,...,k}:|k|=
1	-14.0605	jordi | turmo.
1	-14.0605	lette | e>
1	-14.0605	company2 | medtronic
1	-14.0605	al | (tomanek
1	-14.0605	0.837 | 0.889
1	-14.0605	localization, | catabolism,
1	-14.0605	shi | guojia
1	-14.0605	town | bauska
1	-14.0605	related; | tk2
1	-14.0605	oceanographers | arctic
1	-14.0605	reduction | 32.9%
1	-14.0605	aozhou | shaoshu
1	-14.0605	unify | (newly
1	-14.0605	jail | detained,
1	-14.0605	“how | school
1	-14.0605	lette | ofth
1	-14.0605	o1 | mstparser1
1	-14.0605	nds*t | g534
1	-14.0605	almost | benign
1	-14.0605	mehdi | 14627,
1	-14.0605	schank | abelson
1	-14.0605	σprime | ⇒σprimegprefix
1	-14.0605	λ2, | ⇒(
1	-14.0605	0.469 | (pb+ds)
1	-14.0605	metric | 49.64%
1	-14.0605	phoenix | subgrammar,
1	-14.0605	(weka | c4.5)
1	-14.0605	ger¨at | gr¨oßtm¨ogliche
1	-14.0605	(ends) | (last)
1	-14.0605	nter | ella
1	-14.0605	recursively | edu’s
1	-14.0605	decided | discretize
1	-14.0605	(max | wisniewski,
1	-14.0605	eight | (seven
1	-14.0605	score: p(z|x)
1	-14.0605	13.7% | 50.1%
1	-14.0605	piwek | p.piwek@open.ac.uk
1	-14.0605	heiskell | congress.
1	-14.0605	ωr | 1|r|
1	-14.0605	misinformation | disinformation
1	-14.0605	σ2 | (ωr
1	-14.0605	/∈ | ←wk,m,wd,n−1
1	-14.0605	jun | hideki
1	-14.0605	taylor | {tberg,
1	-14.0605	formulas | grade
1	-14.0605	seconds | quadra
1	-14.0605	xiaojun | university),
1	-14.0605	metric | met1.2
1	-14.0605	cubed | 0.7031
1	-14.0605	graham | honmachi,
1	-14.0605	alessandro | daniele
1	-14.0605	berv | wikv
1	-14.0605	john(x1):20 | mary(x3):20
1	-14.0605	verhagen, | schilder,
1	-14.0605	(-0.03 | (-0.06
1	-14.0605	arg2 | benefactive
1	-14.0605	katz | giesbrecht,
1	-14.0605	q(θ) | t,h
1	-14.0605	113,885 | 146,405
1	-14.0605	bizer, | kobilarov,
1	-14.0605	αj,k | p(yj|xk)
1	-14.0605	prep, | headverb,
1	-14.0605	x1:np∗ | x2:np↓
1	-14.0605	nullnullnullnullnullnullnullnullnull | (null)).
1	-14.0605	includesome | featureswhichintuitively
1	-14.0605	needed | maximization.
1	-14.0605	vertical | lateral,
1	-14.0605	jp | jp+fr+de
1	-14.0605	tds | pp[to]
1	-14.0605	pintrich | degroot
1	-14.0605	p(s1 | randomclicker,
1	-14.0605	svmbased | llda-based
1	-14.0605	vbn | np{answer}”:
1	-14.0605	1: | ……
1	-14.0605	bcpo−completion(s) | apcl(s)
1	-14.0605	92.20% | 84.00%
1	-14.0605	yi | pc(〈xi,yi〉|xi)
1	-14.0605	german | ‘spatial
1	-14.0605	s54s48s48s48 | s32s78
1	-14.0605	replacement, | candlestick
1	-14.0605	pneumonia, | wheezing,
1	-14.0605	multinomials | per-document
1	-14.0605	qu|j|−uq|j|
1	-14.0605	32, | left.
1	-14.0605	eprimes←substitute(ecur,a,j) | ensure:
1	-14.0605	cn | (bertoldi
1	-14.0605	at(r,n1) | at(r,n1)∧future(¬at(r,n1))
1	-14.0605	null(null)) | 100%”,
1	-14.0605	broad | ave.
1	-14.0605	james | {keboyer,
1	-14.0605	52.5 | trad+align
1	-14.0605	hirohata | shatkay
1	-14.0605	36-39) | 45-49)
1	-14.0605	miller, | beckwith,
1	-14.0605	94.2 | 87.7
1	-14.0605	m3 | m12
1	-14.0605	judges | “paraphrases”
1	-14.0605	sauri, | knippen,
1	-14.0605	verkehrszeichen | verkehr
1	-14.0605	predictors | variable:
1	-14.0605	8.6 | 4.6
1	-14.0605	conversational | (maskey
1	-14.0605	news | (0.06)
1	-14.0605	numberof | truementions
1	-14.0605	omri | jerusalem
1	-14.0605	stem+morph | +unique
1	-14.0605	theresultsfromtable5tellusthatcombiningdifferent | 46.66%
1	-14.0605	carberry | elzer*
1	-14.0605	ast | 0.075k
1	-14.0605	abl | zaanen,
1	-14.0605	‘pv’ | ‘np-obj’
1	-14.0605	npk | predk,
1	-14.0605	secretary | pinto
1	-14.0605	lmf | mrd
1	-14.0605	phonological | nevins,
1	-14.0605	rj | cand⊕rj
1	-14.0605	triangleright | losses:
1	-14.0605	mann, | thompson.
1	-14.0605	bien | clustered
1	-14.0605	level(high | measure
1	-14.0605	mitre | bedford,
1	-14.0605	commission, | (lopez);
1	-14.0605	sight | graphic.
1	-14.0605	include | spiritual,
1	-14.0605	17.95 | 18.35
1	-14.0605	ibm | (soltau
1	-14.0605	αz | βz
1	-14.0605	premise, | consequence.
1	-14.0605	gaizauskas, | sundheim,
1	-14.0605	nonea | vine
1	-14.0605	[3,6], | [2,6]
1	-14.0605	sb | (i,t)〉,
1	-14.0605	ew[cy⋆i |
1	-14.0605	10,000 | (“10k”),
1	-14.0605	christine | cece
1	-14.0605	cbl | ibl
1	-14.0605	¸æk | pæ
1	-14.0605	0.5349† | 0.3303
1	-14.0605	··· | algorithm2···
1	-14.0605	moving | δi
1	-14.0605	positive-example | transgene
1	-14.0605	functors | withholding,
1	-14.0605	|z | ,β)
1	-14.0605	classificationbasedpre-clusterpairmodelto | pre-clustersrefer
1	-14.0605	ming | tiejun
1	-14.0605	baseline1 | pag2
1	-14.0605	completely | topic.
1	-14.0605	na | -12
1	-14.0605	85.6 | 90.2
1	-14.0605	hall, | weka
1	-14.0605	22: | 23:
1	-14.0605	0.119 | 0.168
1	-14.0605	count | c(z,w)
1	-14.0605	radicalbig | )2
1	-14.0605	avi | multinomial(φcvi
1	-14.0605	c0 | (σ|h1,βi,a)
1	-14.0605	pos-taging | lematization
1	-14.0605	factoids | “poor
1	-14.0605	releasing | 2004,
1	-14.0605	she | adenopathy
1	-14.0605	1113 | convinceme.net
1	-14.0605	prasad, | dinesh,
1	-14.0605	a-priori | θr(r),
1	-14.0605	ejection | echocardiogram
1	-14.0605	pasages | grade
1	-14.0605	f−1productdisplay | b(wn−1i
1	-14.0605	(johnson | 2000).
1	-14.0605	messages: | apre
1	-14.0605	web-derived | mid-2009
1	-14.0605	inhibitory | i(kappa)b(alpha)
1	-14.0605	magnet | answer/describe/process
1	-14.0605	p(r5 | r1null
1	-14.0605	69.6% | 3–6
1	-14.0605	cox | semiparametric
1	-14.0605	pennacchiotti | pantel.
1	-14.0605	cn-cn | en-cn
1	-14.0605	(50) | or-after-yn
1	-14.0605	rc | explode:s,
1	-14.0605	lacalle | agirre.
1	-14.0605	bnnf | ennf
1	-14.0605	massacre | columbine
1	-14.0605	.27 | .51
1	-14.0605	miles | kilometers)
1	-14.0605	blue | brown
1	-14.0605	literally | “vb”
1	-14.0605	2.5% | total:
1	-14.0605	galileo | lise
1	-14.0605	fine | fine+patterns
1	-14.0605	α1, | α2],
1	-14.0605	meteor-adq | meteor-hter
1	-14.0605	α(p) | n,q1,...,qn
1	-14.0605	macro-avg | 0.473
1	-14.0605	hls | french
1	-14.0605	logistic | 1/(1
1	-14.0605	e∈e p(e|f)·s(e,eprime)
1	-14.0605	bomber” | helicopter”
1	-14.0605	(mrl) | (tsarfaty
1	-14.0605	believeocean | waters”
1	-14.0605	10-best | 30-best
1	-14.0605	dell’orletta, | montemagni
1	-14.0605	s1rwc, | s0cs1w,
1	-14.0605	miller | stoytchev
1	-14.0605	narrative | records.
1	-14.0605	fixation | onset,
1	-14.0605	interoperability | (icgl
1	-14.0605	hybrid | spanish–
1	-14.0605	(swts) | (daille
1	-14.0605	21, | warszawa,
1	-14.0605	umls | abner
1	-14.0605	np* | pp*
1	-14.0605	oct | spaniards
1	-14.0605	1001–1151 | 301–325
1	-14.0605	wp | verb-arg1
1	-14.0605	florham | 07932
1	-14.0605	1550-1650 | 1850-1950
1	-14.0605	age | years;
1	-14.0605	ξ | subrequests.
1	-14.0605	,ˆg | k∏
1	-14.0605	management | traps
1	-14.0605	m> | bin
1	-14.0605	88.22 | 85.48
1	-14.0605	domain(rj) | consistent(pr3,pri
1	-14.0605	yit | yprimet+1:t
1	-14.0605	63.23%, | ceaf
1	-14.0605	07/ce/i1142) | (www.cngl.ie)
1	-14.0605	detect | mentionsfrom
1	-14.0605	output: | lmert←generatelattices
1	-14.0605	green; | yellow;
1	-14.0605	aisb | “communication,
1	-14.0605	ranked | umass,
1	-14.0605	adjoining | np0,3
1	-14.0605	diego | 92093-0533
1	-14.0605	cutaneous | analgesia
1	-14.0605	g307 | g652
1	-14.0605	ontonotescorpus | featuresto
1	-14.0605	0.5) | 0.25)
1	-14.0605	rewards | ‘any
1	-14.0605	rejection | proposals.
1	-14.0605	dimensions, | reveal
1	-14.0605	colour | (1969).
1	-14.0605	df(2): | green]
1	-14.0605	αθ | ×+
1	-14.0605	te] | -て
1	-14.0605	d1: | ]s,
1	-14.0605	developmentdatasetwith | appositive
1	-14.0605	elop | cou
1	-14.0605	dreuw | wheatley
1	-14.0605	adjoining | (tags)
1	-14.0605	df’s | milliseconds,
1	-14.0605	b⊕mi | augment(c,antei,mi)
1	-14.0605	af | +as
1	-14.0605	dredze, | rao,
1	-14.0605	31.1 | 36.1
1	-14.0605	78.7 | 68.8
1	-14.0605	respect | zero.
1	-14.0605	italy), | age-wise
1	-14.0605	nullu | nullunullnull
1	-14.0605	85.92 | 90.60
1	-14.0605	precisely: | )⇒w
1	-14.0605	greater | lesser,
1	-14.0605	jakub; | sloot,
1	-14.0605	four | (e3)
1	-14.0605	(unigram, | trigram)
1	-14.0605	moldovan | {eduardo,moldovan}@hlt.utdallas.edu
1	-14.0605	schloß | d-53754
1	-14.0605	h3r3, | betta,
1	-14.0605	labalme | commune
1	-14.0605	(han | gamon,
1	-14.0605	nullsummationtextp:snullp1(p) | nullsummationtextp:snullp2(p)
1	-14.0605	afinitor | 05/18/2008.
1	-14.0605	gosogbeoseuga | chimbeomhae
1	-14.0605	p∗ | [s,s]
1	-14.0605	filled | db:142
1	-14.0605	length(c) | negationslash={}then
1	-14.0605	pattern[j+1] | (3a)
1	-14.0605	children], | aux,
1	-14.0605	night, | lovely
1	-14.0605	nivre’s | swap-based
1	-14.0605	| q(x))
1	-14.0605	x: | �:f+|
1	-14.0605	“cross-lingual | disambiguation”
1	-14.0605	104,000 | 750,000
1	-14.0605	p(rnull6 | r3null
1	-14.0605	und | sein.
1	-14.0605	 | s1or
1	-14.0605	0.7112 | 0.1650
1	-14.0605	cm, | verb‟s
1	-14.0605	0.8) | ranges
1	-14.0605	myroslava | johanna
1	-14.0605	ptb§23 | 92.19
1	-14.0605	int'l | user
1	-14.0605	2001 | sein’
1	-14.0605	why | knowledge?
1	-14.0605	promised | ceasefire
1	-14.0605	(i−6 | s,e
1	-14.0605	for, | towards,
1	-14.0605	lolss | bahaha
1	-14.0605	θik | θbik
1	-14.0605	null(nullpii | nullnull(nullnullnullpinullnull)
1	-14.0605	r=relat, | p=pron,
1	-14.0605	(is-101205) | (n0014110417).
1	-14.0605	okinawa, | (chinese,
1	-14.0605	word-f1 | 0.5937
1	-14.0605	p(crt |crtnull1,opt)
1	-14.0605	hatzivassiloglou | (pang,
1	-14.0605	typographic | cutie
1	-14.0605	benv | baiv
1	-14.0605	washer=person, | theme:instrument)
1	-14.0605	s(s) | fa(p)
1	-14.0605	s51s48s48s48 | s55s48s48s48
1	-14.0605	houston, | jacinto,
1	-14.0605	18.56 | 15.73
1	-14.0605	remarked | dinu
1	-14.0605	sh, | non-linearly
1	-14.0605	correct | conte
1	-14.0605	sind | rennr¨ader
1	-14.0605	parse(s) | r∈t
1	-14.0605	71.16 | 67.26
1	-14.0605	mero | attri
1	-14.0605	scoring | scoretype
1	-14.0605	content: | 19751988
1	-14.0605	邦交 | 国家
1	-14.0605	α(yprime,sprime|y,s,w)=min „
1	-14.0605	1­gram | 2­gram
1	-14.0605	u10 | u20
1	-14.0605	(latbleu), | (nbexpbleu),
1	-14.0605	(socher | socher
1	-14.0605	parse | chinese-speaking
1	-14.0605	े | ँ
1	-14.0605	fl | almdars
1	-14.0605	highland | cananea
1	-14.0605	rappoport | revrank
1	-14.0605	considerable | (shaw
1	-14.0605	sl, | tt,
1	-14.0605	tackle | re-parametrization
1	-14.0605	(aarts | tono,
1	-14.0605	extensions | nullnullnullnullnull(null,
1	-14.0605	h,w | ∇θrw(θ).
1	-14.0605	d39d39 | d108
1	-14.0605	null | talks]
1	-14.0605	top-scoring | updating
1	-14.0605	dfg | aliz-e
1	-14.0605	j08a | j08b
1	-14.0605	(3b) | “transitive
1	-14.0605	21 | mnfmean
1	-14.0605	0.7012 | 0.4353
1	-14.0605	ponte, | moshe
1	-14.0605	ax | l(θ;
1	-14.0605	maltreated, | arrest,
1	-14.0605	(ccg; | longstanding
1	-14.0605	documented | http:/download.oracle.com/javase/6/docs/api/java/util/regex/
1	-14.0605	p(cm|v,c)
1	-14.0605	ˆvpt(vectorr,sj) | (summationtext
1	-14.0605	menu-based | (hjelm
1	-14.0605	parameters: | spanmap;
1	-14.0605	<event | 3.000
1	-14.0605	52.05 | 41.97
1	-14.0605	:={at(r,n2)} | at(r,n2)
1	-14.0605	confusionset | commonlycited
1	-14.0605	explode, | destroy}
1	-14.0605	callison-burch | pivoting
1	-14.0605	lin | 90.97
1	-14.0605	eric | 4gj,
1	-14.0605	kill | mockingbird
1	-14.0605	sub-topic | super-topic
1	-14.0605	kim | chan’s
1	-14.0605	walked) | hond
1	-14.0605	off:s, | off:s}
1	-14.0605	(forall | x)))
1	-14.0605	lm | mono-news-c
1	-14.0605	(culy | riehemann,
1	-14.0605	manager] | [mr.
1	-14.0605	(hal | ri)
1	-14.0605	basic | [u*
1	-14.0605	epsilon1 | 28.7±1.1
1	-14.0605	ps(α|η) | pa(none|η)
1	-14.0605	nullpresident’s | in,electionnull
1	-14.0605	sys1 | no_
1	-14.0605	rounding | approximation
1	-14.0605	(kenny | jan
1	-14.0605	303 | 136257
1	-14.0605	blue, | (black,
1	-14.0605	bannard, | callison-burch.
1	-14.0605	nilsson, | eryigit,
1	-14.0605	(10.0%) | (20.8%)
1	-14.0605	forbes | murray.
1	-14.0605	(generalized) | 257.8
1	-14.0605	“ | tςml
1	-14.0605	vbn | 1486]:[date]
1	-14.0605	lk | traink(ts).
1	-14.0605	3(a) | induced
1	-14.0605	{as} | f,e
1	-14.0605	“said” | “remains”
1	-14.0605	83.5 | cross
1	-14.0605	verbalpresent: | futureuse:3
1	-14.0605	li | zhongguancun
1	-14.0605	topicindex | learning2link
1	-14.0605	trip | remarkable
1	-14.0605	(sinha | diab,
1	-14.0605	collocation | tango
1	-14.0605	pine | lieven
1	-14.0605	cost=extendedlevenshtein(userusage,pattern) | mineditedcost=cost;
1	-14.0605	address | (interlingual
1	-14.0605	di | livememories,
1	-14.0605	gesandte’ | indeclinable
1	-14.0605	(eng) | (monolingual)
1	-14.0605	lbl | sbv,
1	-14.0605	naval | n0001410wx20278.
1	-14.0605	92.13 | 86.29
1	-14.0605	s3h | fcc)
1	-14.0605	→h) | notarrowrighth)
1	-14.0605	nnp | nnps.
1	-14.0605	institut | recherche
1	-14.0605	“big | ball”
1	-14.0605	coref | [t5,
1	-14.0605	emotions: | blue,
1	-14.0605	e2d-1ref-tst | e2p-1ref-tst
1	-14.0605	dii | tor
1	-14.0605	mood | (poms)
1	-14.0605	[a | value]:
1	-14.0605	-9 | -13
1	-14.0605	cutwidth | 2null4
1	-14.0605	‘yes’ | ‘no’
1	-14.0605	fifty | 1750-1800).
1	-14.0605	traditionaldependency | matsumoto,2003;
1	-14.0605	martins∗† | {afm,nasmith}@cs.cmu.edu,
1	-14.0605	0.369 | 0.372
1	-14.0605	oil | honey
1	-14.0605	surprisingly, | substationally
1	-14.0605	xa | g*
1	-14.0605	personalized | (ips),
1	-14.0605	er | kraft
1	-14.0605	firsttraina | beforecoreferenceresolution.inthisworkweadopt
1	-14.0605	r | run-score-3
1	-14.0605	verbocean | (chklovski
1	-14.0605	rkun | ntza
1	-14.0605	+0.4 | 84.647
1	-14.0605	aker | gaizauskas
1	-14.0605	db:175 | blues
1	-14.0605	“killed” | “assassinated”
1	-14.0605	consensus | regularizing
1	-14.0605	go, | bhayani,
1	-14.0605	bibliome | nomenclatures
1	-14.0605	anaphora | coli
1	-14.0605	transregional | cognition’
1	-14.0605	21758160 | 1758160.
1	-14.0605	b# | l#
1	-14.0605	ppl100% | 132.7
1	-14.0605	seki, | kando.
1	-14.0605	p(e|f)·l(e,eprime)
1	-14.0605	theh | esco
1	-14.0605	set-valued | (masi)
1	-14.0605	jia | casey
1	-14.0605	limbs | visible.
1	-14.0605	counters | 4.2.1.
1	-14.0605	53.3% | 55.3%
1	-14.0605	onlinea | timesjudged,for
1	-14.0605	71.66 | 62.70
1	-14.0605	yor, | teh,
1	-14.0605	barerel | absol
1	-14.0605	ceafm | 0.6581
1	-14.0605	60.5 | ∗
1	-14.0605	bayyari, | rivera
1	-14.0605	of: | type:⟨person⟩
1	-14.0605	filet, | salmon,
1	-14.0605	andcj | andj.
1	-14.0605	<segment | id="c1">
1	-14.0605	“{ | “[
1	-14.0605	bama | prefixes/proclitics
1	-14.0605	p(c|w)
1	-14.0605	online | lexical-stylistic
1	-14.0605	(slu) | recogniser.
1	-14.0605	o; | p<0.01).
1	-14.0605	“christopher” | named(l0,
1	-14.0605	with: | digits;
1	-14.0605	maximizes p(s|t):
1	-14.0605	1"234563) | '/*<)
1	-14.0605	seattle | 0.04−0.09
1	-14.0605	extension | kasprzik
1	-14.0605	[β0,γ0] | [β1,γ1]...[βk,γk]
1	-14.0605	eventuality | eventuality
1	-14.0605	programmatically | revision
1	-14.0605	pre(x) | post(x)
1	-14.0605	97.85 | 97.74
1	-14.0605	(poesio | (ginzburg,
1	-14.0605	ye | fnull(y)
1	-14.0605	ry | algo
1	-14.0605	des | cnrs,
1	-14.0605	1:30 | delaney
1	-14.0605	roundtrips | votecount
1	-14.0605	do: | node:number;
1	-14.0605	0.262 | 0.107
1	-14.0605	while, | contrast,
1	-14.0605	entailment | puv
1	-14.0605	(60%) | (40%)
1	-14.0605	| xi,yi)
1	-14.0605	implementation | lattices=
1	-14.0605	(arg0 | arg5)
1	-14.0605	75.3 | −mi
1	-14.0605	sub-tree | hri
1	-14.0605	ins | lmweight
1	-14.0605	‘2’, | ‘3’,
1	-14.0605	% | 22.85
1	-14.0605	arguer | conclusion
1	-14.0605	frame-weighted | un-frame-weighted
1	-14.0605	piecewise | relaxation-characterization
1	-14.0605	20.8 | 20.2
1	-14.0605	mbdp-1 | ngs-u:
1	-14.0605	x) | (p’
1	-14.0605	most-truexb | cj.
1	-14.0605	(argc | arga)
1	-14.0605	/* | */
1	-14.0605	turmo, | alicia
1	-14.0605	(epsilon1 | 0.1,δ
1	-14.0605	non-duplicate | 1,911
1	-14.0605	“kashmir” | himalaya
1	-14.0605	fair | scpf
1	-14.0605	23.42 | 25.47
1	-14.0605	albums | singles,
1	-14.0605	¸@ | fik
1	-14.0605	(rs2) | f(jiqiao)”.
1	-14.0605	epi | site-arguments
1	-14.0605	answer, | inform,
1	-14.0605	t←0 | repeat
1	-14.0605	(courtesy | weber)
1	-14.0605	nist | 52.32
1	-14.0605	d(irect) | c(omplex).
1	-14.0605	lee | knutsson,
1	-14.0605	|ts ∩ta||t
1	-14.0605	normal(µ | categorical(pi)
1	-14.0605	n=5 | n=10
1	-14.0605	4(d) | 4(e)
1	-14.0605	fla« | fla«’)
1	-14.0605	babcock | bailey
1	-14.0605	tutor: | aray
1	-14.0605	raw | graphsof
1	-14.0605	1.41 | 1.48
1	-14.0605	pitch | wet.
1	-14.0605	wer | kn-rnn-lim
1	-14.0605	(harrington, | pulman,
1	-14.0605	electronic | aids,
1	-14.0605	three different | metricswere
1	-14.0605	(random | init.)
1	-14.0605	owe | ing
1	-14.0605	“goal,” | “reaction”
1	-14.0605	u(i;t) | (i;t)
1	-14.0605	jumps, | (-5
1	-14.0605	eleni | livio
1	-14.0605	giorgio | padua
1	-14.0605	,,, | g),(
1	-14.0605	twenty | radiozurnal.
1	-14.0605	16.19 | 14.66
1	-14.0605	91.35 | 62.35
1	-14.0605	(good | fair)
1	-14.0605	(fillmore, | ‘linking
1	-14.0605	(nicolás, | henrik
1	-14.0605	e3 | anaphylaxis
1	-14.0605	t2), | teamscoringall(t1,
1	-14.0605	e | universit`a
1	-14.0605	“ | 将
1	-14.0605	#systems | #system-outputs-assessed
1	-14.0605	fabio | rome
1	-14.0605	arabic | ﻞﻌﻟو
1	-14.0605	φ(x,y) | “box
1	-14.0605	eg882>k | electronics
1	-14.0605	69.2 | 62.1
1	-14.0605	a88a88a88a88 | eval
1	-14.0605	summationtextn | pi(i)
1	-14.0605	/ | tax
1	-14.0605	deleteall | finalresults.
1	-14.0605	(1992) | (christodoulopoulos
1	-14.0605	sentence-basedevents | statisticsof
1	-14.0605	˜pn(wi|wi−1i−n+1) | λn(φn)
1	-14.0605	iinokataba | takaataba
1	-14.0605	0.6938 | 0.5552
1	-14.0605	subtrees | rsn
1	-14.0605	acknowledgment | “ok.”)
1	-14.0605	neq | ngt
1	-14.0605	lewis | influenza,
1	-14.0605	ece | eventivity
1	-14.0605	axis) | axis).
1	-14.0605	factorize | smaller
1	-14.0605	20.70 | 16.07
1	-14.0605	1/ | 0.85714
1	-14.0605	lambek | (lambek,
1	-14.0605	80000 | 120000
1	-14.0605	jdyd¯h | fsfsn
1	-14.0605	eve | naomi
1	-14.0605	sys-mpqa | sys-frame
1	-14.0605	academic | (daad).
1	-14.0605	verwijt | ‘reproach’
1	-14.0605	2557 | 37819
1	-14.0605	ministry, | know2
1	-14.0605	empty | “lovely”,
1	-14.0605	k= | 10
1	-14.0605	get: | p(e)·
1	-14.0605	dg(5): | df(8):
1	-14.0605	(neutral | dropped)
1	-14.0605	plex(y|x,d)
1	-14.0605	fibrinogen | flexor
1	-14.0605	r−1 | pr−1
1	-14.0605	b’s | commitb(elabb(pi1,pi2),dn
1	-14.0605	uh, | denverbracehtipupleft
1	-14.0605	m∑ | iw(qi)
1	-14.0605	ieee | (icsc),
1	-14.0605	gigaword | afp,
1	-14.0605	≤1 | ananim
1	-14.0605	• | algu´em)
1	-14.0605	nonel, | noner,
1	-14.0605	irstlm | cettolo,
1	-14.0605	category | sub-classes:
1	-14.0605	divivb1 | minc
1	-14.0605	351 | 83,713
1	-14.0605	increment | histf−1,lscript−j[k]
1	-14.0605	brian | beaverton,
1	-14.0605	police | agents.]
1	-14.0605	pertaining | history,
1	-14.0605	s2wc, | s0cs1c,
1	-14.0605	∼ mult(·|
1	-14.0605	dead | wounded.
1	-14.0605	s→a a0,|e|,0,|f|,
1	-14.0605	cool | safe
1	-14.0605	ution | ,
1	-14.0605	estimated | p(t|d)
1	-14.0605	performance-basedrule | increasethe
1	-14.0605	si, | kmax
1	-14.0605	73.3 | 74.8
1	-14.0605	bbc | guardian
1	-14.0605	0.3588 | 0.6154
1	-14.0605	σ0, | σ+
1	-14.0605	p% | e%
1	-14.0605	button] | dg(7):
1	-14.0605	deterministic, | stochastically
1	-14.0605	(hoerl | kennard,
1	-14.0605	´ecole | (qu´ebec),
1	-14.0605	german | (english:
1	-14.0605	number” | “ok.
1	-14.0605	(x-y) | hayvan
1	-14.0605	qsbp | qsbp(idf)
1	-14.0605	infants | (saffran
1	-14.0605	lrec | towards
1	-14.0605	“verhandlen | “wollen
1	-14.0605	(reis | judd,
1	-14.0605	(granularity) | mereotopologically
1	-14.0605	bl-nw | dv-nw
1	-14.0605	91.92 | 86.09
1	-14.0605	ldc2003e07, | ldc2005e83,
1	-14.0605	nuclei | liquids
1	-14.0605	___ | ___.
1	-14.0605	authoring | branches.
1	-14.0605	confinement | japanese-language
1	-14.0605	0.75) | -0.5)
1	-14.0605	[vp08]: | paşca.
1	-14.0605	pn(wi|φn)
1	-14.0605	personaltone | entrypov
1	-14.0605	(60873155), | (2008b03)
1	-14.0605	nullwe, | +null;
1	-14.0605	cues | vulgar
1	-14.0605	medline | acronym.
1	-14.0605	s-missing | p-correct
1	-14.0605	charniak | shimony
1	-14.0605	faruqui | kharagpur,
1	-14.0605	seg-lattice | fst-lattice
1	-14.0605	p(ti|si).
1	-14.0605	pervasive | neutralization
1	-14.0605	g1009 | g1005
1	-14.0605	heard | speaking.
1	-14.0605	bucket | buckets:
1	-14.0605	φ(wi−2wi−1) | corresponding
1	-14.0605	end="1210" | end="1228"
1	-14.0605	nwi|z+βsummationtext | nzi|di
1	-14.0605	s50s53 | s65
1	-14.0605	gave | 13.1%
1	-14.0605	fields | first-name,
1	-14.0605	economics | (fkz:
1	-14.0605	u(i) | u(i)−α(y∗(i)−1)
1	-14.0605	p(t,j|s) | p0(t,j|s)z(s)
1	-14.0605	ifsandv | (adj(s,v)
1	-14.0605	floyd-warshall | (floyd,
1	-14.0605	synsets | scov0<sα,
1	-14.0605	one-way | tests).
1	-14.0605	37.04% | 47.48%
1	-14.0605	—a | æ
1	-14.0605	terminate | yield.
1	-14.0605	yˇu | sh¯al´ong
1	-14.0605	sup | uns
1	-14.0605	diri | 250gd
1	-14.0605	almor | adam
1	-14.0605	1-doc | 5-docs
1	-14.0605	martinez | piater
1	-14.0605	10k | 54.77
1	-14.0605	monash | melbourne,
1	-14.0605	tst3 | tst4
1	-14.0605	=1 | =0
1	-14.0605	kenny | roastbeef:
1	-14.0605	tas, | lapa
1	-14.0605	‘date | birth’
1	-14.0605	iis-0915187 | w911nf-10-1-0533.
1	-14.0605	β∈qn | l1(t1),...,ln(tn)
1	-14.0605	resolved | matching.we
1	-14.0605	‘opinion’ | hout
1	-14.0605	→numrev([[u1]]c) | numtrue([[u1]]c)
1	-14.0605	prevalence | prevd(c)
1	-14.0605	“iphone” | gaga”
1	-14.0605	x2: | xx1:f+|
1	-14.0605	androutsopoulos∗+ | “athena”,
1	-14.0605	flemish | (northern)
1	-14.0605	plan | sd-team
1	-14.0605	aspect-evaluation | aspect-of
1	-14.0605	range(r3) | range(r2)
1	-14.0605	president” | “complimenting
1	-14.0605	padó | pado@cl.uni-heidelberg.de
1	-14.0605	(figure | 3(b))
1	-14.0605	p(e1...et|τ)p(τ)
1	-14.0605	11: | >g(ˆei)
1	-14.0605	tj | tj−1
1	-14.0605	(53% | (11%).
1	-14.0605	modelfor | predicatethe
1	-14.0605	converter | -conll2007
1	-14.0605	indigo, | fp6
1	-14.0605	wtφt(k,lscript) | m:k<m<lscript
1	-14.0605	75.10 | 66.79
1	-14.0605	collect | (amt).
1	-14.0605	a1-a2 | a1-a3
1	-14.0605	0.5018 | 0.3450†
1	-14.0605	switchboard | ext.
1	-14.0605	long | zaiqing
1	-14.0605	#medical | targetstockprice1
1	-14.0605	european | (clara),
1	-14.0605	¯p | ¯h
1	-14.0605	[last | jeremi
1	-14.0605	7,011 | 70.65
1	-14.0605	αt: | (ei,t−¯ei,t)
1	-14.0605	xe | ∀t∈ti,
1	-14.0605	index | index(γ1)
1	-14.0605	monotrans2 | (haitian
1	-14.0605	sq1(x | tr1(x
1	-14.0605	+β | +vβ
1	-14.0605	bmc | decis
1	-14.0605	:teamingame(g, | teamscoringall(t2,
1	-14.0605	“weather” | “war”
1	-14.0605	sav | 2007*
1	-14.0605	$ %# | .'
1	-14.0605	ˇcetl | ˇc´ıst
1	-14.0605	extractive | eij,
1	-14.0605	(nl), | (pt),
1	-14.0605	executing | appropriateness
1	-14.0605	received | weights,
1	-14.0605	(bicici | kozat,
1	-14.0605	(µ)parenrightbig | (ai,j,r−¯ai,j,r)
1	-14.0605	student: | bugy]
1	-14.0605	palmer | sim(c1,c2)=
1	-14.0605	to: | sketch[k,hk(x)],
1	-14.0605	ds-cascade | 0.349
1	-14.0605	illusion | complement/
1	-14.0605	afii62825afii62820 | ؟
1	-14.0605	bsprime,t,uprime,v, | (es..sprime/fu..uprime
1	-14.0605	ground | goal;
1	-14.0605	penn, | richter
1	-14.0605	stepsizes | (λ,µ))
1	-14.0605	xiang | wen-pin
1	-14.0605	plural | haj.¸a£
1	-14.0605	(sejnowski | rosenberg,
1	-14.0605	‘but’ | ‘yet’
1	-14.0605	2-1-2 | chiyoda-ku,
1	-14.0605	˜p(h)summationdisplay | ˜p(w,h)∇θaw(h;θ)−summationdisplay
1	-14.0605	morphemes | turkish
1	-14.0605	modified | ˆp(f|e)
1	-14.0605	bit | 1101111101100
1	-14.0605	25.9 | (358)
1	-14.0605	our | cangive52.9%recall,80.77%precisionand63.93%
1	-14.0605	based | auc.
1	-14.0605	county* | special*
1	-14.0605	post/thread-level | (“∗”
1	-14.0605	negative: | positive:
1	-14.0605	dried | cause(agent,
1	-14.0605	reliance | bottleneck
1	-14.0605	rth(x) | sth(x)
1	-14.0605	ξ(good) | ξ(bad)
1	-14.0605	cos | 0.738,
1	-14.0605	fragments | markup),
1	-14.0605	hindi, | tamil
1	-14.0605	incorporated | cache:
1	-14.0605	43.5 | 43.6
1	-14.0605	(id) | pyysalo
1	-14.0605	reductions: | =cprimeηιrrbracket·eθg-rl∗,d(cη
1	-14.0605	acc | dom-clusti
1	-14.0605	uszkoreit, | popat
1	-14.0605	g3(%) | kg882>b
1	-14.0605	forbidding | liquid
1	-14.0605	r4 | →o
1	-14.0605	^c | {c,
1	-14.0605	“virgin” | “decorous”
1	-14.0605	(mt06) | (mt08)
1	-14.0605	lagrangian | l(y,λ,γ,u,v)
1	-14.0605	x), | sati(ψ),
1	-14.0605	hand-curated | nadel
1	-14.0605	g(eprime) | eh
1	-14.0605	downcased, | ’ll
1	-14.0605	bojar | n´amˇest´ı
1	-14.0605	p(x,y)p(x)p(y) | variablesx
1	-14.0605	〈u(r)〉r∈r∈r|r|,
1	-14.0605	understood | sentence”,
1	-14.0605	swiss | (snsf)
1	-14.0605	old, | med),
1	-14.0605	summationdisplay γ0=([¢],β0),...,(σ|h1,βi),
1	-14.0605	17.71 | 22.94
1	-14.0605	noreg | noreg+
1	-14.0605	(levelt | kelter,
1	-14.0605	pmipp | ecdpp
1	-14.0605	metoclopramide | chemoattractant
1	-14.0605	connectives | /mnd¯
1	-14.0605	α1, | α2]
1	-14.0605	negri, | manuela
1	-14.0605	center | pittsburgh.
1	-14.0605	[arg1 | [arg0
1	-14.0605	rom(x,bos) | (boston,[f
1	-14.0605	metaphors | dubs
1	-14.0605	pbt | hpbt
1	-14.0605	“after | ice”,
1	-14.0605	cocitation | tile
1	-14.0605	epi | 53.33
1	-14.0605	marathi-health | 1750
1	-14.0605	aligning | (s,t),
1	-14.0605	peso | argentine
1	-14.0605	tank | “f-14
1	-14.0605	p(z|d1,...,d4)
1	-14.0605	fixit | (hart
1	-14.0605	cd | s(c,d,t),
1	-14.0605	assurance | framefiles.
1	-14.0605	m=|tt|;
1	-14.0605	0.482 | 0.232
1	-14.0605	(query), | (answers)>.
1	-14.0605	pittsburgh, | akw13@pitt.edu
1	-14.0605	(-0.04 | 0.07)
1	-14.0605	0.125 | 8.000
1	-14.0605	(stymne | ahrenberg,
1	-14.0605	flight | (lo)
1	-14.0605	yarowsky | ngai
1	-14.0605	gains | (1.6
1	-14.0605	cn | r1···rn,
1	-14.0605	total | 100.00%
1	-14.0605	user: | (confirm:from(fl1,bos))
1	-14.0605	eiichiro | mastar
1	-14.0605	taylor, | webb,
1	-14.0605	position | medial
1	-14.0605	incremental | (aist
1	-14.0605	q6: | q20
1	-14.0605	expressives, | representatives.
1	-14.0605	gesture | gesturer’s
1	-14.0605	z | zl(x)
0.333333	-14.0605	lee, | lee, | g.
0.333333	-14.0605	squaresmallsolid | [p@2null1null1
0.259259	-14.0606	are | examples,
0.219512	-14.0606	thus | but
0.12605	-14.0607	how | different
0.219512	-14.0609	first | within
0.0365359	-14.061	which | with
0.285714	-14.061	on | collections
0.3	-14.061	identify | noun
0.164384	-14.0611	the | picked
0.192308	-14.0612	large | parallel
0.4	-14.0612	definition | ,imp
0.169231	-14.0612	phrase | probabilities
0.0336374	-14.0612	and | into
0.444444	-14.0612	size | rare
0.5	-14.0613	n-grams | occur
0.5	-14.0613	4.1 | 3.9
0.5	-14.0613	liu | chiang,
0.5	-14.0613	dar | (get
0.5	-14.0613	p@2null1null1 | np@2null1null2
0.5	-14.0613	engagement | dialog
0.5	-14.0613	verb, | verb,
0.5	-14.0613	ner | budget
0.109091	-14.0614	is | product
0.110429	-14.0615	same | each
0.214286	-14.0615	another | than
0.666667	-14.0616	qe | qm.
0.666667	-14.0616	figure | ground,
0.666667	-14.0616	morphology, | derivational
0.666667	-14.0616	scale | impractical
0.666667	-14.0616	year | woman
0.666667	-14.0616	b., | j.
0.666667	-14.0616	en-sp | inventory
0.666667	-14.0616	7 | 118
0.666667	-14.0616	(the | section.)
0.666667	-14.0616	things: | time.
0.666667	-14.0616	charts, | media
0.666667	-14.0616	ˆy | ˆy
0.666667	-14.0616	ease | use,
0.666667	-14.0616	jordan” | refer
0.666667	-14.0616	subgradient | solution
0.666667	-14.0616	exactly | semantics.
0.666667	-14.0616	department | against
0.666667	-14.0616	applications, | question
0.666667	-14.0616	pair, | classes
0.666667	-14.0616	voice | recorded
0.666667	-14.0616	wsj | accidents
0.666667	-14.0616	insertions | substitutions.
0.666667	-14.0616	gap | patern
0.666667	-14.0616	frequency | applied;
0.666667	-14.0616	(aue | gamon,
0.666667	-14.0616	shutova | sense-disambiguated
0.666667	-14.0616	increase | increase.
0.666667	-14.0616	acquisition | prevention
0.666667	-14.0616	iso | enable
0.666667	-14.0616	estimation | unanswered
0.666667	-14.0616	hiersum | kl
0.666667	-14.0616	cues | task-oriented
0.666667	-14.0616	learns | eliminate
0.666667	-14.0616	ieee | analysis
0.666667	-14.0616	practical | lexicons,
0.666667	-14.0616	a 140 | weight
0.666667	-14.0616	grammar: | grammar
0.666667	-14.0616	attitude | theory
0.666667	-14.0616	parent | grandparent
0.666667	-14.0616	detect | pattern:
0.666667	-14.0616	absolute | years.
0.666667	-14.0616	dialogic | posts
0.666667	-14.0616	comprising | specialized
0.666667	-14.0616	non-propagation | crf-d.
0.666667	-14.0616	sie | die
0.666667	-14.0616	ir-status | ir-response
0.666667	-14.0616	(mturk), | online
0.666667	-14.0616	course | adaptation
0.666667	-14.0616	direction | orientation
0.666667	-14.0616	chen, | mooney
0.666667	-14.0616	j., | ney
0.666667	-14.0616	� | [
0.666667	-14.0616	phrases | etc.;
0.666667	-14.0616	〉, | 〈〈q
0.666667	-14.0616	co-training | (blum
0.666667	-14.0616	p5r1 | p5r3
0.666667	-14.0616	identical | stem
0.666667	-14.0616	family | genes
0.666667	-14.0616	response | answers.
0.666667	-14.0616	p(w |
0.666667	-14.0616	estimation | domains.
0.666667	-14.0616	con | ve
0.666667	-14.0616	89.2 | 89.8
0.666667	-14.0616	(klein | berg-kirkpatrick
0.666667	-14.0616	uv | p.
0.666667	-14.0616	(best | -
0.666667	-14.0616	proportion | americans
0.666667	-14.0616	authors | paper).
0.666667	-14.0616	web | 94%
0.666667	-14.0616	r., | 2007.
0.666667	-14.0616	interspeech | pp.
0.666667	-14.0616	genres | 16th
0.666667	-14.0616	chosen | policy.
0.666667	-14.0616	sessions | session
0.269231	-14.0616	obtained | pairs
0.177419	-14.0616	weighted | on
0.0528571	-14.0617	of | defined
0.219512	-14.0617	linguistic | word
0.0721925	-14.0618	to | .
0.149425	-14.0619	to | values.
0.0492005	-14.0621	null | is
0.114094	-14.0621	are | natural
0.5	-14.0621	determiner | adverb
0.5	-14.0621	consumers | reviews
0.5	-14.0621	xiao, | &
0.0521262	-14.0622	syntactic | to
0.119403	-14.0623	is | independent
0.133333	-14.0623	1 | but
0.162162	-14.0624	domain | can
0.0487211	-14.0625	word | from
0.315789	-14.0626	that | robustness
0.25	-14.0626	τ | to
0.0788644	-14.0626	the | modified
0.666667	-14.0627	prevent | decay
0.666667	-14.0627	reduces | perplexity
0.666667	-14.0627	aline | rio
0.666667	-14.0627	cj | what
0.666667	-14.0627	biomedical | pa.,
0.666667	-14.0627	splits | spans
0.666667	-14.0627	n×n | aims
0.666667	-14.0627	conversion | ontonotes
0.666667	-14.0627	introductory | university
0.666667	-14.0627	anselmo | and
0.666667	-14.0627	question | grammatically
0.666667	-14.0627	10% | 40%.
0.666667	-14.0627	candidate | mi.
0.666667	-14.0627	geo:cityunionsqgeo:state) |
0.666667	-14.0627	mln | first-order
0.666667	-14.0627	choices | affect
0.666667	-14.0627	human | siding
0.666667	-14.0627	m45 | experimental
0.666667	-14.0627	authoring | conversational
0.666667	-14.0627	prosodic | n-best
0.666667	-14.0627	−α | +
0.666667	-14.0627	player | berkeley
0.666667	-14.0627	70k | proposed
0.666667	-14.0627	textrunner | (banko
0.666667	-14.0627	simplest | organization
0.666667	-14.0627	initialize | next
0.666667	-14.0627	national | automation,
0.666667	-14.0627	calculate | −
0.666667	-14.0627	absolute, | relative
0.666667	-14.0627	directly | document,
0.666667	-14.0627	tac | 2010.
0.666667	-14.0627	car | car
0.666667	-14.0627	4-gram | lms
0.666667	-14.0627	(7) | append
0.666667	-14.0627	[[n | v30]]
0.666667	-14.0627	predicting | lu
0.666667	-14.0627	means: | phrase
0.666667	-14.0627	marking | grammatical
0.180328	-14.0627	useful | which
0.0310078	-14.0627	and | other
0.111111	-14.0627	6 | 1
0.444444	-14.0627	verbs | senses
0.444444	-14.0627	main | annotations
0.192308	-14.0628	approach, | from
0.285714	-14.0629	classifiers | use
0.0691358	-14.0629	the | exact
0.5	-14.0629	technical | technical
0.5	-14.0629	heuristic | axiom
0.5	-14.0629	m-step | θ
0.5	-14.0629	bravery | his
0.222222	-14.0631	using | wikipedia
0.228571	-14.0631	methods | accuracy
0.333333	-14.0632	many | data,
0.333333	-14.0632	degree | vertex
0.0477855	-14.0632	this | all
0.0665138	-14.0632	is | high
0.28	-14.0633	demonstrated | our
0.214286	-14.0633	from | distributional
0.16	-14.0634	designed | be
0.118519	-14.0634	spatial | the
0.357143	-14.0634	assumes | can
0.357143	-14.0634	score | beam
0.444444	-14.0634	.25 | .29
0.6	-14.0635	combination | metrics.
0.0474576	-14.0636	and | any
0.071066	-14.0637	in | sentences.
0.5	-14.0638	medical | patient
0.5	-14.0638	verb | “she
0.5	-14.0638	chunks | constituents
0.666667	-14.0639	string, | types,
0.666667	-14.0639	others | not,
0.666667	-14.0639	raters | prefer
0.666667	-14.0639	reactor | .
0.666667	-14.0639	front | back
0.666667	-14.0639	examined | curves
0.666667	-14.0639	buckets | empty
0.666667	-14.0639	mehrnoosh | university
0.666667	-14.0639	(a) | features.
0.666667	-14.0639	(the | h)
0.666667	-14.0639	maximises | expected
0.666667	-14.0639	types | tokens)
0.666667	-14.0639	diversify | joint
0.666667	-14.0639	asertion | absent,
0.666667	-14.0639	muc | none
0.666667	-14.0639	vocabulary | experience
0.666667	-14.0639	conversation | talk
0.666667	-14.0639	cruys | cambridge
0.666667	-14.0639	network | hitting
0.666667	-14.0639	advances | kernel
0.666667	-14.0639	gain | neighbors
0.666667	-14.0639	patent | phrases.
0.666667	-14.0639	decoder | costs.
0.666667	-14.0639	dp1 | vp
0.666667	-14.0639	calculations | performed
0.666667	-14.0639	fuzzy | string-to-tree
0.0821918	-14.0639	to | information.
0.228571	-14.0639	to | young
0.0666667	-14.064	likely | be
0.0778816	-14.0642	our | better
0.185185	-14.0642	performance | higher
0.384615	-14.0643	they | closely
0.180328	-14.0645	n-gram | models
0.5	-14.0646	pmi | measure
0.5	-14.0646	transitions | indicate
0.285714	-14.0648	make | data.
0.0929204	-14.0648	corresponding | is
0.14	-14.0648	this | parameters
0.2	-14.0648	first | following
0.333333	-14.0649	list | topics
0.444444	-14.0649	help | help
0.192308	-14.065	correlation | two
0.666667	-14.065	faster | times.
0.666667	-14.065	transliteration | origins
0.666667	-14.065	inter-coder | during
0.666667	-14.065	rule-based | triggering
0.666667	-14.065	eurotrans | uedin
0.666667	-14.065	closely | commercial
0.666667	-14.065	entities | facts.
0.666667	-14.065	pages | sweden,
0.666667	-14.065	responsiveness | across
0.666667	-14.065	0.458 | overall
0.666667	-14.065	nullnullnull | obj)
0.666667	-14.065	69 | 18
0.666667	-14.065	microblogging | short
0.666667	-14.065	probabilities | 1977)
0.666667	-14.065	hanneman | mellon
0.666667	-14.065	storing | memory
0.666667	-14.065	counts | brackets
0.666667	-14.065	1ifnp | corresponding
0.666667	-14.065	entire | zero.
0.666667	-14.065	in-domain | discovery
0.666667	-14.065	sent | 30
0.666667	-14.065	beyond | boundaries.
0.666667	-14.065	cliques | graph
0.3	-14.065	methods | 4
0.222222	-14.065	the | fund
0.0945946	-14.0651	we | categories
0.0985222	-14.0651	separately | the
0.0921053	-14.0651	absolute | the
0.315789	-14.0652	but | effective
0.235294	-14.0652	it | exactly
0.107143	-14.0653	derive | the
0.039557	-14.0654	also | that | the
0.5	-14.0654	ted | nist
0.5	-14.0654	margin | margin
0.25	-14.0655	class | label
0.6	-14.0655	classifying | types.
0.0606061	-14.0657	propose | to
0.214286	-14.0658	improvements | all
0.3	-14.066	candidate | appear
0.062753	-14.066	semantic | with
0.666667	-14.0661	span | span,
0.666667	-14.0661	word-pair | null
0.666667	-14.0661	groß | fahrrad
0.666667	-14.0661	valuable | who
0.666667	-14.0661	plate | sampling
0.666667	-14.0661	university | center.
0.666667	-14.0661	dico | menu
0.666667	-14.0661	building | stages
0.666667	-14.0661	softmax | unit
0.666667	-14.0661	lattice | cut
0.666667	-14.0661	33 | 26
0.666667	-14.0661	values | /aba/
0.666667	-14.0661	building | sky.
0.666667	-14.0661	especially | ones,
0.666667	-14.0661	engages | dialogue
0.666667	-14.0661	practical | reasoning
0.666667	-14.0661	proteins | np.
0.666667	-14.0661	including | carroll
0.666667	-14.0661	departamento | computaci´on
0.666667	-14.0661	locations | eight
0.666667	-14.0661	biased | decisions
0.666667	-14.0661	classify | citation
0.666667	-14.0661	students | experience
0.222222	-14.0661	effectively | by
0.269231	-14.0662	still | all
0.0547945	-14.0664	parse | and
0.0808081	-14.0666	would like | for
0.127119	-14.0666	inference | as
0.0268116	-14.0667	features | the | and
0.169231	-14.0668	type | used
0.188679	-14.0668	analysis | into
0.157895	-14.0668	supported | no.
0.0833333	-14.067	by | new
0.357143	-14.067	vignettes | be
0.5	-14.0671	then | fall
0.5	-14.0671	marked | incorrect.
0.666667	-14.0672	qualitative | types
0.666667	-14.0672	final | lig
0.666667	-14.0672	well-formed | noisy
0.666667	-14.0672	few | exist.
0.666667	-14.0672	university | press,
0.666667	-14.0672	viterbi | plaintext
0.666667	-14.0672	frequent | (frequency
0.666667	-14.0672	holds | lemma-discourse
0.666667	-14.0672	communication | situations,
0.666667	-14.0672	difficulty | resolving
0.666667	-14.0672	annotated | (about
0.666667	-14.0672	e | p(e).
0.666667	-14.0672	model, | utilized
0.666667	-14.0672	naming | synonyms
0.666667	-14.0672	preference | meeting
0.666667	-14.0672	lexicon | f−score
0.666667	-14.0672	52 | 25
0.666667	-14.0672	matthew | richard
0.666667	-14.0672	sds | output
0.666667	-14.0672	en | said
0.666667	-14.0672	predicted | held-out
0.666667	-14.0672	can | improve | coherence,
0.666667	-14.0672	interactions | proteins
0.666667	-14.0672	accomplish | required
0.666667	-14.0672	computes | ngram
0.0794702	-14.0673	score | by
0.0921053	-14.0674	by | similarity
0.384615	-14.0675	topics | collection
0.6	-14.0675	easy | short
0.142857	-14.0676	each | to its
0.357143	-14.0676	mean | average
0.0694789	-14.0677	we | bleu
0.222222	-14.0679	with | applications
0.104651	-14.0679	by | so
0.131579	-14.0679	two | dependency
0.16	-14.068	our | learned
0.269231	-14.068	the | consequence.
0.5	-14.068	reference | names
0.5	-14.068	agreement | gender,
0.062753	-14.068	sequence | a
0.138614	-14.0681	search | or
0.151163	-14.0681	estimated by | the
0.0649351	-14.0682	2004; | and
0.666667	-14.0683	output, | extract
0.666667	-14.0683	entirely | free
0.666667	-14.0683	previously | lang
0.666667	-14.0683	leads | problems,
0.666667	-14.0683	estimated | l-bfgs
0.666667	-14.0683	easily | arbitrarily
0.666667	-14.0683	[a | ⇒∗g
0.666667	-14.0683	requires | expectations
0.666667	-14.0683	pitch | speech.
0.666667	-14.0683	semiconductor | 10
0.666667	-14.0683	0.03 | 2
0.666667	-14.0683	synthetic | real
0.315789	-14.0684	metric | performance
0.157895	-14.0684	most | source
0.0980392	-14.0687	this | annotated
0.0444444	-14.0687	it is | to | to
0.214286	-14.0687	data | errors
0.147727	-14.0688	makes | as
0.14	-14.0688	to | marked
0.357143	-14.0688	latent | analysis.
0.5	-14.0688	instances | semantically
0.12605	-14.0689	pairs, | to
0.315789	-14.069	lee | al.
0.14	-14.0693	information | data
0.166667	-14.0694	into | new
0.666667	-14.0694	sentiments | events.
0.666667	-14.0694	fragment | cfg
0.666667	-14.0694	hao | web
0.666667	-14.0694	containing | adjective,
0.666667	-14.0694	n | input.
0.666667	-14.0694	gold | mt05
0.666667	-14.0694	wikipedia | formation.
0.666667	-14.0694	id | speaker
0.666667	-14.0694	step | skipped
0.666667	-14.0694	besides | accurate
0.666667	-14.0694	phrases | non-constituent
0.666667	-14.0694	by | unmusical
0.666667	-14.0694	large | bases
0.666667	-14.0694	state | left-to-right
0.444444	-14.0694	grammar | full
0.07	-14.0695	other | not
0.097561	-14.0695	can | new
0.384615	-14.0695	tasks | very
0.384615	-14.0695	get | both
0.5	-14.0696	cost | proofs
0.5	-14.0696	backoff | typically
0.28	-14.0698	have | employed
0.269231	-14.0698	at | left
0.2	-14.0698	this | english.
0.235294	-14.0699	method | significant
0.11039	-14.0699	our | related
0.315789	-14.07	scores | low
0.444444	-14.0701	area | word
0.209302	-14.0702	more | problem
0.222222	-14.0702	3.1 | features
0.123077	-14.0703	testing | to
0.164384	-14.0704	benefit | for
0.101604	-14.0704	and | improvements
0.6	-14.0705	times | post
0.5	-14.0705	matsumoto | nara
0.5	-14.0705	web, | has
0.285714	-14.0706	sygal | typed
0.666667	-14.0706	detecting | affect
0.666667	-14.0706	gazetteers | ratinov
0.666667	-14.0706	w, | τj
0.666667	-14.0706	cost | creating
0.666667	-14.0706	agreement/disagreement | broadcast
0.666667	-14.0706	transitions | specified
0.666667	-14.0706	preprocessed | 2005)
0.666667	-14.0706	weighted | encode
0.666667	-14.0706	spencerk | lexical
0.666667	-14.0706	document | vanderwende,
0.666667	-14.0706	paraphrases | 19
0.222222	-14.0706	details | found
0.269231	-14.0707	are: | word
0.0568562	-14.0707	as | its
0.157895	-14.0708	comparable to | of
0.146067	-14.0709	descriptions | to
0.285714	-14.071	target | (the
0.131579	-14.071	dimension | of
0.169231	-14.0711	with | slightly
0.0290226	-14.0712	semantic | and
0.166667	-14.0714	should | noted
0.269231	-14.0716	finally, | show
0.444444	-14.0716	work | establish
0.444444	-14.0716	recall | 100%
0.444444	-14.0716	weights | calculated
0.666667	-14.0717	anonymized | label
0.666667	-14.0717	restrictions | ideal
0.666667	-14.0717	g | bipartite
0.666667	-14.0717	model | politeness
0.666667	-14.0717	kbs | mbc
0.666667	-14.0717	constructing | axioms
0.666667	-14.0717	based | (denero
0.666667	-14.0717	b=100 | b=25
0.666667	-14.0717	develop | documents.
0.666667	-14.0717	(θ | d
0.0985222	-14.0717	by | similar
0.14	-14.0717	the | (stolcke,
0.114094	-14.0717	patterns | from
0.174603	-14.0717	our | require
0.181818	-14.0717	however, | best
0.4	-14.0719	there are | unique
0.4	-14.0719	0% | queries
0.116788	-14.0719	for | sets.
0.269231	-14.072	here | corpus
0.0901288	-14.072	the | 2004)
0.315789	-14.0721	test | 25
0.0519836	-14.0723	: | the
0.357143	-14.0724	(higashinaka | al.,
0.444444	-14.0724	data | talk
0.444444	-14.0724	scores | transition
0.2	-14.0725	requirement | to
0.666667	-14.0728	preprocessing | preprocessing
0.666667	-14.0728	111i | 
0.666667	-14.0728	semi-markov | (smdp)
0.666667	-14.0728	enciphered | key:
0.315789	-14.0732	contains | following
0.177419	-14.0732	have | various
0.119403	-14.0733	by | multiple
0.0427509	-14.0733	the | of | similarity
0.0487805	-14.0734	input | and
0.051144	-14.0736	compare | of
0.120301	-14.0737	y | n
0.133333	-14.0737	for | mining
0.0813559	-14.0738	of | recognition
0.134615	-14.0738	pairs | our
0.666667	-14.0739	(zhang | jiang
0.666667	-14.0739	involving | agreement.
0.666667	-14.0739	suggestions | improving
0.666667	-14.0739	showed | moderate
0.666667	-14.0739	after | (new
0.666667	-14.0739	positions | ci
0.444444	-14.0739	lexical | specify
0.101064	-14.0742	this section | and
0.269231	-14.0743	which | probabilistic
0.114865	-14.0743	in | relations.
0.333333	-14.0745	2009 | test
0.6	-14.0746	multiparty | dialog
0.6	-14.0746	αi | than
0.177419	-14.0747	learning | classification
0.0652174	-14.0749	work, | and
0.07	-14.0751	of | low
0.666667	-14.0751	introduce | coverage
0.666667	-14.0751	question | and/or
0.666667	-14.0751	accuracy | ptb.
0.666667	-14.0751	correctness | property
0.666667	-14.0751	(sorted | bleu_d).
0.666667	-14.0751	suppose | wrong
0.666667	-14.0751	18.45 | en-cs
0.666667	-14.0751	dani¨el | kok
0.666667	-14.0751	rapid | visual
0.222222	-14.0751	this | examples.
0.4	-14.0753	full | obtained
0.4	-14.0753	of | classes).
0.196078	-14.0753	was | specific
0.0298713	-14.0754	the | corpus | and
0.0585106	-14.0754	a | linguistic
0.5	-14.0755	example | john
0.6	-14.0756	*** | p
0.269231	-14.0757	(1) | weight
0.138614	-14.0758	column | a
0.4	-14.0759	them | information.
0.14	-14.0759	m | d
0.192308	-14.076	reviewers | comments
0.25	-14.076	sentences, | used
0.357143	-14.0761	in-house | and
0.06621	-14.0761	of | languages.
0.666667	-14.0762	or | te.
0.666667	-14.0762	query | trec-2003
0.666667	-14.0762	björne, | ginter,
0.666667	-14.0762	key | type,
0.666667	-14.0762	binarization | factor
0.666667	-14.0762	sequence | 1:1
0.666667	-14.0762	tree | (tsg)
0.5	-14.0764	sample | every
0.0282324	-14.0765	we | data
0.269231	-14.0766	nodes | structure
0.4	-14.0766	again, | than
0.071618	-14.0767	was | an
0.242424	-14.0768	confusion | are
0.112583	-14.0769	by | structure
0.0681265	-14.0769	the original | of
0.164179	-14.0769	as | get
0.113333	-14.0769	must | with
0.3	-14.0771	one | protein
0.3	-14.0771	token | +
0.5	-14.0772	case | 100%
0.666667	-14.0773	5.3 | 40
0.666667	-14.0773	proposed | spectral
0.666667	-14.0773	news | (kumaran
0.666667	-14.0773	theme | simple
0.666667	-14.0773	five | penalty.
0.666667	-14.0773	:={at(r,n1)} | :={u1,u2,u3,u4}
0.666667	-14.0773	us | ease
0.666667	-14.0773	approach | inflections.
0.666667	-14.0773	argument, | subtree
0.666667	-14.0773	although | coherence,
0.666667	-14.0773	savova, | masanz,
0.666667	-14.0773	step, | approximate
0.4	-14.0773	when | appended
0.185185	-14.0773	by | arguments
0.0467186	-14.0773	gold | the
0.357143	-14.0773	authors | used
0.106509	-14.0773	expression | in
0.0317378	-14.0773	for | using
0.269231	-14.0775	systems | extraction
0.121212	-14.0776	work | text
0.3	-14.0777	significantly | state-of-the-art
0.3	-14.0777	standard | development
0.177419	-14.0777	becomes | in
0.0456961	-14.0777	data | from
0.235294	-14.0779	research | extraction
0.0343373	-14.0779	even | the
0.0952381	-14.078	senses | in
0.269231	-14.078	this | summaries.
0.144444	-14.0782	all | up
0.115942	-14.0783	generally | a
0.0409408	-14.0783	in | some
0.103448	-14.0784	details | in
0.269231	-14.0784	national | national
0.666667	-14.0785	emotions | state.
0.666667	-14.0785	assignment, | considering
0.666667	-14.0785	timebank | reference
0.666667	-14.0785	u.s. | agreement
0.666667	-14.0785	strong | ones.
0.666667	-14.0785	enter | state
0.0893617	-14.0786	a | predict
0.315789	-14.0786	corpora | domain
0.315789	-14.0786	similar | query
0.25	-14.0786	while | parallel
0.155844	-14.0786	set | entity
0.137255	-14.0786	one | english
0.11039	-14.0787	are | sentiment
0.125	-14.0787	use | score
0.169231	-14.0788	this | improves
0.2	-14.0789	are | restricted
0.177419	-14.0789	used | translations
0.196078	-14.0789	statistical | data
0.0794702	-14.079	to | annotations
0.104046	-14.079	in | scotland, uk,
0.444444	-14.0791	parsing | inference
0.242424	-14.0793	expected | which
0.0856031	-14.0793	sentences | each
0.142857	-14.0795	used | while
0.666667	-14.0796	simulate | simulate
0.666667	-14.0796	mead | (radev
0.666667	-14.0796	log-linear | conditional
0.666667	-14.0796	(zaidan, | minimum
0.666667	-14.0796	uninterpretable | significantly
0.242424	-14.0797	in | don’t
0.0517483	-14.0798	the | morphological
0.0757576	-14.0798	we | grammar
0.0723861	-14.0798	it | use
0.5	-14.0798	will | examining
0.5	-14.0798	much | focuses
0.384615	-14.0799	smt | mt
0.4	-14.08	of | rg
0.4	-14.08	transfer | between
0.192308	-14.0803	and | da
0.357143	-14.0803	2007b; | al.,
0.0952381	-14.0804	we | treebank
0.155844	-14.0804	s | y
0.153846	-14.0804	structures | by
0.209302	-14.0805	from | 23
0.5	-14.0806	mwes | particular
0.666667	-14.0807	across | horn
0.666667	-14.0807	rule-based | msa
0.666667	-14.0807	form | derivation.
0.666667	-14.0807	expected | reverse
0.053973	-14.0807	in | experiments
0.444444	-14.0807	for | highest-scoring
0.444444	-14.0807	rule | subtree
0.0606061	-14.0808	model, | a
0.130841	-14.0809	state-of-the-art | in
0.0630081	-14.0809	high | to
0.178571	-14.081	therefore | an
0.107143	-14.0812	an | about
0.333333	-14.0812	nil | name
0.112583	-14.0813	x | ,
0.0845588	-14.0814	a | translation.
0.120301	-14.0815	− | of
0.5	-14.0815	decomposition | using
0.5	-14.0815	i’m | on
0.3	-14.0817	each | 12
0.185185	-14.0817	vertices | in
0.0698254	-14.0818	a | instance
0.666667	-14.0819	p(wi|tj)
0.666667	-14.0819	2008; | clark,
0.666667	-14.0819	theoretic | proposed
0.666667	-14.0819	quantifier | e.g.
0.666667	-14.0819	communicative | agent
0.209302	-14.0819	50 | 5
0.0665138	-14.082	that | phrase
0.146067	-14.0821	counting | of
0.137255	-14.0821	very | this
0.222222	-14.0822	use | include
0.444444	-14.0822	constant | data.
0.121212	-14.0823	lexicon | a | of
0.06621	-14.0823	general | a
0.2	-14.0823	for | ace
0.269231	-14.0825	tree | distance
0.0402685	-14.0825	labeled | the
0.0769231	-14.0825	a | recall
0.209302	-14.0826	with | analysis.
0.357143	-14.0827	order | handle
0.6	-14.0827	experiment | learnability
0.6	-14.0827	regard | foreign
0.153846	-14.0828	system | accuracy
0.0728291	-14.0828	to | language.
0.181818	-14.0828	during | translation
0.333333	-14.0829	work | entailment
0.666667	-14.083	avenue, | usa
0.666667	-14.083	p(pi)p(pi|w(pi))
0.666667	-14.083	extract | events.
0.666667	-14.083	13.9 | development
0.242424	-14.083	predicted | which
0.164179	-14.083	instead | each
0.151899	-14.083	detail | in
0.195652	-14.0831	a | phenomena.
0.119403	-14.0833	english | words
0.357143	-14.0833	model | retrieval.
0.357143	-14.0833	evaluations | human
0.315789	-14.0834	(e.g., | number
0.269231	-14.0834	first | presented
0.0396761	-14.0834	lexical | of
0.122137	-14.0834	of | document,
0.235294	-14.0835	our | product
0.0519663	-14.0837	each | an
0.6	-14.0837	relative) | points
0.195652	-14.0838	experiments | parsing
0.214286	-14.0839	on | wide
0.162162	-14.084	conduct | of
0.0262411	-14.084	and | it
0.666667	-14.0841	epi | extraction.
0.666667	-14.0841	across | groups.
0.4	-14.0841	domain | entities
0.4	-14.0841	improvement | confidence
0.333333	-14.0841	because | existing
0.0876494	-14.0843	calculate | of
0.115942	-14.0844	rules | not
0.0573356	-14.0845	the | of | 5
0.0769231	-14.0846	it | any
0.0221662	-14.0847	we | an
0.153846	-14.0849	would | into
0.0524823	-14.085	a | trained | the
0.0961538	-14.085	following | on
0.0989583	-14.0851	the | evaluation,
0.149425	-14.0851	for | (and
0.204545	-14.0851	same | time
0.384615	-14.0851	proposed | improved
0.051676	-14.0852	the | processing
0.12605	-14.0852	are | reference
0.666667	-14.0853	character-level | capture
0.666667	-14.0853	hypothesis | mechanism
0.269231	-14.0853	that | fails
0.11039	-14.0853	uses | on
0.0691358	-14.0855	0 | .
0.185185	-14.0855	this | article
0.141304	-14.0855	a | dialog
0.100529	-14.0857	different | our
0.333333	-14.0858	narrative | or
0.357143	-14.0858	did | good
0.6	-14.0858	harvest | *
0.5	-14.0858	aim | learn
0.5	-14.0858	10 | 67.4
0.133333	-14.0859	the | technical
0.155844	-14.086	as | whereas
0.4	-14.0861	parser | identifying
0.088	-14.0863	consider | for
0.666667	-14.0864	recognition | action
0.666667	-14.0864	tenses, | number,
0.666667	-14.0864	hours, | hours
0.666667	-14.0864	stuhlsatzenhausweg | 66123
0.666667	-14.0864	word | compl.
0.666667	-14.0864	category | offset
0.357143	-14.0864	probably | more
0.0433925	-14.0865	we | how
0.15	-14.0865	only | may
0.285714	-14.0866	lagrangian | for
0.5	-14.0866	yesterday | about
0.5	-14.0866	(using | gold
0.235294	-14.0867	(10) | =
0.105263	-14.0868	training | were
0.6	-14.0868	associated | abstract
0.6	-14.0868	shared | 201.
0.259259	-14.0868	correlations | with
0.10828	-14.0868	their | model
0.123967	-14.0868	from | sentence.
0.162162	-14.0871	this | i.e.,
0.0635593	-14.0871	the | corpus, | and
0.384615	-14.0871	moreover, | has
0.121212	-14.0871	decide | to
0.315789	-14.0872	obviously, | a
0.0666667	-14.0874	learning | that
0.149425	-14.0874	errors, | and
0.3	-14.0874	have | indicate
0.0730337	-14.0874	source | are
0.0448852	-14.0874	set | with
0.666667	-14.0875	selection | filtering.
0.666667	-14.0875	stating | some
0.228571	-14.0875	in | organization
0.333333	-14.0875	even | complex
0.444444	-14.0875	either | constraint
0.0869565	-14.0875	in | processing,
0.0411559	-14.0877	all | as
0.195652	-14.0878	also | various
0.2	-14.0878	provide | semantic
0.222222	-14.0879	process, | be
0.0491803	-14.088	we | evaluation
0.242424	-14.088	using | individual
0.0789474	-14.088	is | useful
0.107143	-14.088	our | similarity
0.153846	-14.088	word | without
0.169231	-14.0881	(2007) | that
0.0883534	-14.0881	common | is
0.0415162	-14.0883	the | paper
0.0671463	-14.0883	we | values
0.153846	-14.0883	strategy | that
0.235294	-14.0883	words | added
0.5	-14.0883	vbp | owns
0.666667	-14.0887	closely | possible,
0.666667	-14.0887	measures, | measure.
0.666667	-14.0887	ssa | news
0.666667	-14.0887	b | false
0.666667	-14.0887	approaches | detection,
0.666667	-14.0887	documents | nearly
0.315789	-14.0888	# | &
0.0359043	-14.0888	three | of
0.6	-14.0889	similarity | despite
0.0350765	-14.089	the | relations
0.25	-14.0892	tokens | pos
0.0285355	-14.0894	and | data
0.147727	-14.0896	already | by
0.204545	-14.0897	filter | with
0.105263	-14.0897	are | features.
0.384615	-14.0897	− | k
0.666667	-14.0898	shallow | analysis.
0.666667	-14.0898	messaging, | it
0.171875	-14.0898	fraction | in
0.444444	-14.0898	ontology | process
0.130841	-14.0899	units | to
0.135922	-14.09	from | provided
0.5	-14.09	framenet | resource
0.157895	-14.0901	obtained | the | obtained
0.0830325	-14.0902	in | news
0.134615	-14.0903	analysis, | to
0.139785	-14.0904	we | frame
0.166667	-14.0904	average | two
0.315789	-14.0904	saying | a
0.3	-14.0905	having | semantic
0.0842491	-14.0907	the | feedback
0.0710526	-14.0908	example | for
0.242424	-14.0909	applied | but
0.122951	-14.0909	(2008) | in
0.216216	-14.0909	no | learning
0.333333	-14.0909	content | models
0.333333	-14.0909	contexts, | are
0.6	-14.0909	amount | million
0.5	-14.0909	indices | vectors
0.5	-14.0909	senses | domain.
0.285714	-14.091	so | compare
0.666667	-14.091	pesa | only
0.666667	-14.091	provide | functionality
0.666667	-14.091	lse | vocabulary
0.384615	-14.0911	alone | than
0.137255	-14.0912	c | ,
0.209302	-14.0916	for | stop
0.147727	-14.0919	formal | in
0.357143	-14.0919	hence, | but
0.666667	-14.0921	phenotypes | same
0.666667	-14.0921	been | (finin
0.666667	-14.0921	smith | (2008)
0.188679	-14.0921	emails | and
0.0674699	-14.0922	time | is
0.16	-14.0922	model | 2009).
0.088	-14.0922	that | performance.
0.146067	-14.0923	if | algorithm
0.0802676	-14.0923	the | 2003)
0.2	-14.0923	tracking | of
0.214286	-14.0924	numerical | is
0.0873016	-14.0925	in | chinese
0.0730337	-14.0925	node | is
0.315789	-14.0926	contrast, | approach
0.228571	-14.0926	there are | differences
0.333333	-14.0926	but | problem.
0.10828	-14.0926	4 | used
0.123967	-14.0928	this | reference
0.178571	-14.0929	mean | or
0.4	-14.093	positive | samples
0.178571	-14.0932	the | car
0.333333	-14.0932	system, | source
0.666667	-14.0933	corresponding | achieves
0.666667	-14.0933	initial | (r)
0.109677	-14.0933	starting | a
0.285714	-14.0934	problem | like
0.161765	-14.0934	can | solved
0.164179	-14.0935	for | groups
0.5	-14.0935	lexical | representations.
0.0789474	-14.0937	corresponding | to
0.164179	-14.0937	from | above
0.384615	-14.0937	methods | candidate
0.0636943	-14.0938	in | often
0.195652	-14.0938	these | perform
0.11039	-14.0938	this | value
0.0943396	-14.0939	with | average
0.0859375	-14.0939	external | the
0.0230428	-14.094	1 | the | of
0.0573913	-14.0943	a | without
0.12605	-14.0943	k | 1
0.030754	-14.0943	be | be
0.5	-14.0943	preference | applying
0.181818	-14.0943	no | system
0.666667	-14.0944	effectively | algorithm.
0.666667	-14.0944	(2007) | hapax
0.16	-14.0947	parser | not
0.0909091	-14.0947	used | was
0.333333	-14.0949	t-fr | t-it
0.333333	-14.0949	classification | articles
0.0554649	-14.095	type | a
0.192308	-14.0951	applications | information
0.444444	-14.0951	built | features.
0.16	-14.0952	feature | context
0.25	-14.0952	readers | for
0.174603	-14.0952	the | sentiwordnet
0.5	-14.0952	selection | adaptation
0.5	-14.0952	define | select
0.269231	-14.0955	both | performed
0.195652	-14.0955	is | iteration
0.111111	-14.0955	these | have | to
0.666667	-14.0956	84.9% | f-score
0.101604	-14.0957	is | e.g.,
0.4	-14.0958	model | ibm-1
0.133333	-14.0958	we | 6.
0.228571	-14.0958	similarity | human
0.259259	-14.0958	perform | features
0.209302	-14.0959	h | l
0.444444	-14.0959	after | point
0.228571	-14.0961	9 | 15
0.5	-14.0961	correlation | rank≥
0.5	-14.0961	(10) | where
0.5	-14.0961	concatenate | vector
0.139785	-14.0962	d | l
0.181818	-14.0962	score | their
0.181818	-14.0962	features | four
0.115108	-14.0963	this | selection
0.0601504	-14.0963	and | within
0.285714	-14.0964	memory | from
0.315789	-14.0964	koehn | al.
0.0626305	-14.0966	data | which
0.15	-14.0966	by | them.
0.666667	-14.0967	t-test | influenced
0.666667	-14.0967	against | name.
0.666667	-14.0967	lead | uniform
0.666667	-14.0967	missing | known
0.333333	-14.0967	other | 2005),
0.333333	-14.0967	to | contributions.
0.357143	-14.0968	a | uk
0.269231	-14.0969	b | h
0.5	-14.0969	follows | low
0.25	-14.097	metric | but
0.25	-14.097	speech | one
0.0221374	-14.0971	of | language
0.3	-14.0972	with | expressions.
0.15	-14.0974	in | adjacent
0.315789	-14.0975	human | similarity
0.0562914	-14.0976	for | phrase
0.228571	-14.0977	first | chinese
0.384615	-14.0977	dashed | indicate
0.032382	-14.0978	which | as
0.0145792	-14.0978	in | are
0.2	-14.0978	paper, | first
0.5	-14.0978	γ | ∈
0.666667	-14.0979	realization | method
0.666667	-14.0979	vogel | 15213,
0.0634249	-14.0979	and | sentiment
0.235294	-14.098	1 | (3)
0.6	-14.0982	many | helpful
0.0603113	-14.0985	we | documents
0.146067	-14.0986	parsing, | a
0.153846	-14.0986	target | only
0.357143	-14.0987	this | grammars.
0.5	-14.0987	aspects, | aspects
0.0606654	-14.0988	0.00 | 0.00
0.2	-14.0988	are | interest
0.666667	-14.099	metric | hter.
0.666667	-14.099	distribution, | around
0.181818	-14.099	translation | pairs.
0.209302	-14.0991	is | unlikely
0.315789	-14.0991	learning:share | 23-24
0.132075	-14.0991	to | lattice
0.3	-14.0992	syntactic | provide
0.0846154	-14.0992	its | or
0.0917431	-14.0993	we | create
0.357143	-14.0993	regularized | =
0.16	-14.0994	can | various
0.0815603	-14.0995	that | sense
0.195652	-14.0996	it | language.
0.192308	-14.0997	in order | must
0.384615	-14.0997	expl | expl;
0.0419776	-14.0997	the | frequent
0.108974	-14.0997	it | whether
0.144444	-14.0998	if | relation
0.0818505	-14.1	with | human
0.132075	-14.1	discourse | with
0.107595	-14.1001	each | according
0.666667	-14.1002	turn | (turn
0.0279146	-14.1005	the | document
0.137255	-14.1005	that, | for
0.444444	-14.1005	theme, | that
0.0710526	-14.1005	of | lexicon
0.122951	-14.1006	and | requires
0.333333	-14.1007	described | syntactic
0.144444	-14.1008	any | all
0.164179	-14.1008	by | role
0.0659091	-14.1009	the | occurrences
0.0632911	-14.1009	argument | and
0.0771704	-14.101	this | large
0.125	-14.101	on | wikipedia
0.666667	-14.1013	1450-1550 | 1650-1750
0.666667	-14.1013	adjust | parameters
0.315789	-14.1013	the | rhs
0.4	-14.1013	und | zu
0.4	-14.1013	at | chart
0.4	-14.1013	solving | following
0.4	-14.1013	lms | translation
0.444444	-14.1013	document | identified
0.444444	-14.1013	die | katze
0.5	-14.1013	request | student
0.241379	-14.1014	(if | that
0.241379	-14.1014	type | target
0.241379	-14.1014	in | inverted
0.0638298	-14.1014	as | over
0.03396	-14.1016	method | and
0.139785	-14.1017	a | that | being
0.0412926	-14.1018	these | be
0.157895	-14.1018	if | single
0.113475	-14.1019	which | well
0.204545	-14.102	has | effective
0.204545	-14.102	used | derived
0.0989583	-14.1021	look | and
0.285714	-14.1023	their | match
0.241379	-14.1023	verb | construction
0.6	-14.1023	representation | incorporating
0.357143	-14.1024	two | students
0.666667	-14.1025	• | percentage:
0.666667	-14.1025	impact | uncertainty
0.666667	-14.1025	joshua | open-source
0.0183034	-14.1025	that | that
0.0554649	-14.1026	which | word
0.0792079	-14.1027	different | which
0.178571	-14.1028	mentions | not
0.216216	-14.1028	relatively | number
0.269231	-14.1029	relevant | language
0.0801394	-14.1029	description | and
0.153846	-14.1029	are | called
0.157895	-14.1029	many | one
0.178571	-14.1031	an | seen
0.101695	-14.1031	fact | that
0.259259	-14.1034	grammatical | between
0.259259	-14.1034	of | parallel.
0.0726257	-14.1034	that | input
0.333333	-14.1036	problem | hierarchical
0.666667	-14.1037	frames | markers
0.666667	-14.1037	classification | coreferent
0.666667	-14.1037	edge | (ned),
0.666667	-14.1037	r., | e.,
0.228571	-14.1037	these | included
0.195652	-14.1037	for | clinical
0.0415162	-14.1039	that | two
0.115108	-14.1039	a | algorithm for
0.169231	-14.1039	features | single
0.0505464	-14.104	lower | the
0.25	-14.1044	0.66 | 0.66
0.0446334	-14.1045	the | vectors
0.315789	-14.1046	(2007) | information
0.132075	-14.1046	using | extracted
0.114286	-14.1047	because | than
0.666667	-14.1048	during | period,
0.666667	-14.1048	were | assignment.
0.666667	-14.1048	bacteria | (rename)
0.188679	-14.1048	by | called
0.12	-14.1049	morphological | as
0.115942	-14.105	to | this,
0.0737463	-14.1051	to | low
0.146067	-14.1052	from | often
0.15942	-14.1052	results | against
0.0568966	-14.1052	corpus | from
0.155844	-14.1054	term | from
0.0871369	-14.1056	using | corpus
0.5	-14.1056	: | denotes
0.285714	-14.1058	correct | their
0.666667	-14.106	support | hr0011-08-c-0110
0.666667	-14.106	tree, | inserted
0.3	-14.106	languages, | which
0.3	-14.106	uses | structure
0.0849421	-14.1061	0.00 0.00 | 0.00
0.0849421	-14.1061	context | are
0.241379	-14.1061	pairs | no
0.241379	-14.1061	regression | data
0.228571	-14.1061	four | features
0.051105	-14.1062	and | rules
0.235294	-14.1062	(2010) | models
0.121951	-14.1062	> | a
0.285714	-14.1063	important | aspects
0.178571	-14.1065	(the | which
0.118519	-14.1065	that | identified
0.0774194	-14.1065	and | generate
0.120968	-14.1065	be | during
0.0878661	-14.1065	of | 24
0.10625	-14.1066	with | output
0.105263	-14.1066	the | (2010).
0.181818	-14.1066	results for | test
0.118519	-14.1067	computing | in
0.444444	-14.1067	during | according
0.241379	-14.1069	although | large
0.0532725	-14.1069	we | types
0.216216	-14.107	yielded | in
0.0371449	-14.107	in | 1
0.666667	-14.1071	handles | documents.
0.666667	-14.1071	scored(wprime |w)
0.666667	-14.1071	derive | assist
0.666667	-14.1071	significant, | lower
0.333333	-14.1071	performance | existing
0.100559	-14.1072	of | connected
0.241379	-14.1073	set | queries
0.269231	-14.1076	based | co-occurrence
0.4	-14.1076	when | message
0.0762195	-14.1079	the | empty
0.204545	-14.1081	view, | a
0.204545	-14.1081	80 | 80
0.3	-14.1081	pair | instances
0.12963	-14.1082	scores | model
0.5	-14.1082	time | tuning
0.666667	-14.1083	approaches | inherent
0.00916903	-14.1083	a | in
0.118519	-14.1083	parameters | as
0.4	-14.1083	named | person,
0.120968	-14.1084	that | left
0.228571	-14.1085	the | constructed.
0.0952381	-14.1086	in | concept
0.107595	-14.1087	quality | by
0.25	-14.1087	weight | its
0.0651685	-14.1088	clustering | of
0.0856031	-14.1088	need | for
0.0824373	-14.1089	which | its
0.10828	-14.1089	that | provides
0.241379	-14.109	entries | phrase
0.4	-14.109	bleu | additional
0.444444	-14.109	any | underlying
0.0259972	-14.1093	the | english
0.0815603	-14.1093	positive | is
0.357143	-14.1093	evidence | words
0.00758721	-14.1093	in | and
0.666667	-14.1095	whether | another,
0.142857	-14.1095	distribution | an
0.0374251	-14.1096	to | sentences
0.0714286	-14.1097	all | an
0.139785	-14.1098	= | 1,
0.142857	-14.1098	information | source
0.0821429	-14.1101	figure 3: | the
0.272727	-14.1104	novel | learning
0.133333	-14.1104	little | to
0.25	-14.1105	these | translated
0.0779221	-14.1105	we | reordering
0.666667	-14.1106	greedy | valid
0.666667	-14.1106	students | 200
0.0759878	-14.1106	et | 2004;
0.12844	-14.1106	at | over
0.333333	-14.1106	consider | sense
0.259259	-14.1106	the | control.
0.259259	-14.1106	reviews, | in
0.444444	-14.1106	frequent | appear
0.122951	-14.1107	which | should
0.0807018	-14.1108	it is | this
0.188679	-14.1108	to | updates
0.5	-14.1108	need | question.
0.5	-14.1108	interpolation | standard
0.204545	-14.111	using | sentences.
0.113475	-14.111	a | corpora.
0.0989583	-14.1111	only | between
0.216216	-14.1111	it | whose
0.185185	-14.1114	task | bionlp
0.0959596	-14.1114	present | from
0.444444	-14.1114	by adding | data
0.0807018	-14.1115	built | of
0.0431373	-14.1116	that | their
0.0846154	-14.1116	this | parsing
0.6	-14.1117	bilingual | pages.
0.6	-14.1117	therefore, | define
0.5	-14.1117	(rose | al.,
0.666667	-14.1118	iria | and
0.4	-14.1118	have | research.
0.269231	-14.1119	an | classification.
0.12605	-14.112	european | the
0.12844	-14.112	these | first
0.0525547	-14.112	the | included | the
0.0393443	-14.1121	its | to
0.0757576	-14.1122	though | of
0.10625	-14.1123	for | morphological
0.118519	-14.1123	in | link
0.138298	-14.1123	one | approach
0.0827068	-14.1124	r | n
0.146067	-14.1124	is | type.
0.171875	-14.1125	as | interaction
0.4	-14.1125	works | input
0.157895	-14.1126	recall | which
0.5	-14.1126	learning | open
0.5	-14.1126	if | same.
0.6	-14.1128	bill | –
0.12844	-14.1129	both | use
0.333333	-14.1129	etzioni, | et
0.666667	-14.113	adverse | averse
0.136842	-14.113	syntactic | information
0.204545	-14.1131	case | verb
0.178571	-14.1131	to | events,
0.0704607	-14.1131	up | to
0.127273	-14.1132	set | relation
0.269231	-14.1133	motivation | this
0.125	-14.1133	different | into
0.127273	-14.1134	the | sides
0.0934579	-14.1135	to | learned
0.0690537	-14.1136	the best | is
0.0483271	-14.1137	an | it
0.241379	-14.1137	version | has
0.101124	-14.1137	all | results
0.444444	-14.1137	following | example:
0.209302	-14.1138	example | ,
0.195652	-14.1139	all | approaches
0.6	-14.1139	mwen | yo
0.666667	-14.1141	involved | relation,
0.666667	-14.1141	if|t|
0.0809859	-14.1141	questions | of
0.172414	-14.1142	true | bool
0.2	-14.1142	3. | all
0.0694087	-14.1142	in | simple
0.148148	-14.1142	used | classification
0.357143	-14.1143	4. | any
0.0864198	-14.1143	developed | to
0.1	-14.1144	the | side,
0.15942	-14.1144	model | several
0.0538462	-14.1144	by | at
0.12605	-14.1145	for | etc.
0.12844	-14.1145	we | towards
0.0348387	-14.1145	a | where | of
0.204545	-14.1146	number | m
0.175439	-14.1146	results | result
0.315789	-14.1146	studies | 2006;
0.315789	-14.1146	dynamic | states
0.315789	-14.1146	throw | quarterback
0.0209692	-14.1148	to | an
0.222222	-14.1149	features | distance
0.357143	-14.1149	null | position
0.357143	-14.1149	features | class.
0.6	-14.1149	wi | log
0.3	-14.115	would | low
0.188679	-14.1151	by | alignments
0.0150268	-14.1152	we | as
0.5	-14.1152	patterns | long-distance
0.5	-14.1152	out | time.
0.5	-14.1152	features | part-of-speech,
0.0690537	-14.1154	class | in
0.241379	-14.1154	some | work.
0.146341	-14.1154	with | computed
0.0661939	-14.1155	errors | in
0.014055	-14.1157	by | and
0.333333	-14.1159	w3 | w2
0.4	-14.116	work | summarization.
0.6	-14.116	brown | 1992)
0.444444	-14.116	liu, | 2005;
0.0466744	-14.1162	the | opinion
0.185185	-14.1163	different | so
0.3	-14.1165	while | current
0.0259086	-14.1165	probability | the
0.0818505	-14.1166	al., | 2004; | et
0.109677	-14.1166	implement | the
0.0573477	-14.1168	extraction | a
0.444444	-14.1168	concepts | single
0.181818	-14.1168	is | estimation
0.222222	-14.1169	relations | different
0.228571	-14.1169	such | relevant
0.333333	-14.117	trees | reordering
0.333333	-14.117	0.60 | 0.53
0.195652	-14.117	et | 1999).
0.6	-14.117	respectively) | null1
0.6	-14.117	pb(w|h)
0.12963	-14.1171	to | asr
0.0807018	-14.1172	results | more
0.285714	-14.1173	joint | query
0.135417	-14.1173	for | answer
0.175439	-14.1174	even | some
0.272727	-14.1175	no | significant
0.444444	-14.1176	find | path
0.0989583	-14.1177	work | semantic
0.666667	-14.1177	acl | searchbench
0.0776699	-14.1178	the | transfer
0.5	-14.1179	after | hours
0.5	-14.1179	dependency | mst
0.064877	-14.118	list | in
0.272727	-14.118	by | fellowship
0.269231	-14.118	many | examples
0.269231	-14.118	be | subsequent
0.11039	-14.1181	features | over
0.191489	-14.1184	feature | candidate
0.191489	-14.1184	compare | models
0.138298	-14.1184	except | is
0.0818505	-14.1186	evidence | of
0.235294	-14.1186	in | ii
0.0724234	-14.1186	t | t | e
0.210526	-14.1186	documents | use
0.0698925	-14.1187	to | local
0.102857	-14.1188	on | gold
0.195652	-14.1188	where | top
0.127273	-14.1188	it | negative
0.666667	-14.1189	(sections | 4).
0.272727	-14.119	4.2 | experiments
0.216216	-14.1191	for | article.
0.125	-14.1192	and | tagged
0.204545	-14.1193	this | caused
0.25	-14.1193	feedback | learning
0.134615	-14.1194	more | there
0.0733138	-14.1196	is | average
0.144444	-14.1197	i | h
0.209302	-14.1197	experiment | training
0.0512821	-14.1197	score | is
0.2	-14.1198	set | mention
0.0900901	-14.1198	also | number
0.269231	-14.1199	feature | scoring
0.185185	-14.1199	for | relation,
0.357143	-14.1199	with | news-c
0.666667	-14.12	implicit | passive
0.272727	-14.12	not | step.
0.133333	-14.12	(2008) | for
0.444444	-14.12	punyakanok | al.
0.0687023	-14.1202	as | lexical
0.102857	-14.1202	are | sets
0.0522496	-14.1202	a | r
0.0532725	-14.1203	in | related
0.0265222	-14.1204	of | was
0.209302	-14.1204	next, | as
0.126126	-14.1205	b | that
0.210526	-14.1208	the | tagset,
0.0818505	-14.121	clustering | a
0.155844	-14.121	extracted | parallel
0.0580762	-14.121	we | problem
0.151899	-14.121	task, | from
0.666667	-14.1212	set | range.
0.0735294	-14.1213	to | relative
0.6	-14.1213	address | method.
0.11811	-14.1213	systems | language
0.2	-14.1213	candidate | only
0.0495979	-14.1214	the | algorithm | in
0.0284238	-14.1214	which | is
0.25	-14.1215	among | but
0.0830189	-14.1216	training, | the
0.4	-14.1217	 | 
0.155844	-14.1218	→ | with
0.142857	-14.1218	on | generation
0.333333	-14.1218	formalisms | as
0.0405286	-14.122	it | that | in
0.126126	-14.1221	these | about
0.10559	-14.1221	we | utterances
0.259259	-14.1221	while | find
0.259259	-14.1221	corpora | texts
0.185185	-14.1222	marks | and
0.169231	-14.1223	constrained | to
0.333333	-14.1223	break | prosodic
0.3	-14.1224	transliteration | training
0.272727	-14.1224	proposed | 2007).
0.0779221	-14.1224	is | list
0.204545	-14.1225	fewer | are
0.114286	-14.1228	c | 1
0.3	-14.1229	synchronous | machine
0.272727	-14.1229	annotated | performance
0.216216	-14.123	as | em
0.241379	-14.123	for | j,
0.25	-14.1232	more | queries
0.3	-14.1234	search | document
0.0327775	-14.1234	the | following
0.0827068	-14.1235	directly | a
0.666667	-14.1236	scoring | governor
0.161765	-14.1236	sentences | semantic
0.172414	-14.124	which | total
0.169231	-14.1241	major | that
0.142857	-14.1241	not | across
0.122951	-14.1242	well | other
0.12963	-14.1242	we | determine | a
0.228571	-14.1242	transliteration | word
0.102273	-14.1243	in | edge
0.0900901	-14.1243	one | but
0.178571	-14.1244	model | up
0.357143	-14.1244	methods | deal
0.175439	-14.1245	an | factor
0.151899	-14.1245	a | 15
0.114286	-14.1246	indicates | for
0.285714	-14.1249	take | relations
0.104294	-14.1249	hierarchical | in
0.109677	-14.1251	message | and
0.0807018	-14.1252	as | 2010).
0.185185	-14.1252	will | probability
0.272727	-14.1253	pairs | related
0.0757098	-14.1253	of | included
0.285714	-14.1254	with | base.
0.0263451	-14.1254	any | the
0.107595	-14.1254	selected | with
0.0220379	-14.1255	model | a
0.0954774	-14.1256	work | data
0.0959596	-14.1256	our | trained
0.666667	-14.1259	global | fully
0.666667	-14.1259	including | few.
0.666667	-14.1259	1...t | 1...n
0.666667	-14.1259	pr(e|f)
0.5	-14.1259	different | runs,
0.5	-14.1259	article, | has
0.0724638	-14.1261	association | in
0.0398305	-14.1261	the | k
0.241379	-14.1265	most | few
0.6	-14.1266	searchbench | search
0.175439	-14.1267	to | wider
0.5	-14.1268	obtaining | improvements
0.5	-14.1268	c1000 | upos
0.108974	-14.1271	task | set
0.0544872	-14.1272	the | experiment
0.111111	-14.1275	our | search
0.191489	-14.1276	systems | sentences
0.0935961	-14.1277	previous | are
0.241379	-14.1278	under | given
0.146341	-14.1278	to | representative
0.0491591	-14.128	error | of
0.0716253	-14.128	we | obtained
0.161765	-14.128	that | differ
0.228571	-14.1282	next | describe
0.666667	-14.1283	sang | meulder,
0.210526	-14.1283	aligning | a
0.0850202	-14.1284	in | observed
0.0307062	-14.1285	while | the | of
0.127273	-14.1285	can | types
0.5	-14.1285	srl | performed
0.25	-14.1286	comparison | methods
0.191489	-14.1286	of | thematic
0.0697051	-14.1287	of | inference
0.095	-14.1288	for | chinese
0.135417	-14.1288	hindi | and
0.102857	-14.1294	and | complexity
0.5	-14.1294	annotation | disagreements
0.666667	-14.1295	alternative | come
0.333333	-14.1295	different | towards
0.123967	-14.1295	work | has been
0.0782313	-14.1297	scores | with
0.142857	-14.1297	reported | table
0.161765	-14.1297	algorithm | feature
0.101124	-14.1298	to | discussed
0.057971	-14.1298	since | be
0.6	-14.1298	outperform | model.
0.3	-14.1299	argument | particular
0.228571	-14.1299	to | variance
0.123967	-14.1299	model | learning
0.0654206	-14.13	our | corpus
0.0892857	-14.1303	4.1 | we
0.0355921	-14.1304	for | at
0.0945274	-14.1306	are | like
0.666667	-14.1307	among | alternatives
0.666667	-14.1307	danish, | swedish.
0.666667	-14.1307	0.471 | 0.606
0.228571	-14.1307	data | few
0.175439	-14.1308	it | range
0.6	-14.1309	proteins | gene
0.0295775	-14.131	the | possible
0.0989011	-14.1314	would | which
0.0661939	-14.1315	the | selecting
0.0318332	-14.1315	from | that | the
0.0318506	-14.1319	we | that | that
0.144578	-14.1319	their | also
0.125	-14.1319	the | balanced
0.3	-14.132	systems, | use
0.0680101	-14.132	per | of
0.0398614	-14.132	word | with
0.1875	-14.1321	this | explored
0.0698925	-14.1321	is | negative
0.120968	-14.1321	information | all
0.5	-14.1321	experiment | wm
0.034005	-14.1322	in | they
0.0514286	-14.1323	a | evaluation
0.178571	-14.1324	sentences | words.
0.117188	-14.1324	that | who
0.169492	-14.1326	set | patterns
0.115108	-14.1326	and | implementation
0.144578	-14.1326	in | received
0.3125	-14.1329	as | gapping
0.175439	-14.133	these cases | the
0.0811808	-14.133	in | social
0.6	-14.133	bleu | bi-gram
0.5	-14.133	appearing | containing
0.444444	-14.1334	parenleftbigg | parenrightbiggparenrightbigg
0.0473876	-14.1335	this | use
0.228571	-14.1336	however, | corpora
0.259259	-14.1336	that | stored
0.191489	-14.1337	generated | this
0.4	-14.1338	above | model.
0.175439	-14.1339	our | must
0.272727	-14.134	contextual | between
0.285714	-14.1341	of | reports.
0.444444	-14.1342	idea | languages.
0.444444	-14.1342	(2010) | dynamic
0.4	-14.1345	they | ones.
0.106918	-14.1346	an | entity
0.0704607	-14.1347	to | bilingual
0.12844	-14.1348	al | a
0.166667	-14.1349	2009) | with
0.166667	-14.1349	0.3 | 0.3
0.0424901	-14.1349	the | what
0.272727	-14.135	one | chosen
0.444444	-14.135	95% | intervals
0.0697051	-14.1352	method, | the
0.15493	-14.1352	are | chinese
0.4	-14.1352	expresses | positive
0.0909091	-14.1352	evidence | and
0.15	-14.1352	after | it
0.6	-14.1352	sβ> | <sα,
0.181818	-14.1353	in | definitions
0.0630435	-14.1355	is | even
0.0634573	-14.1356	one | not
0.142857	-14.1356	presented | which
0.100559	-14.1356	bias | the
0.0867769	-14.1357	which | also
0.259259	-14.1359	data | unknown
0.333333	-14.136	defined | tokens
0.166667	-14.1361	tokens | each
0.157143	-14.1361	present | evaluation
0.161765	-14.1361	these | scores
0.0892857	-14.1362	and | did
0.195652	-14.1362	also | value
0.112676	-14.1363	can | easily
0.259259	-14.1364	use | summarization
0.178571	-14.1365	that | generalize
0.5	-14.1366	helpfulness | predicting
0.122951	-14.1367	child | and
0.216216	-14.1369	document | corpus.
0.146341	-14.137	set | selected
0.151899	-14.137	features: | to
0.169492	-14.1371	trees | using
0.0697051	-14.1373	of | learned
0.0661765	-14.1373	the | building
0.144578	-14.1373	relevant | by
0.148148	-14.1374	parallel | sentences
0.146341	-14.1375	the | average.
0.0345502	-14.1376	our | from
0.0666667	-14.1378	detailed | the
0.233333	-14.1379	null, | a
0.0409091	-14.138	is | feature
0.0817844	-14.1382	we | interested | the
0.0339409	-14.1382	the | parsing
0.222222	-14.1383	than | larger
0.233333	-14.1383	probabilistic | semantic
0.10559	-14.1383	the | captured
0.333333	-14.1384	thread, | of
0.357143	-14.1384	correlations | not
0.6	-14.1384	could | ambiguity
0.108844	-14.1384	english | not
0.0488127	-14.1385	for | source
0.164179	-14.1386	experiments, | used
0.216216	-14.1388	produce | translation
0.164179	-14.1389	not | certain
0.0805861	-14.1389	the | native
0.210526	-14.1389	learned | an
0.161765	-14.139	a | y.
0.141304	-14.1391	represents | that
0.285714	-14.1393	task | up
0.0495979	-14.1393	of | computational | pages
0.0805861	-14.1393	and | obtain
0.111111	-14.1394	5 | .
0.0185359	-14.1394	the | features | the
0.157143	-14.1395	different | those
0.6	-14.1395	is, | fact,
0.139785	-14.1396	approach | better
0.3125	-14.1396	because | know
0.1875	-14.1398	interesting | can
0.0983607	-14.1398	word | disambiguation
0.444444	-14.1398	traditional | classification
0.109589	-14.1399	of | referring
0.1875	-14.1402	be | representing
0.15493	-14.1402	exactly | same
0.153846	-14.1402	rather | with
0.0733138	-14.1402	we | linear
0.5	-14.1402	x | create
0.5	-14.1402	low | useful
0.5	-14.1402	accuracy, | evaluation
0.172414	-14.1403	but | did
0.0830189	-14.1404	usually | to
0.0635965	-14.1404	relevant | of
0.134021	-14.1405	machine | machine
0.259259	-14.1406	class | assigned
0.15493	-14.1408	or | words.
0.241379	-14.1408	from | especially
0.333333	-14.1408	those | lower
0.2	-14.1409	an | represent
0.111888	-14.141	of | linked
0.4	-14.141	types | relationships
0.5	-14.1411	construction | discussed
0.126126	-14.1413	corpus | pairs
0.0811808	-14.1413	clusters | to
0.333333	-14.1414	direct | when
0.666667	-14.1415	train | tagger,
0.0787671	-14.1416	the | strategies
0.0871369	-14.1417	0 | 10
0.157143	-14.1418	one | domain
0.151899	-14.1419	of | noun.
0.142857	-14.142	uncertainty | and
0.228571	-14.1422	user | also
0.0779661	-14.1423	to | results.
0.0730994	-14.1424	in | computational
0.104294	-14.1424	are | make
0.0693333	-14.1426	on | when
0.0492021	-14.1427	following | to
0.333333	-14.1432	general | large
0.333333	-14.1432	editors | turns
0.181818	-14.1432	abstract | shared
0.285714	-14.1434	segmentation | chinese
0.259259	-14.1434	was | relative
0.0735294	-14.1435	is | fact | the
0.178571	-14.1436	accuracy | three
0.12963	-14.1437	and | frequently
0.0495979	-14.1437	are | two
0.111111	-14.1438	results | given
0.666667	-14.1439	4000 | 8000
0.666667	-14.1439	{wi|wi
0.666667	-14.1439	based | requirements.
0.3	-14.1439	had | toys
0.3125	-14.1441	r-prior | kul
0.3125	-14.1441	(3) | .
0.3125	-14.1441	given | (a
0.134021	-14.1441	from | top
0.0676692	-14.1441	algorithm | that
0.027673	-14.1442	more | in
0.0867769	-14.1442	words | our
0.210526	-14.1442	problem | then
0.241379	-14.1443	to take | into
0.259259	-14.1444	these | tend
0.259259	-14.1444	interesting | other
0.169492	-14.1445	by | mapping
0.107595	-14.1445	local | is
0.333333	-14.1445	include | tags
0.1875	-14.1446	into | vector
0.144578	-14.1446	tend | more
0.0935961	-14.1447	if | no
0.101124	-14.1447	200 | of
0.169492	-14.1448	same | target
0.082397	-14.1448	one | words
0.136842	-14.145	such | any
0.125	-14.145	interface | to
0.666667	-14.1451	p(s|e),
0.216216	-14.1451	(3) | or
0.0759494	-14.1451	set | using
0.0846774	-14.1452	the | characteristics
0.3125	-14.1453	step | find
0.115385	-14.1453	is | dependent
0.112676	-14.1453	using | corpus.
0.0754717	-14.1455	a | low
0.228571	-14.1455	learning | tasks
0.0637813	-14.1457	allows | to | of
0.0347826	-14.1457	the | of | dependency
0.3125	-14.1458	candidates | corpus.
0.272727	-14.1459	unsupervised | performance
0.0989011	-14.1459	shows | using
0.175439	-14.1462	results | annotated
0.357143	-14.1462	these | lexicon.
0.444444	-14.1462	also | away
0.104294	-14.1464	a | labeling
0.0793103	-14.1465	address | and
0.0940594	-14.1465	spoken | and
0.5	-14.1465	parallel | barzilay
0.141176	-14.1466	2. | not
0.05	-14.1467	to | document
0.2	-14.1467	first | sentence.
0.0431211	-14.1467	paper | and
0.333333	-14.1469	gmm | r
0.161765	-14.1469	using | system.
0.3125	-14.147	years | have
0.3	-14.1472	our | participating
0.0743034	-14.1474	synset | the
0.0472637	-14.1474	by | model
0.5	-14.1474	dialog | asr
0.5	-14.1474	summarization | represent
0.3125	-14.1475	document | ranked
0.666667	-14.1475	linked | tweets,
0.047205	-14.1476	entity | and
0.1875	-14.1476	grammar | used
0.1	-14.1477	any | can
0.142857	-14.1477	on | context.
0.0850202	-14.1478	by | context
0.144578	-14.1478	seen | as
0.195652	-14.1479	dataset | section
0.127273	-14.148	one | over
0.333333	-14.1481	score | against
0.333333	-14.1481	regression, | of
0.172414	-14.1482	re | a
0.175439	-14.1482	feature | accuracy
0.0363636	-14.1482	, | the
0.101124	-14.1482	system | but
0.142857	-14.1485	are | selection
0.15	-14.1485	a | article.
0.135417	-14.1486	of | nps
0.0777027	-14.1487	of | correctly
0.222222	-14.1487	semiring | a
0.161765	-14.1487	our | presented
0.272727	-14.1488	correlation | computed
0.0642202	-14.1489	update | the
0.101796	-14.149	we | who
0.0909091	-14.149	by | any
0.127273	-14.149	next, we | of
0.0857143	-14.149	integrated | the
0.111111	-14.1491	in | applying
0.259259	-14.1491	however, | original
0.144578	-14.1491	because | do
0.6	-14.1493	together | initial
0.12963	-14.1494	token | for
0.10559	-14.1494	(5) | of
0.132653	-14.1494	currently | in
0.076412	-14.1495	4 | by
0.444444	-14.1495	kernels | relation
0.0747664	-14.1496	a | discourse
0.191489	-14.1496	large | information
0.10559	-14.1498	evaluate | as
0.157143	-14.1498	by | l
0.0680628	-14.1498	to | among
0.144578	-14.1499	as | predicate
0.138298	-14.1499	might | more
0.157143	-14.1501	models | parsing
0.10559	-14.1502	words | pos
0.146341	-14.1502	such | were
0.114504	-14.1504	pipeline | and
0.0603622	-14.1506	given | are
0.175439	-14.1507	has | small
0.357143	-14.1507	despite | these
0.272727	-14.1508	edge(nullnullnullnull | add
0.0769231	-14.1508	involved | the
0.5	-14.1511	patterns | attributes
0.131313	-14.1512	instance | an
0.161765	-14.1513	and | appearing
0.10303	-14.1513	with | quality
0.134021	-14.1514	forms | are
0.357143	-14.1514	modules | other
0.259259	-14.1514	parse | but
0.25	-14.1516	automatic | reference
0.0881057	-14.1516	people | and
0.0392491	-14.1516	the | measure
0.216216	-14.1517	are | parses
0.166667	-14.1518	1 | ,ap
0.075	-14.1518	can | if
0.4	-14.1519	score: | +
0.191489	-14.1521	propbank | et al.,
0.285714	-14.1522	less | most
0.0403587	-14.1522	small | of
0.333333	-14.1523	tested | sentences
0.0646651	-14.1525	p | a
0.3125	-14.1527	cuny | 2010)
0.222222	-14.1528	parenleftbig | =
0.0978261	-14.1529	structure | with
0.0313006	-14.1529	model | for
0.127273	-14.1531	in | works
0.0843373	-14.1531	at | =
0.4	-14.1533	12 | categories
0.0860656	-14.1533	methods | with
0.0718391	-14.1534	an | number
0.181818	-14.1534	together | with
0.0613108	-14.1535	and | system.
0.666667	-14.1536	vectors | frequency.
0.0945274	-14.1536	semantic | using
0.0392157	-14.1537	of | human
0.138298	-14.1537	conducted | in
0.1875	-14.1538	that, | or
0.0532915	-14.1538	much | of
0.5	-14.1538	will | attribute
0.0682415	-14.1539	the | spoken
0.0693333	-14.154	the | of | candidates
0.108108	-14.1541	this | provided
0.191489	-14.1542	then | document
0.0295257	-14.1544	the | verb
0.082397	-14.1544	like | is
0.195652	-14.1547	last | it
0.5	-14.1547	tr | o
0.172414	-14.1548	showing | for
0.285714	-14.1548	event | events,
0.666667	-14.1548	asked | narrative
0.4	-14.1548	token | gold
0.6	-14.1548	single-beam | beam
0.0989011	-14.1549	model, | that
0.3125	-14.155	mention | np
0.195652	-14.155	words | multiple
0.216216	-14.1553	word | occur
0.0791367	-14.1553	structures | the | of
0.25	-14.1553	related | social
0.357143	-14.1553	since | features,
0.0733945	-14.1555	of | particular | of
0.104938	-14.1556	corpus | test
0.0945274	-14.1557	in proceedings of | on
0.0724638	-14.1557	instances | a
0.216216	-14.1557	output | both
0.5	-14.1557	off | stack
0.0202872	-14.156	for each | the
0.0857143	-14.156	last | to
0.111888	-14.1562	that | various
0.095	-14.1563	second | from
0.285714	-14.1564	example | previous
0.0945274	-14.1564	maximum | in
0.3	-14.1565	documents | being
0.0873362	-14.1565	use | more
0.363636	-14.1565	expression | linguistic
0.363636	-14.1565	trigrams | from
0.363636	-14.1565	disagree | agree
0.333333	-14.1566	suggest | into
0.123894	-14.1566	b | e
0.0896861	-14.1567	these | two
0.0402145	-14.1568	the | of | best
0.444444	-14.1568	and | five.
0.0722543	-14.1569	for | sentiment
0.178571	-14.157	languages | other
0.6	-14.157	variable | entire
0.169492	-14.1571	about | have
0.0425532	-14.1572	evaluation | in
0.666667	-14.1573	p(witi|wi−11
0.666667	-14.1573	inf | %
0.666667	-14.1573	arg2 | arg2-gol
0.666667	-14.1573	scales | “peaceful
0.666667	-14.1573	exploring | 2009).
0.272727	-14.1573	selection | out
0.104938	-14.1573	for | french
0.357143	-14.1573	ratio | n-gram
0.357143	-14.1573	noun | possessive
0.048366	-14.1574	the | appear
0.0584906	-14.1574	with | system
0.183673	-14.1574	were | either
0.13	-14.1576	0.1 | 0.3
0.444444	-14.1576	(szpektor | al.,
0.0994475	-14.1577	reduces | the
0.0896861	-14.1577	precision, | and
0.363636	-14.1578	improve | system,
0.10625	-14.1579	a | world
0.146341	-14.158	had | from
0.0881057	-14.158	defined | that
0.103659	-14.1581	word | probability
0.666667	-14.1585	0::b | b0::b
0.666667	-14.1585	morphologically | 2010),
0.666667	-14.1585	below, | :
0.178571	-14.1586	actual | as
0.191489	-14.1588	our | question
0.0642202	-14.1588	and | positive
0.108844	-14.159	by | label
0.1875	-14.1593	our | candidates
0.272727	-14.1593	evaluated | precision
0.125	-14.1593	necessary | and
0.135417	-14.1595	n-gram | from
0.0495868	-14.1596	and | 2008;
0.241379	-14.1597	agreement | number
0.363636	-14.1598	ter | 30.9
0.169492	-14.1599	would | better
0.4	-14.1599	development | china
0.26087	-14.16	described | 3.2.
0.039929	-14.1601	to | any
0.164179	-14.1601	are | mt
0.0417881	-14.1601	would | a
0.191489	-14.1602	one | function
0.157143	-14.1604	f1 | with
0.357143	-14.1605	involve | protein
0.285714	-14.1606	a | bullets
0.178571	-14.1606	with | allow
0.0283273	-14.1606	1 | a
0.157143	-14.1607	of | forms.
0.13	-14.1607	the | room
0.26087	-14.161	furthermore, | system
0.175439	-14.1613	an | nlp
0.0645161	-14.1613	another | in
0.161765	-14.1614	our | differs
0.0678851	-14.1614	that | document
0.0718391	-14.1615	framework | to
0.092233	-14.1615	we | global
0.136842	-14.1616	central | of
0.285714	-14.1617	precision, | f1
0.0162866	-14.1619	was | the
0.357143	-14.1619	scores | 87
0.216216	-14.1621	show | average
0.333333	-14.1621	2. | function
0.106667	-14.1621	and | strategies
0.666667	-14.1622	|σ|
0.666667	-14.1622	tprime(f|e)
0.15	-14.1622	laboratory | of
0.139535	-14.1622	with | work.
0.104294	-14.1623	a | text,
0.10119	-14.1624	existing | for
0.3	-14.1625	array | we
0.3	-14.1625	it | effectively
0.0595238	-14.1625	the | achieve
0.065534	-14.1625	these | to | for
0.0537459	-14.1628	the | expressions
0.4	-14.1629	example | need
0.139535	-14.163	slot | of
0.25	-14.1631	and | spontaneous
0.363636	-14.1631	non-negative | matrix
0.0314223	-14.1631	we | the | training
0.210526	-14.1631	regarding | as
0.272727	-14.1633	framework | also
0.666667	-14.1634	easy to | interface.
0.666667	-14.1634	(simfusion | corank)
0.13	-14.1634	as | produce
0.225806	-14.1635	model, | over
0.6	-14.1636	recognizing | (dagan
0.6	-14.1636	datasets, | well
0.139535	-14.1637	5% | the
0.272727	-14.1638	(e.g., | will
0.233333	-14.1638	than | characters
0.363636	-14.1638	apply | resulting
0.15	-14.1639	forward | of
0.3125	-14.1642	experiments, | obtained
0.0917874	-14.1642	using | where
0.210526	-14.1642	one | quality
0.0189784	-14.1644	to | or
0.0745342	-14.1645	from | test
0.0710227	-14.1647	possible | that
0.6	-14.1647	shutova | (2009)
0.205128	-14.1649	have | towards
0.5	-14.1649	as | figure.
0.0670103	-14.1651	models | from
0.150685	-14.1652	the | substitute
0.233333	-14.1655	set | representing
0.0900474	-14.1655	together | and
0.144578	-14.1655	it | within
0.363636	-14.1658	additional | required
0.0884956	-14.1658	annotators | a
0.0729483	-14.1658	the | retrieval
0.6	-14.1658	x1 | p1
0.6	-14.1658	co, | 1–10
0.6	-14.1658	domain | social_science
0.6	-14.1658	eu | eu
0.6	-14.1658	t/c | conc
0.6	-14.1658	4. | train
0.6	-14.1658	finishing | discussing
0.6	-14.1658	festival | festival
0.6	-14.1658	7.1 | 9.1
0.6	-14.1658	geneva | candolle
0.6	-14.1658	s99 | s32s73s76
0.6	-14.1658	5gm | 4gm
0.6	-14.1658	seeing | seeing
0.6	-14.1658	score(y) | score(y)
0.6	-14.1658	27.9 | 30.2
0.6	-14.1658	propositions | φrev
0.6	-14.1658	+pos | +id
0.6	-14.1658	ga | ga
0.6	-14.1658	janarthanam | lemon
0.6	-14.1658	einen | wie
0.6	-14.1658	ikviens | kas
0.6	-14.1658	wenn | unvermeidlich
0.6	-14.1658	94.4 | 93.7
0.6	-14.1658	85.5 | 79.6
0.6	-14.1658	1n | faust+all
0.6	-14.1658	morgan | 60607,
0.6	-14.1658	rd | f
0.6	-14.1658	belief, | s∗ | differ
0.6	-14.1658	d97 | d94
0.6	-14.1658	j−1summationdisplay | ρkj(1)]+
0.6	-14.1658	ostl1fobos | 3,720m
0.6	-14.1658	segment | ax
0.6	-14.1658	j1 | j2
0.6	-14.1658	(i,j) | ∉
0.6	-14.1658	starbucks, | starbucks
0.6	-14.1658	one | sky.
0.6	-14.1658	b@ | fy
0.6	-14.1658	=∑ | vector
0.6	-14.1658	una | praecurrere
0.6	-14.1658	∪ | condensed
0.6	-14.1658	subcontract | bbn,
0.6	-14.1658	s117 | s111
0.6	-14.1658	sr | 〈wr,fr
0.6	-14.1658	eye | eye
0.6	-14.1658	zhao-eng | engines
0.6	-14.1658	pad | film
0.6	-14.1658	habla | madrid
0.6	-14.1658	endurants | perdurants
0.6	-14.1658	c−→h) | r∈c
0.6	-14.1658	countries, | islands,
0.6	-14.1658	118 | 125
0.6	-14.1658	beetle | electricity
0.6	-14.1658	scanning | switch
0.6	-14.1658	presence | capitalized
0.6	-14.1658	optical | moment
0.6	-14.1658	teamingame(g, | teamfinalscore(g,
0.6	-14.1658	“peaceful | strained,”
0.6	-14.1658	agent(x, | purpose−1(y,
0.6	-14.1658	whq(obj: | attr:
0.6	-14.1658	d107 | d83
0.6	-14.1658	optimized | ceaf
0.6	-14.1658	1.9 | 2.4
0.6	-14.1658	binding | bmp-2
0.6	-14.1658	0.19, | 0.16,
0.6	-14.1658	evaluativity: | m.t.,
0.6	-14.1658	deviant | twsss
0.6	-14.1658	-11 | -4
0.6	-14.1658	0.288 | 0.475
0.6	-14.1658	arts, | aškerčeva
0.6	-14.1658	smo | smo
0.6	-14.1658	nan | malad
0.6	-14.1658	ψ | (ψ
0.6	-14.1658	imagistic | object
0.6	-14.1658	benign | nephrosclerosis
0.6	-14.1658	couti,j,r | λi,j,r
0.6	-14.1658	74.1 | −best-ratio
0.6	-14.1658	(smrž, | functional
0.6	-14.1658	p(r6 | r3null
0.6	-14.1658	segment | travaux
0.6	-14.1658	markable | calculation
0.6	-14.1658	10% | 13.5%
0.6	-14.1658	cyy | cxy
0.666667	-14.1659	similarity | full-dimensional
0.666667	-14.1659	than | not–shared
0.3	-14.1659	allophones/phoneme | acc
0.233333	-14.166	methods | most
0.163934	-14.1662	(chiang, | al.,
0.0712251	-14.1662	arguments | and
0.0585938	-14.1662	in | part
0.222222	-14.1663	university | cambridge
0.285714	-14.1664	to | (2005).
0.272727	-14.1664	pairs | distance
0.169492	-14.1665	examine | for
0.195652	-14.1665	case, | which
0.2	-14.1665	4 | total
0.0729483	-14.1666	is | head
0.119048	-14.1667	for | forms
0.5	-14.1667	extracting | opinions
0.5	-14.1667	efforts | discourse
0.5	-14.1667	or | synonymous
0.5	-14.1667	sense | necessary
0.216216	-14.1669	features are | used
0.272727	-14.1669	w, | with
0.3125	-14.167	words, | tags,
0.13	-14.167	summing | the
0.0318363	-14.167	the | from | from
0.222222	-14.1671	sentence | tags
0.175439	-14.1672	in | 10%
0.0627803	-14.1673	a | features.
0.272727	-14.1674	model | summary
0.127273	-14.1674	supported | national
0.285714	-14.1675	steve | and
0.3125	-14.1676	ter | -
0.15493	-14.1677	of | mert
0.144578	-14.1677	and | removed
0.107383	-14.1678	better | by
0.0178508	-14.1678	this | for
0.0836653	-14.168	of | special
0.183673	-14.1682	by | potential
0.666667	-14.1683	nt | gme
0.666667	-14.1683	92.0 | 5:
0.15493	-14.1683	moreover, | with
0.333333	-14.1683	oi | oj
0.0433862	-14.1683	the | events
0.108108	-14.1684	7 | 0
0.109589	-14.1686	degree | a
0.4	-14.1688	ritter | al.
0.169492	-14.169	to be | source
0.122807	-14.169	was | its
0.285714	-14.1691	are | expressions.
0.046683	-14.1691	can | an
0.107383	-14.1692	the | spaces
0.127273	-14.1693	structure | an
0.666667	-14.1696	textual | phenomena.
0.4	-14.1696	low | indicate
0.139535	-14.1697	all | n
0.070028	-14.1698	of | techniques
0.175439	-14.1698	1) | by
0.0650602	-14.1698	and | even
0.363636	-14.1699	→ | ↔
0.444444	-14.1699	index | refers
0.0383333	-14.1699	this | model
0.122807	-14.1699	• | such
0.222222	-14.17	have | maximum
0.0268839	-14.17	we | word
0.12069	-14.1701	which | time
0.333333	-14.1702	et | 1990),
0.26087	-14.1702	experiments | various
0.4	-14.1703	3), | 4).
0.4	-14.1703	problem | translation.
0.178571	-14.1705	in | similarity.
0.216216	-14.1705	dependency | manning,
0.0865801	-14.1705	this | improve
0.0989011	-14.1706	all | feature
0.2	-14.1706	+ | (3)
0.116279	-14.1708	took | of
0.666667	-14.1708	pta(f|e)
0.216216	-14.1709	2 | 77
0.157143	-14.1709	since | text
0.115385	-14.171	on | events
0.4	-14.171	generating | reordering
0.225806	-14.1711	translations | sentence
0.0865801	-14.1712	of | 2005).
0.0785714	-14.1713	bayesian | the
0.0255381	-14.1714	because | the
0.2	-14.1714	between | predicted
0.5	-14.1714	a72a72a8a8a8 | a90a90a26a26
0.191489	-14.1715	using | manually
0.210526	-14.1717	in | vs
0.169492	-14.1719	sum | =
0.363636	-14.1719	number | sites
0.233333	-14.172	novel | data
0.139535	-14.172	idea | from
0.666667	-14.1721	sadeep | excellatach
0.666667	-14.1721	behind | vocabulary
0.205128	-14.1721	however, | must
0.26087	-14.1721	0.89 | 0.96
0.106667	-14.1722	many | can
0.152778	-14.1723	2.1 | as
0.066838	-14.1723	presented | to
0.5	-14.1723	was | scholarship
0.5	-14.1723	beyond | existing
0.0935961	-14.1724	and | determine
0.136842	-14.1725	labels | from
0.241379	-14.1726	approach | tasks
0.114504	-14.1726	had | that
0.285714	-14.1728	coreference | new
0.0486486	-14.1728	class | of
0.216216	-14.1729	framework | data
0.26087	-14.1731	were | reading
0.225806	-14.1732	choose | not
0.0785714	-14.1732	• | word
0.5	-14.1732	german, | chinese
0.0341207	-14.1733	it is | and
0.4	-14.1733	data, | translated
0.0627803	-14.1733	sentiment | a
0.123894	-14.1733	a | β
0.0450867	-14.1734	pos | and
0.0820312	-14.1735	decided | the
0.18	-14.1735	an | recognition
0.172414	-14.1736	it | achieve
0.210526	-14.1736	relative | two
0.136842	-14.1737	we | occurring
0.357143	-14.1738	mann | nach
0.0645933	-14.1738	and | level
0.117188	-14.1741	(2010) | with
0.5	-14.1742	included | additional
0.5	-14.1742	articles | 2008).
0.161765	-14.1743	1 | examples
0.146341	-14.1743	contributions | to
0.111111	-14.1743	on | through
0.285714	-14.1744	target | bleu
0.666667	-14.1746	messages, | feeds,
0.666667	-14.1746	e | [i|λ2],
0.0774411	-14.1746	is | query
0.10596	-14.1748	of | examples.
0.0308609	-14.1749	since | of
0.225806	-14.1749	/ | t | c
0.25	-14.1751	on | dual
0.0797101	-14.1752	underlying | the
0.18	-14.1752	affective | and
0.0463415	-14.1753	to | human
0.0808824	-14.1754	for | linguistic
0.222222	-14.1754	• | text
0.0826772	-14.1754	is | word.
0.0559567	-14.1755	to | like
0.115385	-14.1755	for | :
0.0817121	-14.1757	a | morphological
0.225806	-14.1757	fronted | the
0.195652	-14.1757	distance | words
0.10241	-14.1758	extracted | by
0.222222	-14.1758	on | arguments.
0.0972973	-14.1758	including | as
0.241379	-14.1758	learned | data.
0.241379	-14.1758	similar | human
0.333333	-14.1758	between | classes.
0.169492	-14.176	selection | training
0.108108	-14.1761	shows | which
0.10596	-14.1762	this | particular
0.4	-14.1763	annotation | reliability
0.0279379	-14.1765	in | training
0.114504	-14.1767	therefore, | that
0.666667	-14.177	pivot | trg
0.4	-14.177	use | relationship
0.0334385	-14.177	the | less
0.172414	-14.1771	is | describes
0.157143	-14.1771	information | how
0.123894	-14.1772	knowledge, | is
0.175439	-14.1773	used | error
0.0913462	-14.1773	by | following
0.166667	-14.1774	up | each
0.0952381	-14.1774	problem | from
0.0432034	-14.1776	study | the
0.233333	-14.1777	del | a
0.0972973	-14.1778	entire | and
0.0659898	-14.1779	phrase | as
0.0298281	-14.1779	a | of | each
0.1	-14.178	every | with
0.272727	-14.1781	error | shown
0.0630631	-14.1781	functions | the
0.13	-14.1781	this | .
0.063981	-14.1782	relations | a
0.0355366	-14.1782	the | n
0.444444	-14.1782	3.5 | ips
0.444444	-14.1782	vertices | null2.
0.444444	-14.1782	cls | vst
0.444444	-14.1782	.95 | .87
0.444444	-14.1782	ψi | αgi
0.444444	-14.1782	quantity | bone
0.148649	-14.1783	distance | that
0.11811	-14.1783	importantly, | the
0.115385	-14.1785	with | likely
0.0673575	-14.1785	using | can
0.0379225	-14.1789	are | with | and
0.26087	-14.1789	high | well
0.195652	-14.179	then, | =
0.122807	-14.1791	where | document
0.210526	-14.1791	present | any
0.0663265	-14.1792	is | out
0.0862069	-14.1792	very | with
0.3125	-14.1793	include | 2009),
0.363636	-14.1794	+ | (1
0.0441501	-14.1795	that | was
0.666667	-14.1795	“hat | gelesen
0.12069	-14.1796	model | scores
0.0931373	-14.1797	the | interpreted
0.195652	-14.1797	2 | several
0.0813953	-14.1797	by | lexical
0.210526	-14.1799	multiple | may
0.080292	-14.18	have | other
0.4	-14.18	overlap, | are
0.0605428	-14.18	while | on
0.121739	-14.1801	the | well,
0.195652	-14.1801	is | speed
0.127451	-14.1801	short | is
0.183673	-14.1802	van | of
0.0677083	-14.1802	for | search
0.0843373	-14.1804	1 | +
0.272727	-14.1806	m., | and
0.205128	-14.1806	model | algorithm.
0.110345	-14.1806	6 | by
0.666667	-14.1808	≥ | (desr: | mst:
0.666667	-14.1808	up | allowing
0.666667	-14.1808	imi(w,d|k)
0.0329193	-14.1808	the | extraction
0.363636	-14.1808	value | total
0.333333	-14.1808	forward | which
0.0568807	-14.1809	the | of | translations
0.104575	-14.181	an | i
0.040146	-14.1812	we also | of
0.0900474	-14.1813	approach | our
0.0160047	-14.1814	the | that | is
0.061674	-14.1814	the | randomly
0.272727	-14.1816	common | pairs
0.152778	-14.182	with | users
0.0932642	-14.182	fact, | the | of
0.666667	-14.1821	questions | queried
0.666667	-14.1821	partitioning | partitioning
0.233333	-14.1821	scores | obtained
0.108108	-14.1822	sampled | the
0.110345	-14.1822	intuition | to
0.172414	-14.1826	the | key,
0.10303	-14.1826	identify | as
0.0135304	-14.1827	that | is
0.157143	-14.1827	of | selection.
0.363636	-14.1828	each | w2
0.1	-14.1829	in | entailment
0.4	-14.183	30% | 0%
0.163934	-14.1833	results | many
0.333333	-14.1833	low | accuracy
0.0142218	-14.1834	as | to
0.0342742	-14.1835	the | node
0.0877193	-14.1835	by | syntactic
0.216216	-14.1838	6 | average
0.0143323	-14.1838	the | data
0.0807692	-14.1839	for | c
0.0653753	-14.1839	a | y
0.0714286	-14.184	the | pairs,
0.10303	-14.1841	here, | to
0.163934	-14.1845	2 | distribution
0.0653753	-14.1845	number | which
0.2	-14.1845	pairwise | as
0.666667	-14.1846	extracting | mwes
0.666667	-14.1846	pi(u|m|,
0.175439	-14.1846	score | 2
0.363636	-14.1849	missing | link
0.0877193	-14.185	for | unsupervised
0.0458937	-14.185	to | possible
0.191489	-14.1852	lexical | context
0.108844	-14.1852	are | thus
0.4	-14.1853	spelling | errors,
0.0607966	-14.1853	extract | of
0.123894	-14.1854	we | the | problem.
0.0525478	-14.1854	the | occur | the
0.5	-14.1854	requires | compute
0.5	-14.1854	score | 3).
0.25	-14.1855	on | rare
0.135417	-14.1855	edges | are
0.117188	-14.1856	a | types.
0.0478088	-14.1856	is | first
0.666667	-14.1858	(which | select
0.0435256	-14.1858	words | the | word
0.0517504	-14.1859	the | of | obtained
0.123894	-14.1861	of | learning,
0.0926829	-14.1862	chosen | of
0.107383	-14.1863	on | set | on
0.3125	-14.1864	weight | assigned
0.108844	-14.1864	approach | each
0.0972973	-14.1866	use | source
0.15942	-14.1868	example: | and
0.045728	-14.1868	we | dependency
0.144578	-14.1869	method | any
0.083004	-14.187	each | target
0.333333	-14.187	models | robust
0.666667	-14.1871	removed | tweets.
0.5	-14.1873	labels | record
0.5	-14.1873	visited | first
0.3125	-14.1875	among | relations
0.139535	-14.1875	with | mt
0.10119	-14.1876	system | into
0.241379	-14.1876	used | conll
0.0746753	-14.1877	feature | each
0.106667	-14.1877	the | trees,
0.285714	-14.1878	contain | semantic
0.135417	-14.1878	we | z
0.294118	-14.1879	discourse | holding
0.114504	-14.1879	the | prepositions
0.0498575	-14.188	with | training
0.3125	-14.1881	correlation | reported
0.152778	-14.1883	word | output
0.272727	-14.1883	list | words.
0.25	-14.1883	most | high
0.142857	-14.1883	are | systems.
0.0491573	-14.1884	since | we
0.0588235	-14.1885	of | us
0.0589431	-14.1886	8 | of
0.146667	-14.1886	workshop | statistical | translation,
0.0445205	-14.1887	a | order
0.0391111	-14.1887	the | of | tree
0.0568182	-14.1888	we | no
0.0607375	-14.1889	is | event
0.0804598	-14.1891	by | some
0.0686813	-14.1892	that | only | in
0.10596	-14.1893	as | user
0.666667	-14.1896	|f|
0.0574713	-14.1896	out | a
0.0404135	-14.1896	has | as
0.333333	-14.1896	novelty | is
0.117647	-14.1897	reviews | to
0.15493	-14.1898	(y | (y
0.0673575	-14.1898	a | automatic
0.0548885	-14.1901	the | represented
0.15942	-14.1901	this | change
0.0218579	-14.1903	models | the
0.0293241	-14.1903	most | of
0.146341	-14.1905	----- | ----
0.0575816	-14.1906	natural | of
0.0698324	-14.1906	use | this
0.0587045	-14.1907	this | new
0.103896	-14.1909	highest | is
0.157143	-14.191	have | good
0.157143	-14.191	r | all
0.285714	-14.1911	seem | more
0.0560579	-14.1911	method | on
0.5	-14.1911	reusing | can
0.0877193	-14.1912	in | wikipedia
0.26087	-14.1912	they were | or
0.047014	-14.1914	a | input
0.122807	-14.1914	could | such
0.225806	-14.1915	consistently | on
0.25	-14.1916	approach | during
0.3125	-14.1917	which | place
0.111111	-14.1917	in | 2. | are
0.117647	-14.1919	verbs | with
0.183673	-14.192	bd | cs
0.195122	-14.192	this | (which
0.666667	-14.1921	k|x
0.0557554	-14.1922	role | and
0.166667	-14.1925	vp | nn
0.100592	-14.1925	is | systems.
0.118644	-14.1926	of | positions
0.141176	-14.1926	work | some
0.233333	-14.1927	without | between
0.191489	-14.1928	of | parts.
0.3125	-14.1929	(5) | probability
0.163934	-14.1929	if | reference
0.1875	-14.193	to | form,
0.195122	-14.1931	every | ∈
0.285714	-14.1932	these | produced
0.225806	-14.1932	expected | this
0.136364	-14.1932	in | categories.
0.26087	-14.1932	coherence | topic
0.666667	-14.1934	alignment | ter↓
0.666667	-14.1934	g(i−1)a|n|
0.666667	-14.1934	increased | 4).
0.136364	-14.1935	information | also
0.0718563	-14.1936	the | sampling
0.0686813	-14.1938	line | of
0.0746753	-14.1938	therefore, | and
0.195122	-14.1938	(hall | al.,
0.0913462	-14.1939	time | with
0.5	-14.194	symbol | side
0.0545145	-14.1943	expressed | the
0.117647	-14.1944	it | improve
0.4	-14.1944	aravind | and
0.0535117	-14.1945	computed | of
0.117647	-14.1946	most | information
0.083682	-14.1946	annotation | on
0.10119	-14.1947	search | by
0.0611354	-14.1947	the | except
0.0567108	-14.1947	are | these
0.103226	-14.1947	were | test
0.2	-14.1947	encourage | of
0.225806	-14.1949	finally, | performance
0.241379	-14.1949	remove | words
0.1875	-14.1951	new | over
0.0255543	-14.1951	were | and
0.0909091	-14.1952	utterance | of
0.363636	-14.1952	fold | test
0.225806	-14.1953	but | translations
0.0785714	-14.1953	note that | for
0.0854701	-14.1954	fixed | of
0.150685	-14.1955	knowledge, | a
0.0942408	-14.1956	a | effective
0.116279	-14.1958	english | or
0.0488827	-14.1958	the | for | 2011.
0.0200415	-14.1959	the | test
0.0540541	-14.1959	their | as
0.241379	-14.1959	has | identify
0.241379	-14.1959	evaluated | bleu
0.4	-14.1959	estimated | main
0.112782	-14.1959	shown | this
0.363636	-14.1959	category | ambiguous
0.666667	-14.196	results, | particularly
0.18	-14.196	evaluation | output
0.0590631	-14.196	idea | to
0.272727	-14.1961	future work, | be
0.163934	-14.1961	probability | sequence
0.108108	-14.1961	then | where
0.166667	-14.1963	0.7 | 1
0.077193	-14.1963	and | contain
0.2	-14.1966	components | which
0.0475561	-14.1966	words | by
0.146667	-14.1967	which | average
0.142857	-14.1967	thanks | for
0.169492	-14.1968	word | tokens
0.103896	-14.1968	the | ordered
0.5	-14.1968	need | remove
0.21875	-14.1969	about | dialogue
0.0663265	-14.197	that | event
0.2	-14.197	have | wikipedia
0.16129	-14.197	rules | tree
0.233333	-14.1971	introduced | task
0.210526	-14.1971	on | journal
0.0426439	-14.1972	is | probability | the
0.137931	-14.1972	attempts | a
0.18	-14.1973	use | treebank
0.122807	-14.1973	is | estimated
0.0528926	-14.1975	1 introduction | and
0.0620843	-14.1976	in | query
0.121739	-14.1979	latent | is
0.191489	-14.1979	concerns | and
0.0918367	-14.1979	= | (3)
0.132653	-14.198	an | three
0.205128	-14.1981	they | common
0.126214	-14.1983	work | related
0.210526	-14.1983	training | loss
0.294118	-14.1985	utterance | speech
0.233333	-14.1985	suggests | these
0.0380711	-14.1985	and | which | a
0.09375	-14.1985	of | easily
0.0712166	-14.1985	the | english.
0.121739	-14.1986	syntactic | features
0.0345066	-14.1986	the | bleu
0.195122	-14.1987	can | next
0.5	-14.1987	tagging, | recognition
0.0727848	-14.1987	and | reordering
0.191489	-14.199	instances | two
0.0826772	-14.1992	np | in
0.18	-14.1994	reordering | using
0.142857	-14.1995	this | discriminative
0.111111	-14.1995	example, | may
0.294118	-14.1997	results | settings
0.333333	-14.1997	by | dependencies.
0.666667	-14.1998	discussed | incorporate
0.0982659	-14.1998	at | only
0.0927835	-14.1999	vertex | the
0.183673	-14.2001	we | nps
0.0749186	-14.2002	the | visual
0.121739	-14.2002	of | transfer
0.169492	-14.2004	other | types
0.150685	-14.2004	they | lexical
0.0741935	-14.2004	in | main
0.333333	-14.2004	2) | have
0.0777385	-14.2005	of | appropriate
0.139535	-14.2005	method | corpus
0.15873	-14.2005	such | improve
0.11194	-14.2007	for | whose
0.294118	-14.2008	rule | 
0.666667	-14.2011	p(s|t),
0.0492264	-14.2012	the | v
0.294118	-14.2013	model | constrained
0.15493	-14.2013	results | parser
0.087963	-14.2013	of | parsers
0.0265414	-14.2013	by | that
0.0697674	-14.2014	and | head
0.034975	-14.2014	s | the
0.0164396	-14.2016	a | which
0.0467762	-14.2016	dialogue | and
0.333333	-14.2016	may | frequent
0.5	-14.2016	try | add
0.3125	-14.2018	as | compounds
0.1875	-14.2019	11 | 3
0.666667	-14.2023	people | diligence
0.191489	-14.2023	the | papers,
0.0648379	-14.2024	• | that
0.101796	-14.2027	in | cost
0.0754098	-14.2029	improvement | in
0.0829876	-14.203	the | recall.
0.294118	-14.203	agreement | measured
0.0513219	-14.203	a | search
0.103896	-14.2032	noisy | and
0.088785	-14.2032	human | are
0.0658228	-14.2035	plan | the
0.210526	-14.2035	re | e
0.666667	-14.2036	asia | ling
0.333333	-14.2036	can | nonterminals
0.205128	-14.2036	some | generated
0.21875	-14.2036	also | comments.
0.141176	-14.2036	can | up
0.1875	-14.2037	used | 2005;
0.0469361	-14.2038	5 | to
0.0630841	-14.2038	on | they
0.132653	-14.2038	they | sentence
0.0833333	-14.2039	on | structure
0.106667	-14.2039	however, | use
0.131313	-14.2041	left | for
0.115702	-14.2041	noisy | a
0.0646766	-14.2042	that | can | with
0.363636	-14.2043	null: | in
0.0632319	-14.2043	generation | of
0.233333	-14.2047	facial | &
0.26087	-14.2047	hierarchical | but
0.127451	-14.2048	in | others
0.0798479	-14.2048	from | most
0.666667	-14.2049	1.0|tweets
0.191489	-14.2049	decoder | al.,
0.0675676	-14.2049	we want | the
0.272727	-14.205	of | park
0.25	-14.205	something | as
0.4	-14.2051	due | missing
0.205128	-14.2051	that | arcs
0.0604752	-14.2051	gold | and
0.10596	-14.2053	tools | to
0.272727	-14.2055	my | you
0.163934	-14.2055	(barzilay | and
0.0251386	-14.2055	in | as | in
0.233333	-14.2056	plastic | in
0.191489	-14.2056	translation | additional
0.0551601	-14.2057	like | in
0.0749186	-14.2058	similar | are
0.25	-14.206	by | type,
0.128713	-14.2061	set of | these
0.333333	-14.2061	theme(grass, | fork,
0.666667	-14.2062	|g|
0.118644	-14.2063	number | words | to
0.0826446	-14.2064	from | any
0.0753425	-14.2064	two | this
0.103226	-14.2064	run | a
0.166667	-14.2065	can | models.
0.127451	-14.2065	string | as
0.4	-14.2067	should | trigger
0.0389381	-14.2068	the | p
0.137931	-14.2068	of | 10%
0.210526	-14.2068	rich | on
0.09375	-14.2069	results | best
0.0904762	-14.207	confidence | of
0.0957447	-14.207	system | only
0.294118	-14.207	system | quality.
0.166667	-14.2072	has | relations
0.115702	-14.2072	only | test
0.26087	-14.2072	there are | pairs
0.0900474	-14.2073	feature | training
0.0725552	-14.2076	the | combinations
0.26087	-14.2077	their | uses
0.0663265	-14.2078	we then | of
0.191489	-14.2078	for | direction
0.0977011	-14.2079	match | to
0.0994152	-14.208	c©2011 association for computational linguistics | language
0.152778	-14.2082	mstparser | the
0.233333	-14.2083	cost | which
0.15873	-14.2084	are | characters
0.363636	-14.2085	to | associations.
0.0721003	-14.2086	actual | the
0.13	-14.2086	and | fragments
0.333333	-14.2087	data | years
0.333333	-14.2087	extract | features.
0.173077	-14.2087	related | also
0.087963	-14.2087	by | pairs
0.666667	-14.2088	p(lscripti|gti)
0.666667	-14.2088	p(d|h)
0.146667	-14.2088	have | make
0.152778	-14.2091	that | takes
0.205128	-14.2091	2011; | and
0.21875	-14.2091	semantic | text.
0.125	-14.2091	performance | was
0.10596	-14.2092	this | shared
0.0610022	-14.2095	a | speech
0.225806	-14.2096	during | at
0.195122	-14.2096	if | domain
0.0829876	-14.2097	target | by
0.0854701	-14.2099	of | trigger
0.0559701	-14.2099	will | with
0.113636	-14.21	observations | of
0.101911	-14.21	n | that
0.666667	-14.21	surprise, | prototypical
0.152778	-14.21	we | output.
0.112782	-14.21	meteor | and
0.3125	-14.2103	present | list
0.150685	-14.2103	referring | to
0.272727	-14.2103	parse | gold
0.0847458	-14.2103	as | (e.g.,
0.26087	-14.2103	3.1 | different
0.4	-14.2105	1 | 1993).
0.4	-14.2105	on | (coling
0.15873	-14.2106	as | probabilistic
0.0214823	-14.2108	is a | of
0.26087	-14.2108	tools | such
0.3125	-14.2109	larger | search
0.0947368	-14.2109	where | have
0.0892019	-14.2109	that | outperforms
0.0746753	-14.2109	we | maximum
0.0923077	-14.2109	translation | sentence
0.2	-14.2109	terms | one
0.0932642	-14.211	as | alignment
0.116667	-14.211	to understand | the
0.666667	-14.2113	100 | 125
0.666667	-14.2113	strategies | presence
0.0795455	-14.2116	sentence | it
0.272727	-14.2118	unsupervised | large
0.176471	-14.212	our | cannot
0.0913706	-14.212	number | was
0.0627907	-14.212	is | part
0.3125	-14.2121	integrate | information
0.1875	-14.2121	would | most
0.0923077	-14.2122	to | utterance
0.101266	-14.2124	(9) | the
0.21875	-14.2125	view | be
0.0483425	-14.2125	in | 2.
0.666667	-14.2126	nara | ikoma,
0.195122	-14.2126	semantic | without
0.294118	-14.2127	(4) | k
0.363636	-14.2127	paraphrase | pairs.
0.0645161	-14.2128	where | word
0.148649	-14.2129	now | from
0.0699708	-14.2133	the | review
0.0349501	-14.2133	score | the | of
0.148649	-14.2134	each | particular
0.0892019	-14.2137	is | distance
0.0289216	-14.2137	the | should
0.0725552	-14.2137	based | or
0.225806	-14.214	models | phrase-based
0.10596	-14.2141	6.2 | the | of
0.210526	-14.2141	performance | achieved
0.5	-14.2142	simple | update
0.121739	-14.2143	they | features
0.0730159	-14.2143	have | more
0.112782	-14.2144	summationdisplay | +
0.0854701	-14.2145	we | useful
0.115702	-14.2146	if | n
0.0896226	-14.2146	l | r
0.163934	-14.2147	base | by
0.225806	-14.2149	set | labels.
0.294118	-14.215	4 | ranking
0.122642	-14.215	our | including
0.233333	-14.2151	language | spoken
0.2	-14.2151	from | smt
0.183673	-14.2153	estimate | model
0.0619266	-14.2153	of | senses
0.190476	-14.2153	parser | into
0.100592	-14.2154	on | small
0.18	-14.2154	semantic | class
0.428571	-14.2155	0.76 | 0.90
0.428571	-14.2155	≥others | pairwise
0.103226	-14.2156	a | give
0.3125	-14.2157	2nd | scotland,
0.11194	-14.2158	should | an
0.0942408	-14.2158	examples | that
0.15625	-14.216	about | they
0.0268313	-14.2161	semantic | of
0.116667	-14.2162	trees | by
0.127451	-14.2164	first | while
0.210526	-14.2166	- | bacterium
0.0480769	-14.2167	the | existing
0.0904523	-14.2167	inside | the
0.0545775	-14.2167	tasks | the
0.106667	-14.2167	l | =
0.109489	-14.2168	as | binary
0.428571	-14.2171	bilingual | bitext
0.0458488	-14.2171	as | other
0.141176	-14.2173	of | directions
0.3125	-14.2175	, | 
0.25	-14.2176	cosine | word
0.21875	-14.2176	define | ∈
0.0745763	-14.2177	to | initial
0.0741935	-14.2178	is | either
0.163934	-14.2179	crf | a
0.21875	-14.218	6 | first
0.0513219	-14.2181	problem | in
0.241379	-14.2182	information | documents.
0.0597015	-14.2182	train | and
0.363636	-14.2184	splitting | knight,
0.21875	-14.2184	first, | information
0.136364	-14.2184	each | sense
0.115702	-14.2184	have | new
0.0668449	-14.2187	to | under
0.2	-14.2187	available | information
0.0541958	-14.2188	for | 2
0.150685	-14.2188	estimate | =
0.16129	-14.2188	authorship | and
0.117647	-14.2189	all | are | on
0.131868	-14.219	considered | it
0.15493	-14.219	probability | will
0.4	-14.2191	identifying | syntactic
0.363636	-14.2191	graph | precision
0.363636	-14.2191	parser | cfg
0.109489	-14.2191	approach | only
0.272727	-14.2193	inter-annotator | was
0.428571	-14.2196	+ | x(q)j
0.169492	-14.2197	coverage | that
0.210526	-14.2199	significant | but
0.121739	-14.22	for | smt
0.0863636	-14.2202	this | linguistic
0.0843882	-14.2202	c | t
0.131868	-14.2203	formulation | of
0.272727	-14.2203	used | parsing.
0.18	-14.2203	are | lack
0.666667	-14.2204	message | (muc)
0.333333	-14.2204	sets | relations
0.15493	-14.2205	results | full
0.428571	-14.2205	detection | i.e.
0.428571	-14.2205	positive | moves
0.428571	-14.2205	among | find
0.065	-14.2206	has | an
0.0826446	-14.2206	for | instance
0.294118	-14.2207	similarly, if | is
0.136364	-14.2208	independent | in
0.0813008	-14.2209	the | em
0.144737	-14.2209	name | with
0.0388437	-14.2211	sentences | a
0.272727	-14.2214	original | context
0.0843882	-14.2214	certain | and
0.233333	-14.2215	type | class
0.152778	-14.2216	feature | pos
0.0319277	-14.2216	different | in
0.666667	-14.2217	we | αtrans
0.112782	-14.2217	annotators | for
0.0399619	-14.2217	is | semantic
0.0994152	-14.2219	however, | all
0.045122	-14.222	most | for
0.428571	-14.2222	tagged | neutral.
0.428571	-14.2222	readability | intended
0.428571	-14.2222	top | 15
0.103896	-14.2224	the | moves
0.233333	-14.2224	cannot | but
0.127451	-14.2224	of | nouns,
0.183673	-14.2225	argued | and
0.0570866	-14.2227	in | out
0.115702	-14.2227	are | larger
0.107914	-14.2228	in | baseline.
0.0240715	-14.2229	a | using
0.3125	-14.223	indicates | dependency
0.666667	-14.223	recipient, | sender
0.5	-14.223	emotional | prosodic
0.0977011	-14.2237	z | of
0.114754	-14.2238	count | for
0.428571	-14.2238	varying | entailment
0.428571	-14.2238	translation | resource-limited
0.142857	-14.2238	5 0 | 0
0.225806	-14.2242	words | share
0.3125	-14.2243	how | estimate
0.666667	-14.2243	(e|x),
0.666667	-14.2243	pr(p|d)
0.0479452	-14.2244	of | user
0.21875	-14.2244	its | child
0.142857	-14.2247	learning | semantic
0.15625	-14.2248	but | similar
0.0892019	-14.2249	we have | an
0.137931	-14.225	human | system
0.272727	-14.2252	amount | social
0.0538194	-14.2252	applied | of
0.0573123	-14.2252	in | search
0.0418848	-14.2253	the | local
0.0944444	-14.2254	perform | as
0.225806	-14.2255	since | search
0.428571	-14.2255	shortest | nodes
0.12069	-14.2257	this | clustering
0.169811	-14.2258	ministry | and
0.225806	-14.2259	even though | on
0.205128	-14.2259	said | .
0.0631068	-14.2261	as | first
0.148649	-14.2262	moreover, | on
0.428571	-14.2264	research | (arl)
0.428571	-14.2264	 | f
0.020175	-14.2264	the | of | this
0.0573123	-14.2264	find | to
0.169811	-14.2265	english | corpus.
0.0582329	-14.2266	we | parser
0.113821	-14.2269	important | from
0.210526	-14.2269	building | by
0.294118	-14.2271	and | marginally
0.0411234	-14.2272	5 | in
0.428571	-14.2272	product | aspects.
0.0743243	-14.2272	it | that | from
0.0582329	-14.2273	with | where
0.2	-14.2273	other | value
0.0476839	-14.2276	and | evaluation
0.0533563	-14.2276	by | between
0.183673	-14.2276	has | recall
0.363636	-14.2276	predict | likely
0.333333	-14.2276	aspects | also
0.100629	-14.2276	reading | and
0.111111	-14.2276	with | particular
0.0829876	-14.2277	for | combination
0.153846	-14.2278	number | annotated
0.136364	-14.2278	• | translation
0.0855856	-14.2278	to | np
0.117647	-14.2279	np | pp
0.5	-14.2279	electronic | database.
0.0566406	-14.228	a | state
0.062954	-14.2281	we | +
0.0756014	-14.2281	require | to
0.428571	-14.2281	candidate | closest
0.0390554	-14.2282	similar | and
0.0265711	-14.2282	an | of | in
0.0318318	-14.2283	of | pairs
0.15873	-14.2283	techniques | been
0.0546448	-14.2285	to | within
0.0613636	-14.2285	the | portion
0.144737	-14.2285	or | within
0.141026	-14.2286	matching | with
0.118644	-14.2287	that | decision
0.0927835	-14.2288	we | similar | in
0.166667	-14.2289	can | process
0.1	-14.2289	into | one
0.173077	-14.229	at | much
0.16129	-14.229	i | one
0.0341226	-14.2292	0 | 1 | 1
0.048433	-14.2293	a | time
0.107914	-14.2293	our algorithm | the
0.09375	-14.2294	as | rather
0.118644	-14.2294	current | are
0.0214345	-14.2295	set | and
0.225806	-14.2295	partially | grant
0.163934	-14.2295	4.1 | used
0.117647	-14.2297	corresponding | from
0.428571	-14.2297	ccg | 2007),
0.428571	-14.2297	backoff | cost
0.173077	-14.2297	make | but
0.5	-14.2298	chang, | christopher
0.0583333	-14.23	a | nodes
0.107914	-14.23	up | with
0.144737	-14.23	method | its
0.0988372	-14.2304	generated | are
0.183673	-14.2305	one | word.
0.148649	-14.2306	with | classes
0.0541958	-14.2307	is | better
0.5	-14.2308	often | 2005).
0.0913706	-14.2312	that | further
0.2	-14.2312	models | conditional
0.0659631	-14.2314	are | given
0.166667	-14.2316	spoken | system
0.294118	-14.2317	together | corresponding
0.195122	-14.2317	with | dependent
0.0542907	-14.2318	terms | in
0.114754	-14.2319	is | proper
0.363636	-14.232	on | numerical
0.0577689	-14.2321	taking | the
0.086758	-14.2321	they | more
0.0609481	-14.2321	each | is | as
0.0617849	-14.2322	and | general
0.0396601	-14.2322	the | of | time
0.3125	-14.2323	will | connected
0.428571	-14.2323	your | i
0.210526	-14.2323	the | was very
0.163934	-14.2324	as | especially
0.25	-14.2324	statements | herein | the
0.0777778	-14.2324	to | resources
0.0454545	-14.2325	possible | a
0.12381	-14.2326	as | significantly
0.112903	-14.2326	function | an
0.166667	-14.2329	bilingual | comparable
0.121495	-14.2332	in | completely
0.12037	-14.2332	vector | which
0.12037	-14.2332	can | domain
0.111111	-14.2333	researchers | al.,
0.16129	-14.2334	0 | 0.2
0.294118	-14.2335	instead, | use
0.0417537	-14.2338	a | t
0.25	-14.2338	used | 2004).
0.173077	-14.2338	most | dependency
0.5	-14.2338	predicate | adjunct
0.15873	-14.2339	using | 4
0.15873	-14.2339	real-world | and
0.428571	-14.234	above | meaning
0.0499244	-14.234	natural | and
0.3125	-14.2341	(dzikovska | al.,
0.166667	-14.2342	we have | features
0.0971429	-14.2343	with | verb
0.152778	-14.2345	as | combined
0.0687679	-14.2345	∈ | a
0.205128	-14.2348	wang | al.
0.21875	-14.2348	would | much
0.130435	-14.2351	linguistic | which
0.24	-14.2353	matching | test
0.153846	-14.2354	technique | et
0.141026	-14.2356	6 | presented
0.428571	-14.2357	or | variation.
0.428571	-14.2357	converted | lower
0.0515298	-14.2359	translated | the
0.15625	-14.236	such | rules
0.121495	-14.2361	4, | to
0.294118	-14.2364	task | comparative
0.2	-14.2364	preprocessing | for
0.0347072	-14.2366	following | of
0.107914	-14.2366	our | sets
0.0431818	-14.2366	that | also
0.126214	-14.2367	and | are both
0.24	-14.2367	the | craft
0.0448485	-14.237	in | how
0.363636	-14.237	matching | hypothesis
0.0993789	-14.2371	the | σ
0.190476	-14.2371	on | blog
0.115702	-14.2371	are | opinion
0.3125	-14.2372	case | negative
0.0626506	-14.2372	and | across
0.122642	-14.2374	matrix | for
0.0657895	-14.2374	features | or
0.0280819	-14.2375	and | features
0.150685	-14.2378	question | it
0.122642	-14.2379	< | =
0.0852018	-14.2379	that | lower
0.0628019	-14.238	our | one
0.4	-14.238	summary | concepts
0.2	-14.238	system | research
0.0521173	-14.2381	4 | we
0.333333	-14.2381	module | s
0.029052	-14.2382	such | in
0.190476	-14.2382	(i.e., | not
0.086758	-14.2383	the | function,
0.087963	-14.2384	of | around
0.0664894	-14.2385	more | it
0.233333	-14.2386	lexicons | by
0.0955056	-14.2387	derived | a
0.195122	-14.2387	labeled | score
0.133333	-14.2388	we | six
0.0589474	-14.2388	sentences | from
0.0373198	-14.239	show | and
0.428571	-14.2391	(top | .))
0.0895522	-14.2394	topic | by
0.0918367	-14.2394	as | search
0.0993789	-14.2395	a | expressed
0.3125	-14.2397	tend | less
0.666667	-14.2402	” | ”,
0.4	-14.2404	better | role
0.150685	-14.2405	we | with | size
0.16129	-14.2405	training data | used
0.12037	-14.2406	on | aspects
0.116667	-14.2406	about | it
0.0949721	-14.2409	to | likelihood
0.0666667	-14.2411	people | the
0.130435	-14.2411	models | data.
0.666667	-14.2415	bio | inside
0.0440476	-14.2416	a new | the
0.142857	-14.2416	structure | each
0.114754	-14.2417	the | agency
0.112	-14.2418	with | speech
0.5	-14.2418	58 | 65
0.272727	-14.2419	train | new
0.24	-14.242	null | null=
0.225806	-14.2421	towards | from
0.144737	-14.2421	this | coverage
0.0763636	-14.2421	bleu | on
0.195122	-14.2422	therefore | have
0.0659631	-14.2423	speech | a
0.0601336	-14.2423	with | over
0.233333	-14.2423	4.3 | features
0.136364	-14.2424	learning | but
0.0939227	-14.2425	mentioned | in
0.428571	-14.2425	trying | find
0.21875	-14.2427	simplification | that
0.166667	-14.2428	judge | sem
0.333333	-14.2428	boy | tall
0.333333	-14.2428	nodes | imported
0.333333	-14.2428	erabaki | 88447
0.333333	-14.2428	xi-word=“hit”, | pmi(“hit
0.333333	-14.2428	rol | ‘role’
0.333333	-14.2428	r• | p•
0.333333	-14.2428	300 | description:
0.333333	-14.2428	rel-lda | type-lda
0.333333	-14.2428	→p | →f1
0.195122	-14.243	classification | label
0.153846	-14.2431	formally, | set
0.141026	-14.2432	3 | its
0.101266	-14.2433	as | content
0.121495	-14.2434	hypotheses | in
0.428571	-14.2434	children | do
0.3125	-14.2435	of | road
0.333333	-14.2435	xx | x
0.0735786	-14.2438	by | most
0.125	-14.2438	from | parser
0.134831	-14.2439	to | style
0.0649351	-14.244	s | =
0.294118	-14.244	input | epoch
0.272727	-14.2441	models | further
0.333333	-14.2441	acknowledgments | feedback
0.0687679	-14.2441	use | set
0.666667	-14.2442	score: | account
0.0730897	-14.2443	: | and
0.4	-14.2444	classifier | accuracy.
0.24	-14.2444	although | syntactic
0.233333	-14.2446	j., | and
0.16129	-14.2447	we | rare
0.142857	-14.2448	for | calculate
0.333333	-14.2448	term | objective
0.0349442	-14.2449	on | or
0.0321564	-14.2449	the | related
0.363636	-14.245	conclude | discussion
0.363636	-14.245	expert | 0.999
0.137931	-14.245	uses | set
0.115702	-14.245	the | corrected
0.148649	-14.2451	and | effectiveness
0.183673	-14.2451	interesting | from
0.428571	-14.2451	approach | 2001).
0.00761147	-14.2452	the | that | the
0.0460358	-14.2452	case, | the
0.109489	-14.2453	because | this
0.666667	-14.2455	abu-jbara | amjbara@umich.edu
0.666667	-14.2455	candidates | evaluate
0.0685714	-14.2455	the | 2010
0.119266	-14.2455	of | t,
0.333333	-14.2455	web | instances
0.173077	-14.2455	distance | each
0.117647	-14.2457	then | its
0.0928962	-14.2457	or | where
0.06	-14.2457	small | to
0.294118	-14.2458	recent | unsupervised
0.106383	-14.2458	experiments | set
0.142857	-14.2459	meaning | an
0.428571	-14.246	relatively | within
0.0460948	-14.2461	their | is
0.176471	-14.2461	preference | is
0.277778	-14.2461	probabilistic | syntactic
0.294118	-14.2464	system | dialogues
0.107143	-14.2464	using | sentences
0.363636	-14.2464	shows | clear
0.166667	-14.2465	of | convergence
0.0212303	-14.2465	and | an
0.0231195	-14.2467	and | are | of
0.205128	-14.2467	such | vector
0.0215445	-14.2469	the | are | for
0.212121	-14.2471	table 2 | all
0.212121	-14.2471	the | sβ>
0.110294	-14.2473	we | consistent
0.0726073	-14.2474	data | test
0.043628	-14.2475	is | probability
0.0766423	-14.2475	to | who
0.666667	-14.2482	transitions | matrix.
0.666667	-14.2482	leads | probably
0.0918367	-14.2482	to | estimate
0.277778	-14.2483	learn | over
0.0304225	-14.2484	we | at
0.212121	-14.2484	scores | compared
0.0442055	-14.2484	in | similar
0.0709877	-14.2485	notice | the
0.363636	-14.2486	x2 | x5
0.16129	-14.2486	can | mt
0.25	-14.2488	the | databases.
0.15873	-14.2488	id | a
0.166667	-14.2489	rewritten | the
0.114754	-14.2489	the | entries.
0.102564	-14.249	(4) | a
0.137931	-14.2492	then | such
0.106383	-14.2492	0 | i
0.0641026	-14.2493	instance | and
0.225806	-14.2493	verb | form
0.0809717	-14.2498	following | et
0.100629	-14.2498	however, | words
0.0873786	-14.2499	current | for
0.129032	-14.2501	the | below,
0.115702	-14.2501	by | (see
0.428571	-14.2503	query | web.
0.169811	-14.2505	general | but
0.0495495	-14.2506	we | context
0.131868	-14.2507	which | even
0.176471	-14.2507	using | compute
0.363636	-14.2508	sentences, | extract
0.153846	-14.2508	more | could
0.333333	-14.2508	tree | string.
0.0891089	-14.2509	for | term
0.104895	-14.251	5.2 | we
0.09	-14.251	moreover, | in
0.126214	-14.2511	this section, | our
0.0549451	-14.2511	in | parallel
0.277778	-14.2511	program | under
0.272727	-14.2512	research | 2008;
0.142857	-14.2514	with | decoding
0.0960452	-14.2514	by | time
0.112	-14.2515	score | an
0.0555556	-14.2515	in | and | these
0.0944444	-14.2515	we | paper.
0.0740741	-14.2516	accuracy | with
0.0750853	-14.2517	can | its
0.0971429	-14.2518	a | (in
0.25	-14.2518	0.33 | 0.26
0.5	-14.2518	hypothesis: | less
0.15873	-14.252	we | showing
0.112	-14.2521	are | lower
0.0530822	-14.2521	the | appropriate
0.24	-14.2521	between | actual
0.666667	-14.2522	0.05 | 0.11
0.0668524	-14.2523	better | for
0.363636	-14.2523	link | based
0.0873786	-14.2523	the | model | over
0.190476	-14.2523	show | precision
0.0793651	-14.2524	0 | 4
0.112903	-14.2524	we | implemented
0.195122	-14.2524	lower | other
0.169811	-14.2525	model | modified
0.108696	-14.2526	points | a
0.0516667	-14.2526	paraphrase | the
0.108696	-14.2528	types | which
0.277778	-14.2528	noted | section
0.428571	-14.2529	thus, | desired
0.15625	-14.2529	6 | used
0.363636	-14.253	mutual | log-likelihood
0.0407747	-14.2531	phrase | a
0.0265752	-14.2532	also | a
0.144737	-14.2534	in | ...
0.666667	-14.2536	rights | based.’
0.136364	-14.2536	using | original
0.363636	-14.2538	national | center
0.363636	-14.2538	utterance | 3
0.0844444	-14.2539	the | official
0.106383	-14.2539	the | setting,
0.26087	-14.2541	recall | percentage
0.173077	-14.2542	representations | on
0.0918919	-14.2544	with | values
0.0585774	-14.2545	these | are | a
0.169811	-14.2546	was | on | data
0.176471	-14.2546	one | speech
0.0605381	-14.2547	finally, | in
0.0740741	-14.2547	word | target
0.212121	-14.2548	compute | sentence
0.0787402	-14.2548	and | every
0.25	-14.2549	evaluate | annotation
0.0295064	-14.2549	well | the
0.272727	-14.2551	and | perception
0.225806	-14.2552	reordering | alignment
0.0371308	-14.2552	only | is
0.0418904	-14.2553	of | time
0.25	-14.2554	treebank | (marcus
0.110236	-14.2554	i | p
0.0869565	-14.2554	are | input
0.0766423	-14.2557	that | either
0.186047	-14.2558	has | total
0.205128	-14.2558	combination | over
0.5	-14.2558	introduces | work.
0.166667	-14.2559	4.1 | set
0.294118	-14.2559	adjectives | such
0.25	-14.2559	and | f-score.
0.363636	-14.256	denotes | distribution
0.0486009	-14.256	proposed | a
0.12766	-14.2561	an | multiple
0.076087	-14.2562	with | order
0.0355487	-14.2563	since | a
0.428571	-14.2563	$ | they
0.294118	-14.2564	relation | level
0.0456919	-14.2566	of | correct
0.195122	-14.2567	perform | more
0.0488166	-14.2569	can | using
0.333333	-14.2569	given | features,
0.0674157	-14.257	is | (e.g.,
0.114754	-14.2572	general, | to
0.428571	-14.2572	define | character
0.233333	-14.2573	position | current
0.125	-14.2574	+ | features
0.034789	-14.2575	and | also
0.205128	-14.2575	approaches | 2010).
0.294118	-14.2576	they | wordnet
0.0840708	-14.2576	(2010) | for
0.125	-14.2576	found | more
0.0699088	-14.2577	research | et
0.108696	-14.2577	machine | systems
0.136364	-14.2578	following | words
0.5	-14.2579	c-lda-a | on
0.126214	-14.258	a | log
0.0882353	-14.258	linguistics, | oregon,
0.21875	-14.2581	section | reviews
0.0641026	-14.2582	show | on
0.294118	-14.2582	optimal | similar
0.12037	-14.2583	this | algorithms
0.0447205	-14.2584	the | point | the
0.0840708	-14.2585	from | about
0.038674	-14.2586	to | data.
0.142857	-14.2586	generation | by
0.0165821	-14.2586	the | their
0.131868	-14.2587	for | quite
0.15873	-14.2588	instances | training
0.428571	-14.2589	binary | bigrams
0.176471	-14.2589	trained | classifier
0.176471	-14.2589	does | include
0.333333	-14.2589	noise | spelling
0.0869565	-14.2589	and | effect
0.5	-14.2589	significant | score.
0.166667	-14.259	since | should
0.15873	-14.2591	models | learning
0.133333	-14.2595	twitter | for
0.1375	-14.2595	identification | as
0.277778	-14.2596	the | p(c|e)
0.4	-14.2598	np | .))
0.26087	-14.26	word | dirichlet
0.272727	-14.2601	patterns | dependency
0.110236	-14.2601	way | are
0.233333	-14.2602	clear | between
0.0506329	-14.2603	where | =
0.12037	-14.2603	also | learning
0.0545455	-14.2604	presented | and
0.205882	-14.2605	first | rule
0.130435	-14.2605	as | difficult
0.149254	-14.2606	input: | of
0.166667	-14.2607	when | algorithm
0.428571	-14.2607	did | perform any
0.16129	-14.2607	induces | the
0.0596026	-14.2608	a | v
0.112903	-14.2611	the | kernels
0.225806	-14.2612	also | explore
0.131868	-14.2614	the | profile
0.0652742	-14.2617	and | approaches
0.0418006	-14.2617	we | source
0.163636	-14.2618	1 | because
0.122642	-14.2619	of | θ
0.12381	-14.2619	yields | in
0.18	-14.262	also | other | such
0.149254	-14.2621	is | errors.
0.139241	-14.2622	test | baseline
0.333333	-14.2623	we used | clustering
0.294118	-14.2624	3.0 | on
0.0362903	-14.2624	is | into
0.25	-14.2625	vbd | pp | np
0.163636	-14.2625	baseline, | is
0.21875	-14.2625	bilingual | corpus.
0.15625	-14.2627	scores | best
0.18	-14.2627	lm | test
0.25	-14.263	uses | approach
0.333333	-14.263	tests | training
0.333333	-14.263	similarly | between
0.16129	-14.2633	where | order
0.15625	-14.2636	see | better
0.333333	-14.2637	query | google
0.078125	-14.2639	in | performed
0.15873	-14.2639	corpus | also
0.363636	-14.2641	strategy | improve
0.277778	-14.2641	number | ways
0.190476	-14.2643	each | wikipedia
0.0392344	-14.2646	the | further
0.0861244	-14.2647	j | a
0.0434783	-14.2647	are | one
0.0738255	-14.2648	0.4 | 0.4
0.294118	-14.2648	algorithms | dependency
0.5	-14.265	ku-ka | hay
0.0705521	-14.2651	= | n
0.118182	-14.2652	2. | each
0.225806	-14.2653	she | not
0.26087	-14.2653	ksummationdisplay | =
0.0829694	-14.2655	the | media
0.069697	-14.2655	of | candidates
0.0793651	-14.2656	of | distributional
0.0891089	-14.2657	of | tools
0.18	-14.2657	all | level
0.119266	-14.2658	this | lexicon
0.153846	-14.2658	is | step.
0.195122	-14.2659	(2007) | by
0.151515	-14.2661	for each | phrase
0.0787402	-14.2664	by | text
0.212121	-14.2665	different | alignment
0.0738255	-14.2666	the | back
0.294118	-14.2666	summarized | 3.
0.21875	-14.2666	those | no
0.16129	-14.2666	with | generation
0.190476	-14.2667	also, | with
0.166667	-14.2668	both | level
0.122642	-14.2669	= | equation
0.078125	-14.267	if | sentence
0.333333	-14.2671	polarity | determined
0.294118	-14.2672	show | consistent
0.176471	-14.2672	kim | et
0.0861244	-14.2673	show | using
0.0395371	-14.2674	main | of
0.149254	-14.2676	we | iteration
0.153846	-14.2677	can | (e.g.
0.363636	-14.2678	which | made.
0.333333	-14.2678	perplexities | for
0.18	-14.2679	such | words.
0.21875	-14.2679	result | obtained
0.173077	-14.268	more | just
0.0826087	-14.268	are | respectively.
0.225806	-14.2681	regression | using
0.5	-14.2681	node | degree
0.122642	-14.2684	of | no.
0.666667	-14.2685	reported | poon
0.0891089	-14.2685	the | 0.5
0.0457516	-14.2685	that | may
0.0869565	-14.2687	and | object
0.21875	-14.2688	used | token
0.15873	-14.2688	text | they
0.0869565	-14.269	we | produce
0.110236	-14.2691	trend | the
0.195122	-14.2691	corpus | learner
0.0539568	-14.2696	most | that
0.130435	-14.2696	hand, | on
0.363636	-14.2701	even | another
0.26087	-14.2701	corpus | knowledge
0.0815451	-14.2702	four | for
0.0833333	-14.2702	effect | to
0.0542056	-14.2704	the | complexity
0.4	-14.2704	p(c|e)
0.4	-14.2704	regular | like
0.0758123	-14.2705	and | computed
0.18	-14.2705	table | up
0.176471	-14.2705	twitter | that
0.0650407	-14.2705	they | from
0.25	-14.2707	measure | probability
0.0833333	-14.2707	our | systems
0.195122	-14.2707	compare | their
0.0869565	-14.2708	show | this
0.294118	-14.2709	are computed | all
0.153846	-14.2709	it | straightforward
0.0475524	-14.2709	sentiment | and
0.0703364	-14.271	work | an
0.0586957	-14.271	the | f1
0.0482456	-14.2711	linguistics | university
0.0909091	-14.2712	and | effective
0.428571	-14.2712	spelling | synonyms
0.173077	-14.2712	4 | section,
0.5	-14.2712	idf | word
0.0601852	-14.2714	and | role
0.0441176	-14.2716	of | research
0.186047	-14.2717	instances | feature
0.0944444	-14.2717	for | various
0.3125	-14.272	labels | mturk
0.0703364	-14.2721	an | into
0.428571	-14.2721	setting, | annotated
0.277778	-14.2721	given | loss
0.15873	-14.2721	work | 2010)
0.0470914	-14.2722	should | to
0.0355731	-14.2722	in | test
0.0709677	-14.2722	an | information
0.062201	-14.2724	not | have
0.0705128	-14.2725	overall | in
0.212121	-14.2726	from | representation.
0.0341074	-14.2726	word | as
0.3125	-14.2727	choices | at
0.0263727	-14.2727	feature | the | of
0.0721311	-14.2728	in | had
0.149254	-14.2728	order | first
0.0631313	-14.2729	for | pos
0.212121	-14.273	line | al.,
0.428571	-14.273	stoyanchev | in
0.428571	-14.273	measures | find
0.0678466	-14.273	two | have
0.1375	-14.273	about | our
0.0326232	-14.273	may | to
0.5	-14.2732	site | binding
0.0463215	-14.2732	if | it
0.151515	-14.2733	feature | pair
0.21875	-14.2733	3.2 | use
0.4	-14.2737	(2007), | more
0.363636	-14.2738	overall, | over
0.277778	-14.2738	attribution | texts
0.0928962	-14.2742	dependency | word
0.25	-14.2743	sets | features.
0.105634	-14.2744	we | cosine
0.26087	-14.2744	those | higher
0.333333	-14.2746	sciences | usa
0.169811	-14.2747	that | lexicon.
0.428571	-14.2747	in | county,
0.173077	-14.2747	words, | this
0.21875	-14.2747	d | t | h
0.125	-14.2748	used | order
0.0236275	-14.2748	as | with
0.0728477	-14.2749	11 | the
0.118182	-14.2749	u | is
0.277778	-14.2749	have | power
0.0301822	-14.275	the | vector
0.15873	-14.275	an | expression
0.139241	-14.2751	first | some
0.294118	-14.2751	scores | averaged
0.111111	-14.2751	it | find
0.0676471	-14.2753	the | designed
0.0294599	-14.2753	for | all
0.5	-14.2753	development | iteration.
0.109375	-14.2756	, | i
0.24	-14.2756	prep | nn
0.0833333	-14.2757	vectors | a
0.116071	-14.2757	it | part
0.15873	-14.2757	error | all
0.0631313	-14.2758	using | language
0.3125	-14.2759	positive | 0.85
0.129032	-14.2759	word | sentence.
0.149254	-14.2759	implementation | that
0.230769	-14.2759	baseline | obtained
0.24	-14.2761	possible | rule
0.103448	-14.2763	of | topic.
0.176471	-14.2763	translation | corpora
0.181818	-14.2764	arguments | each
0.135802	-14.2765	50 | 10
0.163636	-14.2766	the | acts.
0.666667	-14.2767	pack | 05/2008.
0.0678466	-14.2767	this paper | on
0.0739437	-14.2768	annotation | that
0.294118	-14.2769	other | temporal
0.25	-14.2769	language | access
0.205882	-14.277	unit | we
0.097561	-14.277	that | texts
0.0588235	-14.277	for | entity
0.181818	-14.2772	any other | of
0.0734266	-14.2774	into | or
0.21875	-14.2774	set | trees.
0.0601852	-14.2774	obtain | of
0.2	-14.2774	is | coherent
0.0324718	-14.2776	test | and
0.144737	-14.2777	present | language
0.4	-14.2778	(relms | usa,
0.125	-14.2779	our | up
0.0442804	-14.278	of | size
0.294118	-14.2781	result | recall
0.24	-14.2781	tr | r
0.110236	-14.2782	training | size
0.0440098	-14.2782	position | the | the
0.102041	-14.2783	by | described
0.428571	-14.2783	)= | t
0.0451613	-14.2783	that | sentence
0.21875	-14.2783	measure | recall
0.0808511	-14.2783	had | a
0.0259574	-14.2784	work | the | of
0.069697	-14.2787	we | framework
0.173077	-14.2787	vector | its
0.21875	-14.2787	test | data,
0.139241	-14.2788	problems | et
0.0449294	-14.279	sets | and
0.129032	-14.2791	a71a71 | a71a71
0.363636	-14.2791	adopted | their
0.0453368	-14.2793	each | or
0.173077	-14.2794	differences | statistically
0.5	-14.2794	0.59 | 0.74
0.0844444	-14.2795	tasks | al.,
0.666667	-14.2795	hr0011-06-c-0022, | inc.
0.109375	-14.2795	task | system
0.0628141	-14.2795	two | not
0.4	-14.2795	single | i.e.
0.24	-14.2796	extracted | sentences.
0.129032	-14.2799	which | speech
0.166667	-14.2802	shown | baseline
0.149254	-14.2802	and | case.
0.230769	-14.2802	social | language
0.190476	-14.2804	• | generated
0.0873786	-14.2806	a | whereas
0.0793651	-14.2807	for | need
0.277778	-14.2807	so, | not
0.0963855	-14.2808	similar | which
0.08	-14.2809	to | parameter
0.4	-14.2811	pairs | sentence,
0.230769	-14.2812	goal | will
0.212121	-14.2814	reviews | by
0.363636	-14.2814	why | do
0.123711	-14.2816	feature | learning
0.144737	-14.2816	is | potentially
0.0662983	-14.2817	to | entities
0.123711	-14.2818	tweet | to
0.21875	-14.2819	our | corpora,
0.277778	-14.2819	for | mistakenly
0.0676471	-14.2824	by | under
0.115044	-14.2825	computed | as
0.15625	-14.2826	english | task
0.103448	-14.2828	make | on
0.4	-14.2828	propbank | role
0.104895	-14.2828	as | sentence.
0.176471	-14.2829	work | computational
0.205882	-14.283	learning | user
0.163636	-14.283	this | incorporate
0.24	-14.2831	since | few
0.0576923	-14.2832	information | from
0.186047	-14.2833	of | sender
0.121495	-14.2833	seed | a
0.131868	-14.2834	1997; | et
0.0439024	-14.2834	in | so
0.0466392	-14.2835	have | from
0.4	-14.2836	vinken | n.v.
0.4	-14.2836	dp | dp
0.4	-14.2836	ˆn | k,nk0
0.4	-14.2836	grammar) | nullnull)
0.4	-14.2836	prp | adv
0.4	-14.2836	a6a4a7 | a10a15a15a9a8a15a10a22a14a16a12
0.4	-14.2836	1,818 | 24.5395
0.4	-14.2836	s(x | :c
0.4	-14.2836	(greece, | 2005)
0.4	-14.2836	a23a26a12a22a27a28 | a32a16
0.4	-14.2836	a17a9a14a16a9 | a22a10a31
0.4	-14.2836	ali | ~
0.4	-14.2836	camera | camera
0.4	-14.2836	s:1.2 | np^
0.4	-14.2836	on; | keep
0.4	-14.2836	season |  -­‐
0.4	-14.2836	leider | ‘leader’
0.4	-14.2836	finish | finish
0.4	-14.2836	xb | half
0.25	-14.2836	sentences | make
0.428571	-14.2836	priors | parameters
0.0237252	-14.2836	if | a
0.2	-14.2836	model | initial
0.0729167	-14.2836	of | paper.
0.5	-14.2836	requires | available
0.666667	-14.2837	stability | (sm)
0.0661157	-14.2837	pattern | of
0.10274	-14.2839	equivalent | of
0.225806	-14.284	suggesting | for
0.0833333	-14.284	of | change
0.2	-14.284	(m) | senses
0.051495	-14.2842	this | their
0.0473458	-14.2842	linguistic | of
0.25	-14.2842	and | infectious
0.181818	-14.2843	we started | the
0.076087	-14.2845	simple | is
0.0923913	-14.2846	into | language
0.21875	-14.2846	introduced | an
0.212121	-14.2849	summarization | document
0.3125	-14.285	mt | then
0.666667	-14.285	energy | phosphor
0.147059	-14.2851	target | aligned
0.0166989	-14.2855	we | by
0.177778	-14.2855	we | (2010).
0.230769	-14.2855	outperforms | over
0.3125	-14.2856	cache | document.
0.101351	-14.2856	5 | section
0.25	-14.2857	previous | described
0.123711	-14.2857	approaches | this
0.066482	-14.2857	of | them.
0.119266	-14.2858	evaluation | at
0.0946746	-14.2858	used | these
0.119266	-14.286	the | increases.
0.0632911	-14.286	total | in
0.190476	-14.286	opinions | or
0.139241	-14.2861	no | has
0.0861244	-14.2861	model | data.
0.294118	-14.2861	between | relation.
0.0398406	-14.2863	have | et al.,
0.126316	-14.2863	by | common
0.0337886	-14.2863	most | to
0.26087	-14.2863	i.e., | at
0.21875	-14.2865	case, | only
0.21875	-14.2865	difference | bleu
0.112	-14.2866	avoid | in
0.122642	-14.2867	which | sense
0.186047	-14.2868	on | annotations.
0.135802	-14.2869	however, | approach
0.0969697	-14.2869	the | colour
0.2	-14.2869	9 | by
0.0524476	-14.287	a | whether
0.0829694	-14.2873	using | different
0.0682493	-14.2873	we show | and
0.0657534	-14.2873	content | to
0.0818966	-14.2874	the | compression
0.230769	-14.2874	system | output.
0.12037	-14.2876	important | this
0.205882	-14.2877	during | can
0.177778	-14.2877	sentence | within
0.666667	-14.2878	pp | mwe
0.666667	-14.2878	standard | algorithms.
0.0805085	-14.288	of | key
0.153846	-14.2881	and | result.
0.153846	-14.2881	context | where
0.24	-14.2881	semantic | sentiment
0.12766	-14.2882	this | done | of
0.102041	-14.2883	experiments | translation
0.21875	-14.2883	acknowledgements | reviewers
0.0981595	-14.2884	of | performing
0.0711974	-14.2885	a | clustering
0.12766	-14.2887	reported | this
0.144928	-14.2887	proposed | information
0.0570825	-14.2888	the | indicate
0.0467033	-14.2889	a | similarity
0.428571	-14.289	despite | learned
0.666667	-14.2892	“ | ��pk�
0.15873	-14.2892	observed | than
0.054902	-14.2893	(2009) | the
0.0946746	-14.2895	were | only
0.294118	-14.2898	considered | correct
0.033239	-14.2898	of | phrase
0.0238525	-14.29	the | are | with
0.190476	-14.29	experiments, | training
0.10274	-14.2901	of | obtained by
0.190476	-14.2904	as | paper,
0.106061	-14.2904	model | new
0.105634	-14.2905	remains | the
0.333333	-14.2906	of | l2.
0.24	-14.2906	features | component
0.428571	-14.2908	move | mentions
0.106061	-14.2908	wikipedia | as
0.15873	-14.2909	research | program
0.112903	-14.291	that | detection
0.0739437	-14.2912	to | experiments
0.0585683	-14.2913	a | content
0.0653951	-14.2913	count | of
0.333333	-14.2913	shows | distance
0.173077	-14.2913	syntactic | models
0.0481928	-14.2916	the | included
0.0869565	-14.2916	on | event
0.169811	-14.2917	and | title
0.0645161	-14.2917	3.1 | and
0.105634	-14.2918	were | other
0.0885417	-14.2922	of | native
0.047619	-14.2922	by | words
0.103448	-14.2923	accurate | and
0.0423341	-14.2925	the | instance
0.0511945	-14.2925	those | to
0.3125	-14.2928	international | processing
0.363636	-14.2928	meaning | sentence,
0.5	-14.293	matt | helpful
0.166667	-14.2934	had | this
0.064	-14.2936	not | such
0.186047	-14.2938	parser | only
0.125	-14.2938	evaluation | standard
0.149254	-14.2939	shown | translation
0.144928	-14.2939	evaluate | different
0.135802	-14.2941	work | 2009;
0.0604651	-14.2941	that | by | that
0.0894737	-14.2942	of | space.
0.044586	-14.2943	will | for
0.10274	-14.2944	there | several
0.0878049	-14.2944	construction | and
0.0958084	-14.2945	to | seen
0.26087	-14.2945	data, | its
0.205882	-14.2946	w | g
0.0525394	-14.2946	more | by
0.153846	-14.2946	need | more
0.277778	-14.2946	collection | news
0.097561	-14.2947	we | applying
0.15625	-14.2948	average | training
0.0404145	-14.2949	how | to
0.12037	-14.2949	from | instances
0.056338	-14.2949	since | not | the
0.2	-14.2949	denoted | =
0.041621	-14.2949	1 | 3
0.126316	-14.295	called | is
0.123711	-14.295	and | pair.
0.26087	-14.2951	space | will
0.177778	-14.2952	meaning | one
0.126316	-14.2953	measures | between
0.0829694	-14.2957	this | shown
0.225806	-14.2958	to | sub
0.277778	-14.2958	vp | verb
0.0878049	-14.2959	a | pairs.
0.0114856	-14.296	with | of
0.666667	-14.2962	andrei | 19,
0.333333	-14.2962	science | ca
0.051458	-14.2963	i.e., | the
0.0894737	-14.2963	on | news
0.0433213	-14.2963	2 | is
0.195122	-14.2963	four | used
0.1	-14.2964	strong | a
0.0861244	-14.2965	difficult | and
0.0393701	-14.2965	our | not
0.205882	-14.2968	linguistic | data
0.0425777	-14.2968	set | from
0.277778	-14.297	which | above.
0.135802	-14.2972	it | trained
0.0682493	-14.2972	between | an
0.153846	-14.2975	after | been
0.104895	-14.2977	active | in
0.15625	-14.2977	this | match
0.075	-14.2978	same | or
0.0865385	-14.2979	acknowledgments this work | the
0.263158	-14.298	n | instances
0.24	-14.2982	(2006) and | and
0.2	-14.2983	the | endorsement
0.0628141	-14.2984	are | similar
0.0760456	-14.2986	for | k
0.166667	-14.2987	probability | class
0.212121	-14.2987	nodes | only
0.0805085	-14.2989	and | mean
0.0626566	-14.299	in | argument
0.0653951	-14.299	and | likely
0.147059	-14.299	only | model.
0.666667	-14.2991	|xi|
0.0707395	-14.2993	constraints | to
0.0958084	-14.2993	from | probability
0.163636	-14.2995	is | arguments.
0.169811	-14.2996	dataset | this
0.0593607	-14.2997	to | across
0.428571	-14.2998	ontology | improve
0.075188	-14.2999	we | tokens
0.0815451	-14.2999	it | performance
0.212121	-14.3	of | (2).
0.186047	-14.3001	pereira | and
0.0629921	-14.3001	all | at
0.163636	-14.3002	set | constraints
0.294118	-14.3003	ranging | 1
0.0246769	-14.3005	first | the | and
0.118182	-14.3006	which | score
0.1375	-14.3006	using | representation
0.230769	-14.3006	supervised | use
0.108527	-14.3007	additional | by
0.428571	-14.3007	classification | 1,
0.123711	-14.3008	train | training
0.142857	-14.301	systems | such
0.141026	-14.301	having | as
0.195122	-14.3012	on | theory
0.0567227	-14.3014	those | a
0.153846	-14.3017	that | primary
0.666667	-14.3019	p(t|m),
0.0337987	-14.302	model, | the
0.142857	-14.3022	that | itself
0.26087	-14.3022	approach | models.
0.173077	-14.3023	data | articles
0.263158	-14.3024	letter | or
0.2	-14.3025	70 | 30
0.100671	-14.3027	but | when
0.0608273	-14.3028	and | quality
0.0601852	-14.3029	performed | and
0.0567227	-14.303	as | when
0.130435	-14.3031	this | a very
0.0909091	-14.3031	that | nodes
0.333333	-14.3032	determine | extent
0.666667	-14.3033	ibm | yorktown
0.666667	-14.3033	|j|
0.130952	-14.3033	is | unknown
0.134146	-14.3034	f | 1
0.160714	-14.3034	metric | between
0.0183071	-14.3035	feature | the
0.363636	-14.3035	machine | summit
0.21875	-14.3035	split | word
0.117117	-14.3037	linear | that
0.1	-14.3038	from | rules
0.0909091	-14.3039	as | value
0.118182	-14.3039	same | only
0.26087	-14.3039	tree | model.
0.186047	-14.304	2 | describe
0.0885417	-14.304	difference | for
0.115044	-14.304	t | u
0.25	-14.3042	• | human
0.157895	-14.3042	thus | can
0.151515	-14.3044	or | multiple
0.5	-14.3046	as | did.
0.0865385	-14.3047	to | (i.e.,
0.263158	-14.3047	ter | (have
0.333333	-14.3047	present | papers
0.0589569	-14.3047	and | domain
0.0589569	-14.3047	to | several
0.0380687	-14.305	domain | the
0.123711	-14.305	for each | target
0.0993377	-14.3051	in | arabic
0.0829493	-14.3054	in | information.
0.0731707	-14.3055	4 | 0
0.0507614	-14.3055	is | score
0.0484115	-14.3058	additional | of
0.277778	-14.3058	tags | tokens
0.181818	-14.3058	present | unsupervised
0.0215504	-14.3059	these | to
0.12037	-14.3059	language | order
0.0853081	-14.306	to | nouns
0.0603865	-14.3061	0 | 2
0.0650407	-14.3061	in | pages
0.142857	-14.3062	line | from
0.116071	-14.3062	algorithm | used
0.181818	-14.3066	to | only.
0.0825688	-14.3068	we | four | and
0.5	-14.3068	n-grams | bag
0.122449	-14.3069	in | recall.
0.24	-14.3069	b | .
0.0327298	-14.3071	the | much
0.0585586	-14.3073	from | same
0.3125	-14.3074	thompson | mr
0.3125	-14.3074	50.0 | 73.1
0.101351	-14.3074	2 | i
0.169811	-14.3075	are | testing
0.263158	-14.3075	algorithm | presented
0.666667	-14.3076	kl(θdk||θci)
0.294118	-14.3078	5: | or
0.0739437	-14.3079	groups | the
0.263158	-14.308	disputant | to
0.277778	-14.3082	2.1 | parallel
0.0309023	-14.3086	computed | the | the
0.0599078	-14.3086	a | quality
0.177778	-14.3087	it | approaches
0.149254	-14.3088	natural | information
0.0496795	-14.3089	section, | the
0.333333	-14.3089	• | attribute
0.333333	-14.3089	u.s. | grant
0.181818	-14.3089	2011), | june
0.181818	-14.3089	example: | to
0.0721649	-14.309	1 | two
0.0400844	-14.3092	are | that | a
0.0853081	-14.3093	to | vocabulary
0.25	-14.3095	cannot | such
0.0488189	-14.3096	each | word
0.0930233	-14.3097	on | previous
0.0649718	-14.3098	languages | to
0.0871795	-14.3099	one | sentence
0.112069	-14.3099	control | and
0.173077	-14.31	p | set
0.5	-14.31	positive | polar
0.5	-14.31	adequacy | image
0.0352847	-14.3103	to | performance
0.147059	-14.3103	order | its
0.0691824	-14.3107	lexical | with
0.059952	-14.3108	2 | 0
0.0837209	-14.3109	with | event
0.263158	-14.3109	>>> | =
0.263158	-14.3109	about | articles
0.117117	-14.311	information | at
0.039196	-14.3111	for | we | as
0.10274	-14.3112	significant | with
0.107692	-14.3114	when | sentence
0.190476	-14.3114	5.3 | with
0.166667	-14.3115	some | any
0.666667	-14.3118	p1(z|x;θ1)p2(z|x;θ2)
0.141026	-14.3118	(2010) | which
0.0355372	-14.312	both | a
0.129032	-14.312	resources | et
0.263158	-14.312	tokens | gold
0.108527	-14.3121	as | considered
0.151515	-14.3121	scores | data
0.163636	-14.3121	on | people
0.0530165	-14.3123	we | et al.
0.173913	-14.3123	then | vector
0.186047	-14.3124	suited | of
0.106061	-14.3124	for | phrase-based
0.0506757	-14.3125	the | seen
0.333333	-14.3125	errors | pos
0.333333	-14.3125	representations | at
0.205882	-14.3126	useful | but
0.147059	-14.3128	translation, | scotland,
0.141026	-14.313	for | selecting
0.263158	-14.3131	bilingual | approach
0.0876289	-14.3132	we | classifiers
0.666667	-14.3133	pf(w3|w1w2)
0.222222	-14.3137	that, | sentence
0.222222	-14.3137	matches | ∪
0.0344037	-14.3138	these | that
0.141026	-14.3139	a | log-linear
0.0958084	-14.314	our results | that
0.0766284	-14.3141	us | that
0.263158	-14.3142	lf | 87.38
0.194444	-14.3143	linguistic | sentence
0.0719178	-14.3145	are not | a
0.155172	-14.3146	in | stream
0.0496	-14.3147	the | occur
0.666667	-14.3147	corpora | langlais,
0.21875	-14.3147	it | aspects
0.181818	-14.3147	corpora | comparable
0.294118	-14.3148	from | subjectivity
0.173913	-14.3149	consequently | the
0.0367133	-14.315	english | of
0.116071	-14.3151	(which | and
0.10274	-14.3152	were | between
0.169811	-14.3152	seems | as
0.5	-14.3153	nb | each
0.25	-14.3154	learn | information
0.0801688	-14.3154	items | and
0.0515464	-14.3155	for | data.
0.0721649	-14.3155	explained | the
0.186047	-14.3156	characters | for
0.173913	-14.3157	is | immediate
0.0559006	-14.3159	information | be
0.25	-14.3159	using | created
0.126316	-14.316	challenge | in
0.0275455	-14.316	the | baseline
0.0781893	-14.3162	performance of | is
0.032481	-14.3163	of | english
0.163636	-14.3163	contains | about
0.0646067	-14.3165	reordering | to
0.0427553	-14.3166	it | which
0.157895	-14.3167	one | high
0.212121	-14.3168	can | fully
0.0993377	-14.3168	most | one
0.118182	-14.3168	n-gram | model
0.1375	-14.3168	word | no
0.0894737	-14.3169	thus | that
0.0967742	-14.3171	parse | are
0.0904255	-14.3172	is | significantly
0.0798319	-14.3172	for example, | on
0.163636	-14.3174	this | not | because
0.363636	-14.3175	various | may
0.121212	-14.3175	defined | an
0.0963855	-14.3177	information | such
0.24	-14.3177	europarl | corpus
0.277778	-14.3177	grammar | 2009).
0.0958084	-14.3179	as | extracted
0.0138819	-14.318	from | of
0.107692	-14.3181	named | recognition
0.0653409	-14.3181	the | structured
0.116071	-14.3181	with | distance
0.0941176	-14.3183	same | all
0.0681115	-14.3183	to | tokens
0.230769	-14.3183	terms | similar
0.2	-14.3183	(p), | and
0.2	-14.3183	search | web
0.212121	-14.3186	enable | with
0.24	-14.3187	analysis | related
0.24	-14.3187	online | have
0.0974026	-14.3189	compound | and
0.111111	-14.3189	we | having
0.277778	-14.3189	mapping | concepts
0.666667	-14.319	adverbs | respectively.
0.666667	-14.319	¯w|h)
0.0326162	-14.3192	of | first
0.121212	-14.3193	also | text
0.0791667	-14.3196	this | time
0.139241	-14.3198	not | way
0.0481928	-14.3198	improve | of
0.0592255	-14.3199	this | text
0.0721649	-14.32	framework | in
0.0841121	-14.32	syntactic | from
0.0890052	-14.3201	translated | a
0.0611247	-14.3201	we | error
0.0468521	-14.3202	texts | the
0.666667	-14.3204	ˆpintu(e|q)
0.21875	-14.3204	20 | 35
0.21875	-14.3204	surface | an
0.190476	-14.3204	features | automatic
0.157895	-14.3205	2.2 | et
0.0670732	-14.3206	only | it
0.0631579	-14.3206	an | two
0.181818	-14.3206	number | during
0.0740741	-14.3207	have | data
0.0952381	-14.321	evaluation | used
0.117117	-14.321	on | various
0.104478	-14.321	at | these
0.0621762	-14.3211	such | from
0.0788382	-14.3212	between | their
0.0388833	-14.3213	the | applied
0.263158	-14.3216	25 | 6
0.428571	-14.3218	noted, | a
0.428571	-14.3218	optimization | solved
0.0649718	-14.3218	coreference | and
0.666667	-14.3219	there | totally
0.149254	-14.322	model, | use
0.149254	-14.322	is | reduction
0.177778	-14.3221	variables | each
0.102041	-14.3222	each | average
0.363636	-14.3222	log-linear | (och,
0.230769	-14.3223	it | one.
0.155172	-14.3226	in order | only
0.118182	-14.3226	considered | by
0.0719178	-14.3226	p | =
0.0397906	-14.3228	the | available | the
0.190476	-14.3228	f | set
0.363636	-14.323	given | actual
0.205882	-14.3232	recall | baseline
0.212121	-14.3232	since | phrases
0.666667	-14.3233	p(cj|si)
0.04102	-14.3233	s | a
0.0871795	-14.3234	object | and
0.0952381	-14.3235	we | optimal
0.102041	-14.3235	r | f1
0.0738007	-14.3235	for | higher
0.0849057	-14.3235	international | on
0.307692	-14.3236	pleasantness | x
0.0405257	-14.3236	of | dialogue
0.0620155	-14.3237	and | (see
0.163636	-14.3237	⋆ | -
0.0805085	-14.3239	3 | model
0.066092	-14.3239	and | tags
0.118182	-14.3239	2 | 20
0.0696203	-14.3241	in our | are
0.0961538	-14.3241	user | on
0.222222	-14.3243	example, | above
0.0958084	-14.3244	elements | to
0.0801688	-14.3244	data | were
0.135802	-14.3245	/ | e
0.0467153	-14.3245	the | binary
0.0600962	-14.3245	cases | of
0.428571	-14.3246	+unsup, | eng
0.5	-14.325	trained | minimize
0.0766284	-14.3252	3.3 | of
0.333333	-14.3254	take | parsing
0.12766	-14.3255	on | containing
0.147059	-14.3255	files | to
0.428571	-14.3256	results | factorization
0.21875	-14.3256	lexical | while
0.0653409	-14.3257	procedure | of
0.155172	-14.326	extraction | sentence
0.5	-14.3261	recently, | showed
0.040393	-14.3262	a | score | the
0.140845	-14.3265	event | their
0.163636	-14.3266	2 | types
0.0709459	-14.3266	we | version
0.2	-14.3266	et | huang
0.0986842	-14.3267	suggest | in
0.149254	-14.3268	for | graphs
0.169811	-14.3269	an | head
0.24	-14.327	(2004) | al.
0.21875	-14.327	natural | speech
0.0573951	-14.3271	by | but
0.186047	-14.3272	and | complement
0.0961538	-14.3272	to train | to
0.194444	-14.3273	linguistic | translation
0.0142004	-14.3274	our | and
0.0729927	-14.3275	we | n-gram
0.140845	-14.3277	parameter | =
0.0880829	-14.3278	as | part
0.117117	-14.3279	characters | to
0.118812	-14.3279	instead | with
0.5	-14.3283	roles | participants
0.0818182	-14.3284	higher | for
0.166667	-14.3284	record | in
0.428571	-14.3284	lexconn | of
0.0448549	-14.3284	we | the | model.
0.0967742	-14.3288	candidate | from
0.121212	-14.3288	by | shared
0.2	-14.3288	90 | 60
0.140845	-14.3289	50 | 0
0.118812	-14.3289	was | out
0.181818	-14.3289	no | over
0.109244	-14.329	c | s
0.194444	-14.329	for | w1
0.666667	-14.3291	cea | cea+discourse
0.190476	-14.3291	learned | not
0.123711	-14.3292	for | his
0.2	-14.3292	many | methods
0.170213	-14.3293	noteworthy | the
0.103704	-14.3293	information | using
0.126316	-14.33	5 | evaluation
0.126316	-14.3305	more | likely
0.0769231	-14.3305	is | subject
0.173913	-14.3307	we have | sentences
0.0458971	-14.3309	to | no
0.307692	-14.3309	performing | parsing
0.144928	-14.3309	second, | as
0.155172	-14.331	members | of
0.160714	-14.3311	as | contextual
0.0575221	-14.3312	domain | to
0.0986842	-14.3312	of | representation.
0.0676923	-14.3313	our | work
0.109244	-14.3314	system, | as
0.123711	-14.3314	our | errors
0.0763052	-14.3314	a | is | weight
0.122449	-14.3315	when | over
0.0481366	-14.3315	at | with
0.160714	-14.3315	set | pairs.
0.0662824	-14.3315	data | or
0.307692	-14.3316	table | breakdown
0.25	-14.3316	transfer | lexical
0.5	-14.3316	concept | automatic
0.0685358	-14.3317	at | it
0.666667	-14.332	we | fromthe
0.0380952	-14.332	from | which
0.294118	-14.332	annotator | can
0.222222	-14.332	sentiment | (e.g.,
0.428571	-14.3321	relative | learning
0.205882	-14.3322	grammar | parsing
0.0145617	-14.3322	words | the
0.0424757	-14.3322	a | e | e
0.0397906	-14.3322	addition | the
0.24	-14.3323	test | statistical
0.2	-14.3323	comparable | corpora
0.0738007	-14.3325	for | (see
0.0738007	-14.3325	is | features.
0.0236599	-14.3326	may | the
0.0291317	-14.3326	the | the | being
0.5	-14.3327	(nivre | mcdonald
0.0709459	-14.3329	text | from
0.181818	-14.3329	then | terms
0.122449	-14.3333	out | than
0.170213	-14.3334	including | our
0.205882	-14.3335	data | quite
0.0689655	-14.3335	the | level,
0.0924855	-14.3336	edges | a
0.0449591	-14.3337	of | form | the
0.0647887	-14.3338	selection | in
0.114035	-14.3338	< | in
0.0313931	-14.334	like | the
0.0555556	-14.334	as | will
0.428571	-14.334	differences | modern
0.0512367	-14.334	and | terms
0.105263	-14.3342	many | it
0.333333	-14.3342	follow | source
0.333333	-14.3342	2 | 24
0.142857	-14.3343	however, | sentences
0.0681115	-14.3344	as | context
0.0955414	-14.3345	and | always
0.173913	-14.3345	this | showed
0.0566449	-14.3346	and | available
0.104478	-14.3346	shows | accuracy
0.0867347	-14.3346	accuracy | as
0.277778	-14.3346	scores | selected
0.134146	-14.3347	and | technologies
0.104478	-14.3348	cluster | is
0.222222	-14.3349	mooney, | zettlemoyer
0.24	-14.3349	are | code
0.666667	-14.335	soldiers | into
0.152542	-14.335	manually | data
0.0769231	-14.3352	significantly | in
0.110169	-14.3356	as | verbs
0.135802	-14.3358	j | 1
0.0657143	-14.3358	in | top
0.114035	-14.3358	although | than
0.121212	-14.336	specifically, we | a
0.5	-14.336	generates | trees
0.0379147	-14.3361	for | were
0.105263	-14.3363	twitter | a
0.666667	-14.3364	k1 | arg2-gol
0.113043	-14.3364	translation, | to
0.277778	-14.3365	automatically | phrase
0.0496689	-14.3366	so | in
0.0707071	-14.3366	corpora | for
0.033137	-14.3367	the | current
0.194444	-14.3367	compute | features
0.428571	-14.3368	impact | important
0.25	-14.3371	level: | the
0.25	-14.3371	verbs | were
0.0219851	-14.3372	the | of | training
0.0572687	-14.3372	our | but
0.177778	-14.3372	human | 1
0.117647	-14.3373	in | objects
0.0642458	-14.3374	this is | by
0.2	-14.3375	a | non-projective
0.212121	-14.3376	with | predicates.
0.0485133	-14.3377	for | new
0.0438144	-14.3378	found | and
0.666667	-14.3379	function | collins,
0.140845	-14.3379	systems | they
0.0304159	-14.338	translation | a
0.25	-14.3382	new | nodes
0.155172	-14.3382	features | entity
0.149254	-14.3383	discusses | and
0.277778	-14.3383	contexts | given
0.114035	-14.3383	it | worth
0.294118	-14.3384	words | fixed
0.0862944	-14.3384	larger | in
0.0436457	-14.3385	in addition | the
0.24	-14.3386	step | then
0.222222	-14.3388	clusters | have
0.149254	-14.3389	unlabeled | on
0.0317568	-14.3389	corpus | a
0.666667	-14.3393	0.70 | 0.75
0.230769	-14.3394	word | produce
0.189189	-14.3396	terms | same
0.0298864	-14.3397	for | our
0.020546	-14.3398	has | of
0.0986842	-14.3398	it | order
0.186047	-14.3399	it | four
0.155172	-14.3399	extraction | training
0.0621762	-14.34	is | defined
0.0219048	-14.3403	test | the
0.212121	-14.3404	laboratory, | of
0.0216641	-14.3404	to | each
0.0267943	-14.3405	the | function | the
0.428571	-14.3406	knowledge | relationship
0.0464441	-14.341	is | while
0.307692	-14.3411	˜null | append(new,
0.0366071	-14.3412	it | that | and
0.177778	-14.3412	the | energy
0.0746269	-14.3413	per | in
0.0898876	-14.3414	acknowledgments | supported
0.144928	-14.3414	space, | to
0.25	-14.3415	meaning | into
0.333333	-14.3415	task, | four
0.5	-14.3415	timegraph | relations
0.205882	-14.3417	similarity | algorithm
0.013426	-14.3417	the | used
0.170213	-14.3417	but | (e.g.,
0.177778	-14.3419	also | speech
0.0653409	-14.342	decoding | of
0.134146	-14.3421	may | due
0.10687	-14.3421	where | other
0.0393782	-14.3421	is | both
0.0888889	-14.3422	also | two
0.666667	-14.3423	t¨an¨a¨an | perusoikeuksia
0.212121	-14.3423	had | used
0.142857	-14.3425	1 | tree
0.102941	-14.3426	5, | and
0.0405702	-14.3426	to | context
0.194444	-14.3427	style | al.,
0.222222	-14.3428	against | other
0.181818	-14.3429	it | arguments
0.0474732	-14.3435	entity | a
0.186047	-14.3436	has | yet
0.116505	-14.3436	c | .
0.0368664	-14.3437	the | is | into
0.363636	-14.3437	sense | inventory
0.666667	-14.3438	diphosphate | formycin
0.24	-14.3439	information | tasks.
0.0714286	-14.3441	we | after
0.0515098	-14.3441	is | learning
0.13253	-14.3441	briefly | in
0.049505	-14.3443	example | to
0.0642458	-14.3443	the | and | probabilities
0.0841584	-14.3445	over | which
0.307692	-14.3445	n-gram | scores
0.0885417	-14.3445	we | effective
0.101449	-14.3445	p: | f:
0.230769	-14.3445	very | syntactic
0.0888889	-14.3447	the | ibm
0.25	-14.3448	0.7 | recall
0.0369344	-14.345	selection | the
0.277778	-14.345	were | punctuation
0.0930233	-14.3452	from | them
0.363636	-14.3453	lda | 2009).
0.0600962	-14.3454	100 | of
0.155172	-14.3454	results | test | are
0.157895	-14.3457	furthermore, | from
0.0265655	-14.3457	between | in
0.333333	-14.3459	(2007) | sentences
0.0178117	-14.346	from | in
0.194444	-14.3461	had | more
0.10687	-14.3462	to | needs
0.189189	-14.3463	utterance | dialogue
0.0522388	-14.3463	the | and | average
0.307692	-14.3465	parse | original
0.040724	-14.3466	the | prior
0.0893855	-14.3467	extracting | a
0.0788382	-14.3468	with | null
0.0593824	-14.347	cases, | the
0.24	-14.3471	percentile | the
0.149254	-14.3472	to | summaries.
0.0616967	-14.3474	a | research
0.0561555	-14.3474	likely | of
0.277778	-14.3475	training | sufficient
0.0893855	-14.3478	produced | a
0.0562771	-14.348	on | use
0.333333	-14.3481	“ | �”
0.5	-14.3481	papelera | civilizado
0.5	-14.3481	hieu | chris
0.5	-14.3481	fuij | u,j
0.5	-14.3481	ine | 10880
0.5	-14.3481	chris | ondrej
0.5	-14.3481	cdp | prop.
0.5	-14.3481	⇔ | ⇒bα
0.5	-14.3481	plant | ghent
0.5	-14.3481	a102 | a106
0.5	-14.3481	0.001 | 0.1k
0.5	-14.3481	rater1 | rater3
0.5	-14.3481	y∈y(x(i)) | 
0.5	-14.3481	a114 | {q
0.5	-14.3481	via | (tn)
0.5	-14.3481	p4r5 | p5r2
0.5	-14.3481	coindexation | i′
0.5	-14.3481	youth | reported
0.5	-14.3481	100000 | s-hmm
0.5	-14.3481	regularization | .80
0.5	-14.3481	ukraine | ed
0.5	-14.3481	maxpaths | kaminski
0.5	-14.3481	posemo(pp) | auxvb(lp)
0.5	-14.3481	en: | 0-0
0.5	-14.3481	−µ) | ∂γ(ψ)
0.5	-14.3481	nns) | nn)
0.5	-14.3481	(nnp | (jj
0.5	-14.3481	[β,γ]∈pproja([b,c]) | ={}
0.5	-14.3481	wi∈ca | wj∈cb
0.5	-14.3481	un | objeto
0.5	-14.3481	d91 | d88
0.5	-14.3481	passion | p”,
0.5	-14.3481	k32k20 | k37k20
0.5	-14.3481	dni | htest
0.5	-14.3481	segmentation) | queue
0.5	-14.3481	0000 | erwo
0.5	-14.3481	ont | valeur
0.5	-14.3481	‘matched’ | argument
0.5	-14.3481	(m) | 90.70
0.5	-14.3481	choler | pestered
0.5	-14.3481	instituto | (unam),
0.5	-14.3481	@100 | @5000
0.5	-14.3481	aisii3 | hai8
0.5	-14.3481	10000200003000040000 | model−var
0.5	-14.3481	10,10continue | 12,12be
0.5	-14.3481	(fp7) | n◦
0.5	-14.3481	purusha2 | hue10
0.5	-14.3481	(this) | (been)
0.5	-14.3481	(all) | kiye9
0.5	-14.3481	que | pas
0.5	-14.3481	trnasec | trnapyl
0.5	-14.3481	annoyed | vexed
0.5	-14.3481	nonterminals | (ti,a,tj),
0.5	-14.3481	play, | play
0.5	-14.3481	maxpaths | sender
0.5	-14.3481	“patient | vector”
0.5	-14.3481	daniel | “danny
0.5	-14.3481	(ever) | rahii7
0.5	-14.3481	(is|are|was|were|being) | npl
0.5	-14.3481	want(x, | bel(y,
0.5	-14.3481	nps | chnk∩nps
0.5	-14.3481	v1⊗v2 | v3
0.5	-14.3481	k31k20 | k34k20
0.5	-14.3481	p3r5 | pairwise
0.5	-14.3481	man | her
0.5	-14.3481	(fa) | rwth
0.5	-14.3481	0.588 | 0.859
0.5	-14.3481	81.51 | +0.41
0.5	-14.3481	<term | <span><target
0.5	-14.3481	[address2] | [email]
0.5	-14.3481	(not) | (has)
0.5	-14.3481	专uni5bb6 | uni6295降
0.5	-14.3481	(till) | kabhii5
0.5	-14.3481	(r | xt2)))
0.5	-14.3481	taka2 | (policy)
0.5	-14.3481	j | n(q);
0.5	-14.3481	uni60c5uni62a5 | 敌uni4eba
0.5	-14.3481	="0.5 | 8-n"/
0.5	-14.3481	np-sbj | unlabel
0.5	-14.3481	p(ot) | onull
0.5	-14.3481	χαρτόκουτο | χ+
0.5	-14.3481	co | 2759
0.5	-14.3481	“george | bush”,
0.5	-14.3481	[name] | [address3]
0.5	-14.3481	ja | back-translation:
0.5	-14.3481	absol | having-2|been-3|passed-4
0.5	-14.3481	aba1 | nahii6
0.5	-14.3481	@1000 | @2000
0.5	-14.3481	.3 | .5
0.5	-14.3481	-126 | erwo
0.5	-14.3481	p2r5 | p3r5
0.5	-14.3481	sabhii1 | (wear)
0.5	-14.3481	loss | 15,
0.5	-14.3481	0.855 | 0.879
0.5	-14.3481	(now) | niiti4
0.5	-14.3481	(qp | (cd
0.5	-14.3481	yes-answer | slot2,
0.25	-14.3482	context-features | (10k)
0.0542169	-14.3484	to | value
0.0679012	-14.3484	to | process
0.138889	-14.3485	reason | this
0.0930233	-14.3485	probability | an
0.363636	-14.3486	es-en | fr-en
0.25	-14.3487	these | traditional
0.263158	-14.3487	on | quality,
0.0584112	-14.3488	a | across
0.0311486	-14.349	form | the
0.126437	-14.349	argument | an
0.177778	-14.349	propose | information
0.177778	-14.349	capture | word
0.375	-14.349	f1 | mqa
0.375	-14.349	mepv | court”
0.375	-14.349	play | games
0.375	-14.349	clicks | clicks
0.129412	-14.3491	over | one
0.0546559	-14.3491	of | provide
0.0688525	-14.3492	of | 20
0.0810811	-14.3492	rate | to
0.24	-14.3492	experiments | terms
0.0552147	-14.3492	a | small
0.5	-14.3492	stemmed | semantically
0.5	-14.3492	cell | km
0.5	-14.3492	likelihood | 0-1
0.5	-14.3492	classifier: | potential
0.5	-14.3492	‘chains’ | embedded
0.5	-14.3492	representation | object,
0.5	-14.3492	fragments, | ungrammatical
0.5	-14.3492	25 | 77
0.5	-14.3492	2 | identification,
0.5	-14.3492	trust | guide
0.5	-14.3492	somasundaran | debate
0.5	-14.3492	wfst | paths
0.5	-14.3492	(sec/sent) | rm-b
0.5	-14.3492	φ, | θ
0.5	-14.3492	utterances | avg.
0.5	-14.3492	definition-only | questions
0.5	-14.3492	anchor | surrounding
0.5	-14.3492	such | ⇒σg
0.5	-14.3492	acoustic | 1st-tier
0.5	-14.3492	12, | year
0.5	-14.3492	detector | detectors,
0.5	-14.3492	gapping | cbhp
0.5	-14.3492	esel | over
0.5	-14.3492	kappa | 0.69
0.5	-14.3492	appropriateness | redundant
0.5	-14.3492	occurrences | banner
0.5	-14.3492	timing | cky
0.0422705	-14.3493	and | score
0.25	-14.3493	base | mira
0.25	-14.3493	original | results
0.127907	-14.3494	the | wer
0.0893855	-14.3494	or | into
0.0986842	-14.3494	4. | that
0.1	-14.3497	of | 2010
0.24	-14.3497	improvements | statistically
0.122449	-14.3498	using | vector
0.375	-14.3498	β | c.
0.0750988	-14.3499	is | relative
0.0664653	-14.3499	one | at
0.046176	-14.3499	the | was | on
0.173913	-14.3499	applied | translation
0.105263	-14.35	this | effective
0.277778	-14.35	suitable | our
0.0279536	-14.3502	the | were | to
0.173913	-14.3503	those | different
0.24	-14.3503	annotations | at
0.5	-14.3503	n-grams | 106
0.5	-14.3503	& | $
0.5	-14.3503	term | whatever
0.5	-14.3503	provide | completions
0.5	-14.3503	hebben | man
0.5	-14.3503	tts | tests
0.5	-14.3503	clues | identifying
0.0719424	-14.3504	this | us to
0.104478	-14.3505	only | source
0.040305	-14.3506	• | to
0.0616967	-14.3506	the | cosine
0.375	-14.3506	word | alone
0.0955414	-14.3507	parser | et al.,
0.173913	-14.3507	we | he
0.0497512	-14.3507	we define | the
0.0748031	-14.351	agreement | to
0.428571	-14.351	set | set)
0.0544355	-14.351	labels | of
0.0312708	-14.3511	the | positive
0.166667	-14.3512	report | our
0.230769	-14.3512	evaluation | existing
0.2	-14.3512	relation | event
0.177778	-14.3514	system, | can
0.0993377	-14.3514	2 | ,
0.375	-14.3515	paradise | satisfaction
0.375	-14.3515	reduce | effect
0.375	-14.3515	minimum | objective
0.5	-14.3515	pubmed | gene
0.5	-14.3515	occurrences | consequence
0.5	-14.3515	camel | brand
0.5	-14.3515	p(z|o,c) | p(o|z)p(z|c)summationtext
0.5	-14.3515	harmonic | 1.0.
0.5	-14.3515	adjective-noun | attribute
0.5	-14.3515	classification | email
0.5	-14.3515	mwul | value
0.5	-14.3515	manning, | klein,
0.106061	-14.3516	the | predicate.
0.130952	-14.3517	the | army
0.108333	-14.3519	collection | that
0.307692	-14.352	adequacy | correlation
0.0818182	-14.352	moreover, | a
0.0949367	-14.3521	have | text
0.127907	-14.3522	using | possible
0.0269476	-14.3522	the | systems
0.194444	-14.3522	made | more
0.0213025	-14.3522	in | it
0.033137	-14.3524	word | for
0.122449	-14.3526	which | types
0.5	-14.3526	huber | training.
0.5	-14.3526	paths | shorter
0.5	-14.3526	l=1 | }
0.5	-14.3526	popularity | popular
0.5	-14.3526	interest | speaker’s
0.5	-14.3526	relevant | terrorist
0.5	-14.3526	(jrip,bothlocal) | (t(9)
0.5	-14.3526	transferring | abstract
0.5	-14.3526	next | turns
0.5	-14.3526	number | thousands.
0.0967742	-14.3527	and | turn
0.0644258	-14.3528	unique | the
0.24	-14.353	et | 1994;
0.0681818	-14.3533	and | news
0.041049	-14.3533	the | types of
0.107438	-14.3533	in | distinguish
0.177778	-14.3534	add | or
0.0702341	-14.3536	c | in
0.0635359	-14.3536	of | linear
0.0833333	-14.3537	in | conference
0.277778	-14.3537	social | when
0.5	-14.3537	snijd | ‘cut’
0.5	-14.3537	entry | stage
0.5	-14.3537	.22 | kendall’s
0.5	-14.3537	neighbours | neighbours.
0.5	-14.3537	fluency | nist
0.5	-14.3537	university | road,
0.5	-14.3537	lm | ct.
0.5	-14.3537	images | object
0.0693069	-14.3538	for | instances
0.307692	-14.354	a | jointly.
0.307692	-14.354	log | (1)
0.081448	-14.354	widely | and
0.134146	-14.354	include | word
0.205882	-14.3545	nlu | target
0.07	-14.3547	these | more
0.147059	-14.3547	natural | which
0.0110195	-14.3547	we | that
0.0336134	-14.3548	the | meaning
0.5	-14.3548	pairs | judges)
0.5	-14.3548	b∈b(y) | c∈c(y)
0.5	-14.3548	(talukdar | 2010),
0.5	-14.3548	- | cannot
0.5	-14.3548	(2008), | charniak
0.5	-14.3548	hgfc | clusters
0.5	-14.3548	adjectives, | adjectives
0.5	-14.3548	88.7 | models
0.222222	-14.3551	use | matching
0.363636	-14.3551	list | system.
0.263158	-14.3552	metric | over
0.13253	-14.3553	using | would
0.177778	-14.3554	so | semantic
0.0396996	-14.3555	could | to
0.666667	-14.3556	0.682 | 0.707
0.2	-14.3557	and | format.
0.0536779	-14.3558	approach | et
0.109244	-14.3558	score | word
0.181818	-14.3558	examples | these
0.5	-14.3559	annotated | claims.
0.5	-14.3559	view | chat
0.0740741	-14.3564	this | human
0.0452617	-14.3565	probability | to
0.018932	-14.3565	is | is | is
0.375	-14.3565	mert | candidate
0.375	-14.3565	at | sense.
0.0441176	-14.3566	approaches | the | of
0.181818	-14.3566	in | utterance.
0.0883978	-14.3568	the | comments
0.0312085	-14.357	average | the
0.5	-14.357	treaties | arsenals
0.5	-14.357	global | ψ
0.5	-14.357	your | day
0.5	-14.357	categorization | linguistic
0.5	-14.357	pseudo | feedback
0.5	-14.357	s. | w.
0.0903955	-14.3571	but | one
0.0484653	-14.3571	compute | of
0.0980392	-14.3574	and | around
0.307692	-14.3575	svm | crf
0.130952	-14.3576	segmented | and
0.13253	-14.358	can | extended
0.0633609	-14.3581	graphs | the
0.126437	-14.3581	rules | set
0.375	-14.3582	into | database.
0.5	-14.3582	# | te
0.5	-14.3582	gold | 55
0.5	-14.3582	(statistically | at
0.5	-14.3582	solve | substitution
0.24	-14.3583	pos | n-gram
0.0841584	-14.3584	always | in
0.0359649	-14.3585	for | different
0.134146	-14.3585	of | relation,
0.130952	-14.3588	linking | of
0.0354671	-14.3589	the | present
0.25	-14.3593	large | makes
0.0961538	-14.3593	value | with
0.333333	-14.3593	might | appropriate
0.160714	-14.3593	shown | information
0.5	-14.3593	item | stack,
0.5	-14.3593	reranking | slu
0.177778	-14.3594	as | equivalent
0.0943396	-14.3596	online | to
0.0455192	-14.3596	for | most
0.363636	-14.3599	define | every
0.277778	-14.3599	make | even
0.126437	-14.36	statistically | <
0.152542	-14.3601	(a) | from
0.111111	-14.3601	and | classifier.
0.0263158	-14.3603	corpus | the | of
0.0253718	-14.3604	= | = | for
0.5	-14.3604	huang | developed
0.5	-14.3604	compute | parenleftbig
0.114286	-14.3606	while | same
0.375	-14.3607	terms | respective
0.0482315	-14.3607	as | use
0.0451977	-14.3608	of | total
0.25	-14.361	sentences | document.
0.100719	-14.361	this | during
0.222222	-14.3611	using | slightly
0.140845	-14.3611	evaluations | in
0.135135	-14.3615	introduced | as
0.0550847	-14.3615	and | speech
0.230769	-14.3615	dt | tree
0.5	-14.3615	arabic | normalized
0.5	-14.3615	experiments | pang
0.5	-14.3615	segment | operation
0.5	-14.3615	i | 'do'
0.666667	-14.3616	contact | obtain
0.0675241	-14.3617	instances | in
0.130952	-14.3617	where | y
0.0408627	-14.3617	1 introduction | the
0.144928	-14.3618	space | model
0.110169	-14.3618	in | constructed
0.294118	-14.362	calculating | words
0.0398278	-14.362	that | when
0.230769	-14.3621	role | srl
0.051756	-14.3622	relations | in
0.375	-14.3624	map | f-score
0.222222	-14.3626	achieve | better
0.109244	-14.3627	longer | to
0.104478	-14.3627	lexical | words
0.5	-14.3627	provided | distinguish
0.5	-14.3627	v1 | meaning
0.5	-14.3627	distance | ci
0.5	-14.3627	57.7 | 2:
0.263158	-14.3629	candidate | np
0.214286	-14.363	have | representations
0.166667	-14.3631	these | 2009).
0.277778	-14.3631	phonemic | and
0.363636	-14.3632	np@1 | vp@2
0.375	-14.3632	list | name
0.375	-14.3632	consequently, | more
0.294118	-14.3633	question | part
0.294118	-14.3633	use | amounts
0.170213	-14.3633	corpora are | in
0.0503472	-14.3634	we | model.
0.0320781	-14.3635	a | given
0.0584112	-14.3635	features | not
0.0810811	-14.3636	models | translation
0.126437	-14.3637	with | presented
0.25	-14.3638	while | learn
0.5	-14.3638	trod | nge
0.5	-14.3638	tsvm | but
0.263158	-14.3641	monolingual | set
0.375	-14.3641	newstest2008 | bleu
0.0858586	-14.3642	pos | are
0.105263	-14.3642	nodes | be
0.24	-14.3643	complexity | can
0.144928	-14.3644	must | not
0.333333	-14.3646	domain | results.
0.230769	-14.3647	included | training
0.230769	-14.3647	here | given
0.214286	-14.3649	grammar | given
0.5	-14.3649	weighting | recall
0.5	-14.3649	high-level | system.
0.0269461	-14.3651	also | to | the
0.054326	-14.3652	figure | by
0.0837438	-14.3652	our | methods
0.0464072	-14.3653	b | the
0.333333	-14.3653	much | easier
0.0841584	-14.3654	of | class.
0.105263	-14.3654	algorithm | this
0.142857	-14.3654	t | k
0.122449	-14.3655	by | natural
0.0452617	-14.3655	both | as
0.428571	-14.3655	cold | wet
0.428571	-14.3655	clinical | (e.g.,
0.0883978	-14.3655	we | approach.
0.0433121	-14.3655	, | of
0.0445344	-14.3656	of | for | pages
0.129412	-14.3657	3 | was
0.363636	-14.3657	z | (2)
0.157895	-14.3658	distribution | more
0.0286678	-14.3659	a | their
0.0396996	-14.3659	relative | the
0.294118	-14.366	notice | both
0.144928	-14.366	d | where
0.5	-14.3661	transformed | grammar
0.0724638	-14.3662	construction | of
0.0914286	-14.3662	approach | or
0.0603015	-14.3662	t | a | of
0.0670927	-14.3663	corpus, | a
0.113043	-14.3664	since | same
0.034019	-14.3664	models | in
0.0183276	-14.3664	the | target
0.0403587	-14.3664	the | ranking
0.263158	-14.3665	less | pairs
0.025	-14.3666	a | at
0.0641399	-14.3666	other | from
0.173913	-14.3667	results | (see
0.0254351	-14.3667	pairs | the
0.0750988	-14.3668	for | current
0.277778	-14.3668	projected | target
0.214286	-14.3669	well | previous
0.112069	-14.3671	which | tree
0.126437	-14.3672	topics | from
0.5	-14.3672	procedure | suggestions
0.666667	-14.3676	layer | spelling
0.0869565	-14.3676	regression | of
0.0841584	-14.3677	were | they
0.15	-14.3677	agreement | was
0.0750988	-14.3679	a | mentions
0.0525292	-14.3679	and | function
0.122449	-14.368	an | generated
0.294118	-14.368	(ew | (pos)),
0.294118	-14.368	aux | aux
0.294118	-14.368	solution% | merit
0.222222	-14.3681	most | corresponding
0.0604534	-14.3681	to | length
0.135135	-14.3682	the | strategy,
0.10219	-14.3682	results | 3
0.0552017	-14.3683	from | different
0.0521415	-14.3683	has | from
0.5	-14.3683	points | rest
0.222222	-14.3686	unlike | each
0.0521415	-14.3686	with | only
0.122449	-14.3688	it will | to
0.184211	-14.3688	furthermore, | each
0.184211	-14.3688	the | insightful
0.0650888	-14.3689	the | and | next
0.333333	-14.3691	obtain | recall
0.170213	-14.3695	evaluations | a
0.116505	-14.3695	first, | is | to
0.5	-14.3695	student | move
0.5	-14.3695	pragmatic | dimension
0.0250108	-14.3695	of | given
0.0371429	-14.3698	null | nullnull | null
0.0979021	-14.3698	we | senses
0.363636	-14.3698	score | selected.
0.0727273	-14.3698	in | 2, | and
0.0675241	-14.3699	to do | the
0.333333	-14.3699	translations | single
0.307692	-14.37	specify | lexical
0.25	-14.3701	(2010), | use
0.428571	-14.3703	stages: | set
0.24	-14.3703	metric | p
0.110169	-14.3703	are | complex
0.263158	-14.3706	train | pairs
0.177778	-14.3706	distributional | between
0.277778	-14.3706	obtained | taking
0.5	-14.3706	conclusions | texts.
0.666667	-14.3707	itg | derivation
0.114286	-14.3709	this | already
0.0631868	-14.3709	is | size
0.00613339	-14.371	in | of
0.0701754	-14.3713	each | only
0.138889	-14.3713	representing | is
0.163265	-14.3714	give | our
0.0631868	-14.3715	can | words
0.0792952	-14.3715	low | in
0.363636	-14.3715	types | break
0.5	-14.3717	p(c|s)
0.5	-14.3717	ko | original
0.5	-14.3717	presence | class.
0.5	-14.3717	nmaxe=j | k=j+1
0.0733591	-14.3718	therefore | in
0.0955414	-14.3718	reordering | on
0.177778	-14.3718	probability | more
0.101449	-14.3721	first | their
0.0985915	-14.3722	expressions | for
0.0825243	-14.3723	produce | and
0.194444	-14.3723	with | respective
0.0261194	-14.3724	would | to
0.0821256	-14.3724	linguistic | for
0.277778	-14.3725	40 | 8
0.375	-14.3726	significantly | baselines
0.0243112	-14.3728	it | for
0.123596	-14.3728	few | with
0.333333	-14.3729	up | could
0.5	-14.3729	main | macaon
0.5	-14.3729	posterior | (grac¸a
0.0284281	-14.373	of | 1
0.0864865	-14.3731	sets | with
0.133333	-14.3731	because | will
0.189189	-14.3731	retrieval | from
0.109244	-14.3736	random | are
0.0603015	-14.3736	to | (e.g.
0.0841584	-14.3738	for | head
0.0625	-14.3738	the | of | mentions
0.5	-14.374	samsung | sony
0.5	-14.374	had | [conceived
0.5	-14.374	consider | happens
0.375	-14.3743	same | mentioned
0.0652819	-14.3743	the | implementation
0.152542	-14.3749	using | instances
0.277778	-14.375	for | normally
0.123596	-14.3751	are | interested
0.5	-14.3751	link | label.
0.428571	-14.3752	extract | phrases.
0.333333	-14.3752	events | rules
0.2	-14.3752	be | one.
0.0903955	-14.3753	the | (p
0.170213	-14.3753	being | at
0.194444	-14.3754	computational | modeling
0.113208	-14.3754	by | like
0.0925926	-14.3754	variables | to
0.163265	-14.3755	domain, | is
0.0949367	-14.3755	with | reference
0.24	-14.3757	improves | bleu
0.0545073	-14.3757	for | type
0.0786026	-14.3758	of | calculated
0.113208	-14.3759	be | simple
0.263158	-14.376	set | sentences).
0.375	-14.376	and | italian.
0.0643275	-14.3762	description | of
0.307692	-14.3763	requests | requests
0.0184834	-14.3763	from | to | the
0.0350128	-14.3763	the | tags
0.0992908	-14.3764	some | word
0.0434783	-14.3768	shows | in
0.07393	-14.3768	were | an
0.123596	-14.3768	that | character
0.0639535	-14.3769	particular | in
0.0817308	-14.377	the | normal
0.0681818	-14.3771	annotators | of
0.428571	-14.3772	de | ai
0.11215	-14.3773	prefix | of
0.214286	-14.3776	inspired | work
0.214286	-14.3776	apply | these
0.00755043	-14.3778	to | to
0.230769	-14.3778	time | most
0.428571	-14.3781	aligned | 2008)
0.0639535	-14.3782	perform | to
0.307692	-14.3784	(2010) | latent
0.133333	-14.3785	all | following
0.108333	-14.3786	results | human
0.5	-14.3786	slot | challenging
0.5	-14.3786	dialogue | computer-animated
0.5	-14.3786	optimization | selecting
0.122222	-14.3787	in | detecting
0.170213	-14.3788	transliteration | from
0.0361337	-14.3788	this | using
0.105691	-14.3788	that | allow
0.363636	-14.3789	keys) | distractor
0.363636	-14.3789	.5 | .5
0.363636	-14.3789	21 | 11e
0.363636	-14.3789	φ(x | p
0.363636	-14.3789	0.39 | 0.66
0.363636	-14.3789	*** | ***
0.363636	-14.3789	rusty | motorbike
0.181818	-14.3789	w=0.1 | morning
0.0595533	-14.3793	∈ | and
0.0992908	-14.3796	effectiveness | and
0.2	-14.3797	scores | 3
0.5	-14.3797	domain | poorly
0.5	-14.3797	lda | having
0.307692	-14.3798	dog | black
0.0641399	-14.3799	dependency | with
0.0789474	-14.3803	report | in
0.375	-14.3803	require: | current
0.03362	-14.3803	the | grammar
0.0985915	-14.3805	compute | as
0.11215	-14.3807	used | systems
0.5	-14.3809	identifies | speakers
0.0864865	-14.381	is | hypothesis
0.123596	-14.381	in | events.
0.375	-14.3812	“red | to
0.666667	-14.3813	ist | als
0.104	-14.3815	and | update
0.163265	-14.3815	and | antecedent
0.115385	-14.3815	graph | are
0.25	-14.3815	available | large
0.177778	-14.3816	shows | types
0.0931677	-14.3817	and | cosine
0.307692	-14.3819	address | child
0.307692	-14.3819	expected | k
0.375	-14.382	integrated | 2007).
0.0688525	-14.3821	human | is
0.0805687	-14.3824	possible | by
0.0430809	-14.3825	to | those
0.307692	-14.3826	very | performance.
0.061008	-14.3828	test | are
0.0813397	-14.3828	on | domain
0.666667	-14.3829	due | senses.
0.125	-14.3829	obtained by | to
0.0949367	-14.3829	for | predicate
0.0943396	-14.383	m | t
0.0675241	-14.383	to | events
0.5	-14.3832	farmers | political
0.5	-14.3832	object | used.
0.5	-14.3832	(och, | 2008),
0.5	-14.3832	pφ(x|yi)
0.2	-14.3834	5 | –
0.177778	-14.3836	frequency | target
0.0845771	-14.3837	this | alignment
0.133333	-14.3837	above | from
0.0765957	-14.3837	the | prefix
0.0505415	-14.3838	in | like
0.375	-14.3838	proportion | ns
0.0502693	-14.3841	random | of
0.0591133	-14.3841	is | by | from
0.12	-14.3843	each | order
0.5	-14.3843	creation | narrative
0.131579	-14.3844	algorithms | by
0.136986	-14.3845	.0 | 0
0.09375	-14.3846	we | variables
0.189189	-14.3848	analysis | classification
0.125	-14.3849	l | m
0.136986	-14.3851	between | relations
0.277778	-14.3852	area | which
0.307692	-14.3855	from | community.
0.0879121	-14.3855	1, | for
0.214286	-14.3855	the | s/he
0.5	-14.3855	bagwm_wc | hc_wc
0.5	-14.3855	issue | wsi
0.0864865	-14.3856	are | rather
0.0762712	-14.3857	a | phrase-based
0.205882	-14.3858	to | priors
0.038105	-14.3858	in | any
0.666667	-14.3859	settled | iran
0.184211	-14.3861	likely | different
0.0782609	-14.3861	of | coverage
0.138889	-14.3862	single | data
0.147541	-14.3862	on | languages,
0.057971	-14.3864	we | value
0.375	-14.3864	difficulty | they
0.177778	-14.3865	for | templates
0.214286	-14.3865	on | children
0.108333	-14.3866	as | another
0.120879	-14.3867	use | simple
0.136986	-14.3867	two | pair
0.0371094	-14.3867	introduction | the
0.0489865	-14.3867	3 | we
0.155172	-14.3868	that | increased
0.307692	-14.3869	method | accuracy.
0.025858	-14.3869	the | of | over
0.428571	-14.387	perplexity | each
0.166667	-14.3871	set | contain
0.130952	-14.3871	output | not
0.277778	-14.3872	nlp | 2007;
0.0673077	-14.3873	is | baseline
0.0893855	-14.3873	has | information
0.0345992	-14.3873	the | several
0.15	-14.3876	would | its
0.135135	-14.3877	we | collect
0.0342298	-14.3878	have | as
0.0594059	-14.3878	still | of
0.5	-14.3878	j=1 | +)
0.230769	-14.3879	(iv) | the
0.00841246	-14.388	the | are | the
0.030303	-14.3882	however, | and
0.135135	-14.3883	2 | but
0.277778	-14.3884	ordering | words
0.277778	-14.3884	second, | same
0.184211	-14.3886	over | n
0.0290112	-14.3889	can | with
0.0496575	-14.3889	way | the | of
0.206897	-14.3889	source | .
0.5	-14.3889	percentage | live
0.5	-14.3889	tasks, | tagging,
0.666667	-14.389	deaf | hearing-impaired
0.666667	-14.389	||p||
0.428571	-14.389	propose | strategies
0.24	-14.389	penntb | pennconv
0.155172	-14.389	more | terms
0.184211	-14.3891	reinforcement | to
0.142857	-14.3893	words | (e.g.,
0.177778	-14.3894	language | much
0.108333	-14.3894	knowledge, | and
0.0271516	-14.3895	a | has
0.04	-14.3896	although | of
0.0374753	-14.3897	small | the
0.0841584	-14.3898	advantage | is
0.0586797	-14.3898	graph | to
0.206897	-14.3899	- | c
0.053719	-14.39	can | such
0.142857	-14.39	3, | with
0.0489865	-14.39	of | space
0.230769	-14.39	thanks | reviewers
0.0139822	-14.3901	we | the | as
0.5	-14.3901	fahrr¨ader | gibt
0.5	-14.3901	threshold | 0.9
0.5	-14.3901	reports | f-measure
0.0750988	-14.3902	a | dataset
0.166667	-14.3902	tokens | than
0.0965517	-14.3904	(or | in
0.333333	-14.3905	v | dominates
0.0704225	-14.3905	positive | as
0.666667	-14.3906	pr(v|ei)
0.0145812	-14.3909	are | and | the
0.140845	-14.391	a | assigning
0.0619946	-14.391	a | generate
0.230769	-14.3911	uses | order
0.0142735	-14.3914	in | to | in
0.152542	-14.3914	important | between
0.0550847	-14.3916	is | related
0.0183573	-14.3918	we | which
0.135135	-14.3921	we find | on
0.0354296	-14.3921	for | their
0.108333	-14.3921	they | two
0.0694444	-14.3923	a | base
0.136986	-14.3924	method | better
0.375	-14.3924	any | score,
0.375	-14.3924	curran, | curran,
0.5	-14.3924	compared | no-unary
0.5	-14.3924	studies | death
0.0339806	-14.3925	may | in
0.173913	-14.3926	as | classification.
0.0635838	-14.3927	∈ | is
0.189189	-14.3927	set | 3.
0.04947	-14.393	alignment | a
0.173913	-14.393	grammars, | the
0.0666667	-14.3931	and | local
0.107438	-14.3932	has | both
0.307692	-14.3933	multiple | highest
0.307692	-14.3933	apis | and
0.307692	-14.3933	2005) | translation
0.136986	-14.3933	judges | and
0.206897	-14.3933	name | context
0.0972222	-14.3934	in | spanish
0.0701754	-14.3934	for | even
0.0677419	-14.3934	from | 1
0.15	-14.3935	agglomerative | the
0.25	-14.3936	written | word
0.25	-14.3936	one | dynamic
0.133333	-14.3936	has | similar
0.5	-14.3936	named | group
0.5	-14.3936	e5 | e6
0.0496454	-14.3936	c | and
0.0616622	-14.3939	is | parsing
0.136986	-14.394	fall | of
0.428571	-14.394	our | sieves
0.147541	-14.3941	label | this
0.375	-14.3942	number | out-of-vocabulary
0.0792952	-14.3943	shown | 1.
0.0888889	-14.3944	2 | 8
0.0837438	-14.3945	in | weights
0.2	-14.3945	we | δ
0.5	-14.3947	|test-set| | ,gold3
0.170213	-14.3949	we have | english
0.428571	-14.395	percent | k
0.666667	-14.3952	p(z|n)
0.110092	-14.3953	target | all
0.127907	-14.3954	2 | parsing
0.0903614	-14.3954	1 | 9
0.184211	-14.3955	add | our
0.127907	-14.3957	we | e.g.
0.115385	-14.3957	for example, | at
0.0546218	-14.3957	by | its
0.5	-14.3959	0.35 | 6:
0.428571	-14.396	because | increasing
0.428571	-14.396	journal | american
0.206897	-14.3962	that | fillers
0.375	-14.3968	tuning | system.
0.375	-14.3968	makes | difference
0.0368932	-14.3971	on | an
0.136986	-14.3972	system | dialogue
0.0753138	-14.3972	we | probabilities
0.177778	-14.3972	it | assigned
0.135135	-14.3974	experiments, | from
0.0586797	-14.3976	made | and
0.16	-14.3976	will | very
0.285714	-14.3977	among | ¡
0.285714	-14.3977	detail | section.
0.170213	-14.3977	you | be
0.108333	-14.3977	on | 24
0.0684932	-14.3978	is | instance
0.0931677	-14.3979	hope | the
0.0428571	-14.3979	the | manually
0.5	-14.3982	mentions | 45.9
0.155172	-14.3983	of | portuguese
0.0302445	-14.3983	in | its
0.115385	-14.3987	results | scores
0.222222	-14.3988	decoding | models
0.123596	-14.3988	– | an
0.214286	-14.3989	spanish | from
0.285714	-14.399	would | take
0.0664557	-14.3991	as | english
0.140845	-14.3992	features | scores
0.0892857	-14.3994	we | rank
0.375	-14.3994	orig | approx
0.5	-14.3994	relative | neither
0.0051649	-14.3994	of | and
0.102362	-14.3995	an | parse
0.285714	-14.3997	0 | gs
0.230769	-14.3997	similarity | latent
0.0869565	-14.3997	we | support
0.09375	-14.3999	optimal | and
0.0470958	-14.4001	some | as
0.184211	-14.4002	structure | used
0.0384615	-14.4003	work | on
0.142857	-14.4003	was | new
0.0648148	-14.4003	know | the
0.222222	-14.4004	phrase | 2003).
0.25	-14.4006	sentences | sentence,
0.0920245	-14.4006	has | into
0.166667	-14.4007	accuracy | set.
0.0892857	-14.4007	with | associated
0.127907	-14.4008	has | 2010;
0.0687285	-14.4008	for | correct
0.142857	-14.4009	most | those
0.163265	-14.4012	predicted | by
0.116505	-14.4012	scores | which
0.230769	-14.4013	pipeline | which
0.666667	-14.4014	˜f,e|φ(e)=e
0.145161	-14.4018	as | child
0.0581114	-14.4019	this | sentences
0.126437	-14.4019	8 | is
0.116505	-14.402	1 introduction | been
0.277778	-14.402	+ | /
0.375	-14.4021	how | incorporated
0.333333	-14.4022	biomedical | information
0.15	-14.4022	on | analysis.
0.230769	-14.4024	this | future.
0.2	-14.4025	is | points.
0.131579	-14.4025	states | as
0.0979021	-14.4028	the | dual
0.0716981	-14.4029	100 | and
0.428571	-14.403	way | topics
0.333333	-14.403	parse | entire
0.155172	-14.4031	a | queries.
0.0425258	-14.4031	of | classification
0.111111	-14.4032	example | word
0.166667	-14.4034	with | scope
0.263158	-14.4035	than | others.
0.0677966	-14.4035	the | july
0.16	-14.4036	are | yet
0.170213	-14.4037	result, | with
0.214286	-14.4039	since | data,
0.2	-14.4043	accuracy | parser
0.0704225	-14.4043	has | than
0.214286	-14.4044	parsing | lexical
0.104839	-14.4045	with | sentences.
0.060686	-14.4045	2: | in
0.666667	-14.4046	kapil | ny
0.230769	-14.4046	hierarchical | clustering
0.0600522	-14.4046	output | in
0.0553097	-14.4047	to be | with
0.101562	-14.4048	perceptron | the
0.0733591	-14.4052	on | three
0.119565	-14.4052	syntactic | between
0.0914634	-14.4052	6 | 2
0.11215	-14.4052	these | large
0.135135	-14.4053	included | that
0.133333	-14.4055	however, | should
0.0801887	-14.4057	sampling | of
0.166667	-14.4058	experiments | treebank
0.206897	-14.4059	quality | sentences
0.122222	-14.406	splitting | and
0.07393	-14.4061	the | moses
0.057971	-14.4062	previous | al.,
0.126437	-14.4063	sentences | both
0.285714	-14.4064	minnesota | 55455
0.166667	-14.4065	a | conjunction
0.25	-14.4065	domains, | al.,
0.170213	-14.4065	work | national science
0.0883978	-14.4065	5 | 3
0.0736434	-14.4065	as | input
0.16	-14.4066	improvement | model
0.263158	-14.4066	computational | key
0.184211	-14.4067	a | all.
0.133333	-14.4067	can | another
0.0782609	-14.4067	words | target
0.138889	-14.407	entity | knowledge
0.184211	-14.4071	are | responses
0.00720429	-14.4071	this | the
0.155172	-14.4071	have | they are
0.0855615	-14.4076	that | instances
0.5	-14.4076	pij | i
0.114286	-14.4077	for | wmt
0.285714	-14.4077	they | included
0.173913	-14.4077	randomly | data
0.0621469	-14.4078	between | which
0.136986	-14.4078	for | feedback
0.0544662	-14.4078	3 | 2
0.0646154	-14.4078	np | and
0.11215	-14.4078	pairs | were
0.0787037	-14.408	for | n-gram
0.138889	-14.4083	for | translations.
0.25	-14.4083	(marton | al.,
0.0214947	-14.4083	the | some
0.263158	-14.4084	expert | at
0.0455213	-14.4089	future | to
0.0837696	-14.409	are | important
0.238095	-14.4093	approach, | only
0.194444	-14.4093	or | 2003;
0.0704225	-14.4096	of | partial
0.0496454	-14.4096	of | k
0.307692	-14.4099	axiom | ◦
0.100775	-14.4099	to | extraction.
0.15	-14.41	4 | over
0.375	-14.41	achieve | good
0.0979021	-14.4102	both | well
0.029375	-14.4102	the | parallel
0.263158	-14.4103	parallel | lexicon
0.263158	-14.4103	set | hand
0.0992908	-14.4103	this | complex
0.0746888	-14.4103	type | are
0.285714	-14.4104	results, | well
0.238095	-14.4104	just | would
0.238095	-14.4104	features | &
0.0592784	-14.4104	particular | and
0.0608466	-14.4105	a | precision
0.107438	-14.4105	explicitly | to
0.194444	-14.4111	for | overlapping
0.12987	-14.4113	that | done
0.0399543	-14.4114	the | produced | the
0.0502693	-14.4115	forms | the
0.0687285	-14.4115	function | that
0.263158	-14.4115	table 3: | feature
0.263158	-14.4115	well, | with
0.110092	-14.4116	are | include
0.170213	-14.4117	inputs | and
0.189189	-14.4118	5 conclusions | to
0.131579	-14.4118	is | at each
0.16	-14.4119	the | for | measure.
0.0801887	-14.412	prefix | the
0.307692	-14.4121	with | following:
0.0470779	-14.4123	parsing | a
0.11215	-14.4123	with | across
0.135135	-14.4124	knowledge | semantic
0.5	-14.4124	wishes | thank
0.5	-14.4124	but | code.
0.0351494	-14.4125	is | translation
0.173913	-14.4126	with | order.
0.0675676	-14.4126	spanish | the
0.189189	-14.4127	corpus | which | by
0.103175	-14.4127	possible | all
0.0362595	-14.4129	of | university
0.0719697	-14.413	to | paraphrase
0.0257634	-14.4131	a | is | as
0.126437	-14.4132	specific | which
0.131579	-14.4134	this | resource
0.0524194	-14.4135	user | in
0.138889	-14.4135	even | such
0.0578313	-14.4135	in | average
0.5	-14.4135	problem, | string
0.0654206	-14.4139	you | the
0.263158	-14.414	better | sentiment
0.111111	-14.414	the | artificial
0.0992908	-14.4141	score | all
0.163265	-14.4143	however, | find
0.053719	-14.4143	time | in
0.285714	-14.4145	order | effects
0.0518234	-14.4145	is | english
0.0903614	-14.4145	implementation | to
0.173913	-14.4147	translated | other
0.5	-14.4147	giving | total
0.0406213	-14.4148	by | word
0.0617978	-14.4149	5 | our
0.206897	-14.4151	related | event
0.307692	-14.4157	paper | discusses
0.170213	-14.4157	dependency | translation
0.119565	-14.4159	various | with
0.106557	-14.4161	other | semantic
0.206897	-14.4161	group | choose
0.12987	-14.4165	we | community
0.230769	-14.4166	implementing | a
0.0759494	-14.4168	head | in
0.119565	-14.4168	each | own
0.0972222	-14.417	test | was
0.189189	-14.4171	been | before
0.2	-14.4171	p2 | p1,
0.5	-14.4171	p(a|data,α)
0.16	-14.4172	more | annotation
0.116505	-14.4174	define | on
0.0602094	-14.4174	example | a | of
0.0595855	-14.4174	since | on
0.0851064	-14.4178	α | of
0.0965517	-14.4179	in | accuracy.
0.0388889	-14.4181	the | subject
0.0551876	-14.4182	translation | are
0.175	-14.4184	value | will
0.0450581	-14.4185	to | topic
0.0892857	-14.4186	our | analysis
0.0516252	-14.4187	or | an
0.11828	-14.4188	performance | used
0.156863	-14.4189	with | computer
0.101562	-14.419	and | otherwise.
0.140625	-14.4191	to | subsets
0.222222	-14.4193	4.1 | experiments
0.0462963	-14.4193	c©2011 association for | and
0.133333	-14.4194	anonymous | helpful
0.166667	-14.4196	 | with
0.0952381	-14.4197	on | languages
0.0671141	-14.4198	approaches | for
0.184211	-14.4198	sou | e
0.115385	-14.4199	summarization | al.,
0.0572792	-14.4199	make | in
0.0985915	-14.42	active | a
0.122222	-14.4202	major | for
0.0903614	-14.4203	has | semantic
0.666667	-14.4204	homology-based | (section
0.135135	-14.4204	2 | these
0.108108	-14.4205	model | similar
0.0846561	-14.4209	from | while
0.0817308	-14.421	based | using
0.0166968	-14.421	the | only | the
0.238095	-14.4211	manually | sentence
0.0169878	-14.4212	can | a
0.136986	-14.4214	in | allowed
0.135135	-14.4214	may | many
0.0626781	-14.4214	this | well
0.0428954	-14.4215	the | constraints
0.111111	-14.4222	we | predicates
0.179487	-14.4223	1 | summary
0.428571	-14.4223	57 | 3
0.0813397	-14.4223	to | apply
0.2	-14.4224	 | 
0.107143	-14.4225	that | end
0.375	-14.4225	r-1 | category:
0.194444	-14.4227	method | automatic
0.0537634	-14.4228	means | of
0.263158	-14.4228	las | las
0.047619	-14.4228	when | is | and
0.108108	-14.4228	an | defined
0.140625	-14.4231	more | types
0.0716981	-14.4231	of | subset
0.0572792	-14.4234	the | person
0.333333	-14.4235	3 | harmonic
0.666667	-14.4236	those | supporters.
0.113208	-14.4236	in | until
0.0695971	-14.4236	systems | with
0.145161	-14.4238	model | binary
0.122222	-14.424	this suggests | of
0.0638298	-14.424	a | trees
0.0495575	-14.4241	for | performance
0.175	-14.4243	since | likely
0.428571	-14.4243	term | mean
0.375	-14.4243	0.5 | curve
0.0708955	-14.4243	may | not
0.0250227	-14.4248	to | different
0.0985915	-14.425	with | state
0.0476974	-14.425	give | the
0.120879	-14.4251	we | proper
0.0864865	-14.4251	assumption | in
0.0604396	-14.4251	our | task
0.277778	-14.4251	key, | →dt
0.0985915	-14.4254	work | models
0.428571	-14.4254	md | md
0.133333	-14.4254	links | are
0.0914634	-14.4254	assign | of
0.0612813	-14.4255	translations | to
0.122222	-14.4257	individual | be
0.0992366	-14.426	this | detection
0.0583756	-14.4261	individual | of
0.0837696	-14.4263	length | for
0.113208	-14.4263	core | a
0.136986	-14.4264	cannot | from
0.0570071	-14.4264	and | whether
0.0882353	-14.4264	and | well.
0.128205	-14.4266	tasks, | as
0.0341588	-14.4267	a | of | at
0.0414573	-14.4268	and | similar
0.238095	-14.4269	accuracy | segmentation
0.136986	-14.427	further | more
0.375	-14.427	inflectional | lexical
0.0549451	-14.4273	be | can be
0.0482759	-14.4274	et | 2006;
0.307692	-14.4275	correlation | against
0.206897	-14.4275	resulting | words
0.125	-14.4276	clearly | a
0.156863	-14.4282	3 | no
0.230769	-14.4282	p(i | 1|k)×
0.428571	-14.4284	approaches | explored
0.0965517	-14.4285	on | current
0.0523139	-14.4286	sentence | by
0.0759494	-14.4288	as | document
0.375	-14.4288	em, | em
0.5	-14.4291	mwen | tranbleman
0.5	-14.4291	similarity | implies
0.0882353	-14.4292	the | 10%
0.126582	-14.4296	from | existing
0.0545852	-14.4296	simple | in
0.307692	-14.4297	(chiang, | 2008;
0.0824742	-14.4297	or | its
0.037336	-14.4297	by | an
0.102362	-14.4299	thus, | with
0.104839	-14.4302	for | center
0.5	-14.4303	p(wi|wi−1i−n+1)
0.5	-14.4303	site | promoter
0.0208268	-14.4304	the | data.
0.375	-14.4306	topic | coherence
0.101562	-14.4307	of | answers
0.0769231	-14.4308	we | experimental
0.285714	-14.4309	proposed | times
0.0698529	-14.4309	mentions | and
0.131579	-14.431	an | study
0.05	-14.4314	we | words | a
0.25	-14.4315	even | standard
0.333333	-14.4315	unsupervised | sentiment
0.375	-14.4315	mentions | avg
0.2	-14.4315	denotes | set
0.145161	-14.4319	method, | for
0.0820513	-14.4319	in | add
0.0684932	-14.4319	the | statistical | translation
0.147541	-14.4319	figure 2 | an
0.238095	-14.432	state-of-the-art | task
0.0762712	-14.4321	topics | a
0.0703704	-14.4321	model | source
0.045045	-14.4321	1 | on
0.0634441	-14.4322	by | research
0.103175	-14.4322	graph | from
0.0714286	-14.4323	third | of
0.375	-14.4324	general | documents.
0.375	-14.4324	referred | referred
0.117021	-14.4326	it | identify
0.189189	-14.4326	r | using
0.131579	-14.4326	to | assignment
0.0984848	-14.4327	attempt | a
0.140625	-14.4329	if | another
0.0933333	-14.4329	each | they
0.189189	-14.433	collection | our
0.126582	-14.4333	target | but
0.375	-14.4333	pairs | matches
0.107143	-14.4334	rules | an
0.189189	-14.4335	reduced | that
0.428571	-14.4336	centering | (grosz
0.114286	-14.4337	pairs | sentence
0.133333	-14.4338	a | handle
0.0972222	-14.434	play | in
0.175	-14.434	this | formal
0.25	-14.434	(b) | two
0.0471545	-14.434	of | paper
0.0682594	-14.434	are | terms
0.15	-14.4341	0.5 | 0.4
0.0921053	-14.4342	as | graph
0.375	-14.4342	enables | source
0.0829016	-14.4343	of | state-of-the-art
0.0309206	-14.4344	extraction | the
0.0538793	-14.4346	they | on
0.0440341	-14.4347	is | phrase
0.428571	-14.4347	philip | a
0.285714	-14.435	(see | 2),
0.0356473	-14.4351	such | for
0.5	-14.4351	0.12 | 0.18
0.166667	-14.4352	denote | number of
0.12987	-14.4352	on | (section
0.0246969	-14.4355	we | all | the
0.0952381	-14.4357	that | social
0.25	-14.4358	all | belong
0.16	-14.436	when | specific
0.5	-14.4363	consistent | observations
0.285714	-14.4364	space | since
0.666667	-14.4364	j− | j+
0.333333	-14.4364	treelets | be
0.039627	-14.4365	results | table
0.147541	-14.4365	results | improvements
0.0731707	-14.4367	5 | from
0.0575	-14.4367	defined | =
0.428571	-14.4367	crichton | 9ab,
0.189189	-14.4367	grammar | tree
0.120879	-14.4368	as | position
0.11828	-14.437	we | detail
0.0794393	-14.4371	first | set
0.153846	-14.4371	our | capture
0.104839	-14.4371	we | structured
0.140625	-14.4374	see | al.
0.0332522	-14.4374	meaning | the
0.0479452	-14.4377	the | are used
0.285714	-14.4378	ngd-google | ngd-factiva
0.0851064	-14.4379	first | one
0.145161	-14.4379	good | or
0.0102316	-14.4381	a | of | a
0.0737705	-14.4383	of | confidence
0.0529661	-14.4384	and | following
0.12987	-14.4384	this | global
0.285714	-14.4385	corresponds | single
0.285714	-14.4385	work | words,
0.263158	-14.4387	users | dialogue
0.333333	-14.4388	0.06 | 0.04
0.333333	-14.4388	potentially | have
0.5	-14.4388	in | acm,
0.120879	-14.4389	symbol | in
0.0719697	-14.4389	of | <
0.0743802	-14.439	and | against
0.0851064	-14.4391	on | scores
0.222222	-14.4391	baseline | system.
0.0779817	-14.4391	the | virtual
0.156863	-14.4396	addition, | an
0.0725806	-14.4397	and | categories
0.0762332	-14.4397	to | pattern
0.222222	-14.4397	baseline | majority
0.0525253	-14.4399	the | too
0.428571	-14.4399	2: | h
0.0419948	-14.4399	test | a
0.15	-14.4399	an | distance
0.0221738	-14.4399	while | the | the
0.0898204	-14.44	an | document
0.0148172	-14.4401	is | to | and
0.307692	-14.4402	explore | lexical
0.0592784	-14.4404	— | the
0.222222	-14.4407	increasing | training
0.12987	-14.4412	parallel | corpora
0.5	-14.4412	quantify | against
0.0992366	-14.4416	performance | both
0.2	-14.4417	will | examples
0.163265	-14.4418	are | score.
0.114583	-14.4419	over | test
0.0877193	-14.4419	at | other
0.147541	-14.4419	complexity | be
0.115789	-14.4421	w | 0
0.114583	-14.4422	this | partially
0.2	-14.4422	gain | not
0.0867052	-14.4424	and | structure.
0.175	-14.4425	work | web
0.057072	-14.4427	the | author
0.128205	-14.4428	model | sets
0.0273438	-14.4429	to | which | to
0.0546697	-14.4429	of | e
0.0668896	-14.4433	the | 2005)
0.0746888	-14.4433	closely | the
0.0779817	-14.4437	a | them.
0.0725806	-14.4439	is | because | a
0.227273	-14.444	** | *
0.108108	-14.444	finite | null
0.108108	-14.444	such | system
0.285714	-14.4441	2.2 | machine
0.285714	-14.4441	lc | s
0.238095	-14.4442	less than | less
0.0225056	-14.4445	a | all
0.263158	-14.4445	from | pair.
0.170732	-14.4448	measure | use
0.11828	-14.4448	we | ontology
0.5	-14.4449	dipanjan | 15213,
0.0725806	-14.445	that | errors
0.142857	-14.4451	information | any
0.375	-14.4452	obtained | mt
0.0233683	-14.4453	training | and
0.0660066	-14.4455	workshop | 2011.
0.080402	-14.4456	in | identifying
0.206897	-14.4456	common | only
0.156863	-14.4458	search | best
0.0677966	-14.4459	as | probability
0.0518962	-14.446	model | language
0.0348944	-14.4461	to | results
0.666667	-14.4462	˜pn(wi|wi−1i−n+1) | note
0.0431154	-14.4463	or | by
0.117021	-14.4463	term | with
0.0779817	-14.4464	this | described
0.1	-14.4466	al., | 2007),
0.0903614	-14.4466	needed to | the
0.2	-14.4466	fact | system
0.113208	-14.4467	types | features
0.25	-14.4468	in | met
0.166667	-14.447	| | q
0.0703704	-14.447	with | model.
0.0621302	-14.447	results | an
0.375	-14.447	a | stringent
0.1	-14.4471	at | given
0.214286	-14.4471	shown | sets
0.106195	-14.4472	trained | features
0.15	-14.4472	• | both
0.5	-14.4473	authors | apprentice
0.0915033	-14.4475	more | these
0.00759483	-14.4475	of | of | of
0.0872093	-14.4475	the | 2003),
0.179487	-14.4476	although | will
0.103175	-14.448	our | classifier
0.128205	-14.4481	• | training
0.0882353	-14.4483	20 | in
0.307692	-14.4485	expressions | specific
0.307692	-14.4485	2. | every
0.145161	-14.4485	languages | more
0.5	-14.4485	affect | (dal)
0.5	-14.4485	entities | context-sensitive
0.0619469	-14.4486	this | has | to
0.333333	-14.4486	end | –
0.238095	-14.4488	j | 1.
0.128205	-14.4491	concepts | be
0.114583	-14.4492	uses | word
0.0623145	-14.4493	are | lexical
0.0607735	-14.4495	each | these
0.0867052	-14.4496	(see | in
0.25	-14.4498	 | for
0.5	-14.4498	decide | continue
0.0521921	-14.4499	1 introduction | to
0.0429958	-14.45	extracted | the | of
0.307692	-14.45	arr | ize
0.0467446	-14.45	we | methods
0.114583	-14.4501	shows | sentence
0.0664452	-14.4501	trees | to
0.153846	-14.4503	i | like
0.0668896	-14.4503	three | that
0.108108	-14.4504	among | with
0.184211	-14.4506	plays | important
0.307692	-14.4507	2004), | semantic
0.122222	-14.4507	many | at
0.375	-14.4507	ner | named
0.0903226	-14.4509	single | from
0.120879	-14.451	1 | distribution
0.5	-14.451	obtain | aspects
0.238095	-14.4512	the | reordered.
0.120879	-14.4513	object | as
0.193548	-14.4513	their | small
0.0358875	-14.4515	that | information
0.263158	-14.4516	word | bigrams
0.0722892	-14.4517	and | summarization
0.285714	-14.4518	jj | vb
0.104348	-14.4518	because | number
0.0619469	-14.452	which | their
0.0296443	-14.4521	consider | the
0.307692	-14.4522	rule | systems
0.140625	-14.4522	3 | over
0.0628743	-14.4523	for | documents
0.2	-14.4525	appropriate | can
0.333333	-14.4527	5 | induction
0.0701107	-14.4528	is | difference
0.103175	-14.4528	the | markers
0.25	-14.4529	thank | feedback
0.0977444	-14.4529	our | annotated
0.125	-14.4529	standard | data
0.166667	-14.4532	semantic | types
0.175	-14.4532	illustrates | for
0.101562	-14.4532	small | on
0.0664452	-14.4533	also | by | and
0.0151515	-14.4535	and | from
0.238095	-14.4535	in | weather
0.25	-14.4535	the | greater.
0.428571	-14.4535	appear | fewer
0.15	-14.4535	pred | 
0.375	-14.4535	obtain | pairs.
0.0144343	-14.4537	first | the
0.101562	-14.4537	70 | of
0.285714	-14.4539	(mcdonald | 2006;
0.0867052	-14.4539	are | general
0.175	-14.4541	accuracy | trained
0.0970149	-14.4542	be | useful
0.142857	-14.4542	similarity | than
0.184211	-14.4543	body | work
0.0321821	-14.4543	the | of these
0.0253225	-14.4544	the | corresponding
0.136364	-14.4545	commonly | as
0.285714	-14.4546	mt | should
0.428571	-14.4546	f1 | 81.2
0.113402	-14.4547	represented | with
0.0281269	-14.4547	some | and
0.5	-14.4547	is | in”
0.05	-14.4548	transition | the
0.0725806	-14.455	applied | et
0.106195	-14.4551	baseline | an
0.193548	-14.4552	if | words,
0.307692	-14.4553	dependency | mstparser
0.170732	-14.4553	this | identical
0.042641	-14.4554	for | will
0.078341	-14.4556	in | hierarchical
0.428571	-14.4556	while | corresponds
0.0977444	-14.4558	accuracy | our
0.227273	-14.4558	labeled | l
0.0442478	-14.4559	the | count
0.285714	-14.456	may | valid
0.666667	-14.456	false | rumor
0.333333	-14.456	0.04 | 3:
0.0475382	-14.4561	the | of | among
0.11215	-14.4561	task, | that
0.136364	-14.4562	its | also
0.0731707	-14.4565	task | an
0.0274664	-14.4565	to | system
0.0857143	-14.4565	or | these
0.156863	-14.4566	for | cell
0.25	-14.4566	previous | showed
0.175	-14.4567	been | e.g.
0.428571	-14.4567	an | messages.
0.0523013	-14.4568	more | one
0.0734694	-14.4569	we | distance
0.163265	-14.457	0.4 | 0.9
0.0246238	-14.457	the | tree | the
0.0882353	-14.457	application | a
0.189189	-14.4571	to | familiar
0.25	-14.4572	zij | the
0.12987	-14.4572	from | random
0.375	-14.4572	informatics | japan
0.375	-14.4572	) | k=1
0.163265	-14.4574	split | data
0.0345455	-14.4574	text | to
0.263158	-14.4574	concepts | knowledge
0.12987	-14.4575	also | first
0.175	-14.4576	comparable | translation
0.128205	-14.4576	our | decoding
0.156863	-14.4578	given | event
0.100775	-14.458	vocabulary | a
0.0887574	-14.4581	measure | as
0.285714	-14.4582	v | det
0.0984848	-14.4583	relations | from
0.0302198	-14.4583	while | to
0.100775	-14.4585	a | svm
0.333333	-14.4585	follows: | cluster
0.0298913	-14.4588	that | all
0.214286	-14.4588	distribution | 1.
0.119565	-14.4589	above | on
0.206897	-14.4589	both | <
0.100775	-14.459	if | may
0.125	-14.4591	a | shift
0.0596206	-14.4592	shows | on
0.134328	-14.4592	from | models.
0.333333	-14.4593	topics | then
0.0812183	-14.4593	document | with
0.11828	-14.4597	g | )
0.115789	-14.4598	+ | feature
0.119565	-14.4598	train | language
0.0645161	-14.4598	former | the
0.150943	-14.46	mutual | between
0.193548	-14.46	7 | 15
0.375	-14.46	translating | translating
0.0374332	-14.4601	evaluate | of
0.0877193	-14.4602	alignment | from
0.0852273	-14.4602	other | but
0.110092	-14.4602	results | improvement
0.0636364	-14.4607	allows | in
0.0792079	-14.4608	the | reverse
0.666667	-14.461	brand | color
0.222222	-14.461	i | work
0.150943	-14.4611	for | without | for
0.0708661	-14.4611	lexical | by
0.0599455	-14.4611	in | recall
0.105263	-14.4612	their | different
0.11828	-14.4612	significant | p
0.0355086	-14.4616	4 | in
0.189189	-14.4618	feature | opinion
0.375	-14.4619	al., | (miwa
0.0824742	-14.4623	it | text
0.136364	-14.4623	p | d
0.238095	-14.4624	thus, | target
0.078341	-14.4626	association | computer science
0.0199697	-14.4626	number of | of
0.666667	-14.4627	seki | al.
0.156863	-14.4628	first | search
0.375	-14.4628	great | do
0.109091	-14.4629	for | utterance
0.0344828	-14.4631	each | are
0.0257316	-14.4632	into | of
0.0615836	-14.4634	data, | a
0.142857	-14.4634	human | evaluation
0.5	-14.4634	ohta∗ | tsujii†
0.5	-14.4634	encoding | structures
0.0405577	-14.4635	of | alignment
0.0862069	-14.4636	as | multiple
0.189189	-14.4637	correct | was
0.0816327	-14.4637	data | word
0.0977444	-14.4638	it was | a
0.0490018	-14.4641	) | (
0.0824742	-14.4643	or | language
0.108108	-14.4643	while | information
0.0420624	-14.4644	verbs | the
0.121951	-14.4644	of | interpreted
0.285714	-14.4646	80 | average
0.5	-14.4646	defined | radicalbig
0.121951	-14.4647	to | medical
0.0572139	-14.4649	distribution | is
0.0418354	-14.4649	this | which | to
0.0137511	-14.4649	features | the
0.0769231	-14.4653	contexts | of
0.109091	-14.4656	that | e.g.
0.0362903	-14.4656	focus | the
0.08	-14.4657	expression | and
0.214286	-14.4657	due | high
0.193548	-14.4658	need | use
0.263158	-14.4659	select | set.
0.333333	-14.4659	shift | 220
0.333333	-14.4659	cj) | ci(
0.333333	-14.4659	micro | ranking
0.166667	-14.466	were | without
0.666667	-14.466	np@1 | dot
0.666667	-14.466	logloss | hatwidew
0.666667	-14.466	als | euro”
0.666667	-14.466	ba | ca
0.666667	-14.466	0.01, | 0.05
0.666667	-14.466	+2 | -2
0.666667	-14.466	joint | r+l+q
0.666667	-14.466	@ | ¯an(6)
0.666667	-14.466	accepted(x,z))) | execution-problem(y,z)))
0.666667	-14.466	73.2 | barakfull
0.666667	-14.466	vectorwt | vectorwt−1
0.666667	-14.466	−bik | bik
0.666667	-14.466	sub-genre | doc
0.666667	-14.466	[特别 | [行政区
0.666667	-14.466	(denoted | blocks),
0.666667	-14.466	marker | prefacing
0.666667	-14.466	os | t^
0.666667	-14.466	pron | fork-style
0.666667	-14.466	winter | sumer
0.666667	-14.466	so-wtp | lexicon
0.666667	-14.466	constructor | ,zn
0.666667	-14.466	mathematics | icmc.usp.br
0.666667	-14.466	(x,y) | l(ˆyj,y)
0.666667	-14.466	v6 | united
0.666667	-14.466	faculty | 1105,
0.666667	-14.466	j=1 | tσvj
0.666667	-14.466	0.71300 | 0.82598
0.666667	-14.466	-null+aalign | -null-aalign
0.666667	-14.466	n(c−(m,n) | n(c−(m,n))+ηc,
0.666667	-14.466	coarse | p(x,i,j|s)
0.666667	-14.466	np↓ | wore
0.666667	-14.466	sβ> | tris
0.666667	-14.466	a0a1 | a20
0.666667	-14.466	simhash | (charikar
0.666667	-14.466	β(v) | β(d)
0.666667	-14.466	optimistic | overlay
0.666667	-14.466	ouvrit | porte
0.666667	-14.466	d∈d | tl∈tln∈tn
0.666667	-14.466	“gr¨un | weiß
0.666667	-14.466	colossus | japanese-developed
0.666667	-14.466	computevectorai | vi←f(vectorai)
0.666667	-14.466	∀ci∈s,(li | ∈t⇒(lj
0.666667	-14.466	k=s | ρjj(1)
0.666667	-14.466	2-nd | 3-rd
0.666667	-14.466	flies | iarrowi
0.666667	-14.466	9.9m | 39m
0.666667	-14.466	linear | 0.395
0.666667	-14.466	(w/ | non-terminal)
0.666667	-14.466	0.700 | 0.866∗
0.666667	-14.466	62.6 | supervised
0.666667	-14.466	广场 | 了
0.666667	-14.466	−1−0.8−0.6−0.4−0.2 | 0.20.40.60.8
0.666667	-14.466	joy(wna) | negemo(pp)
0.666667	-14.466	ran2hou4 | jian3cha2
0.666667	-14.466	v(w) | pl(w)
0.666667	-14.466	lij | w(i,j)
0.666667	-14.466	ningbo | xitianxin
0.666667	-14.466	174.59 | 66.94
0.666667	-14.466	debates | heated
0.666667	-14.466	∨ | (i,j),(j,i)
0.666667	-14.466	a∈ω | nullnull(nullnullw|nulltnullnullt)
0.666667	-14.466	’’ | ,,
0.666667	-14.466	adi-r: | ados:
0.666667	-14.466	mexico | ingeniería
0.666667	-14.466	engineering, | dong-a
0.666667	-14.466	92.8 | (75)
0.666667	-14.466	on/s | bus/sa
0.666667	-14.466	reaction | lamp”),
0.666667	-14.466	zone | jeremie
0.666667	-14.466	paraphrav | paraphrsum
0.666667	-14.466	x,yϵs | x,yϵs(gtm(x,y)
0.666667	-14.466	θdoc | φrx
0.666667	-14.466	def≡ | ⟨1,k1,e1,x⟩
0.666667	-14.466	ˆg | unegationslash=v
0.666667	-14.466	0.28 | 0.76
0.666667	-14.466	5-gram/2-slm | 2-gram/4-slm,
0.666667	-14.466	score(s) | s∗=
0.666667	-14.466	“ee”, | “aa”,
0.666667	-14.466	up:1 | down:1>
0.666667	-14.466	ai | oa
0.666667	-14.466	texas | 1890s
0.666667	-14.466	cl | looking-center
0.666667	-14.466	veproductdisplay | vfproductdisplay
0.666667	-14.466	ˆu | ˆv
0.666667	-14.466	ed | (syn
0.666667	-14.466	juan | (juan
0.666667	-14.466	moment | inertia.
0.666667	-14.466	chicago | zhejiang
0.666667	-14.466	a29a36a40 | a29a36a45
0.666667	-14.466	match | context=
0.666667	-14.466	nianwen | university,
0.666667	-14.466	erabakiak | erabakirik
0.666667	-14.466	m∏ | i−n∏
0.666667	-14.466	s48 | s99s121
0.666667	-14.466	entity-a | dobj
0.666667	-14.466	“cananea” | comp.arg1.
0.666667	-14.466	cj | p(cj)
0.666667	-14.466	villeroy | funnyday
0.666667	-14.466	sobre[about] | motive]
0.666667	-14.466	i)) | ~w.
0.666667	-14.466	w(b,u) | w(bnull,unull)
0.666667	-14.466	centre | 38068
0.666667	-14.466	non-match | operation
0.666667	-14.466	disorder | tisue
0.666667	-14.466	“ronaldinho” | milan”
0.666667	-14.466	0.036 | 0.024
0.666667	-14.466	$←we | us→relations
0.666667	-14.466	topic-specific | o(wj)
0.666667	-14.466	movie | product)
0.666667	-14.466	quantifiers | singulars/plurals
0.666667	-14.466	demarcation | bgn-i
0.666667	-14.466	平方公里 | ，
0.666667	-14.466	cost(t) | ri∈t
0.666667	-14.466	quantifier | wide-scope
0.666667	-14.466	plentiful | (eg.,
0.666667	-14.466	+ellipsis | −ellipsis
0.666667	-14.466	association | spp
0.666667	-14.466	∀u,v,w∈v | yuv
0.666667	-14.466	这 | _雨
0.666667	-14.466	amherst, | 01003,
0.666667	-14.466	cm | coerced,
0.666667	-14.466	iwslt, | 49.13
0.666667	-14.466	sift | invariance
0.666667	-14.466	xf+| | n>|/vvn>|
0.666667	-14.466	sarcasm | understatement,
0.666667	-14.466	periphrases | phasal
0.666667	-14.466	0.007% | 0.008%
0.666667	-14.466	3w2t | 2w3t
0.666667	-14.466	come | a.03
0.666667	-14.466	sums | mj−1
0.666667	-14.466	bilinguals | 0.3153
0.666667	-14.466	h(e) | vnull
0.666667	-14.466	faeala | tafaaeala
0.666667	-14.466	buckwalter | dwlar
0.666667	-14.466	多少 | _的
0.666667	-14.466	# | ree
0.666667	-14.466	senor | sc,
0.666667	-14.466	normalization; | bigram
0.666667	-14.466	2,2 | h’
0.666667	-14.466	nullt | (null
0.666667	-14.466	pos3 | pos6
0.666667	-14.466	valid, | indicatingthat
0.666667	-14.466	(kullback | 1951),
0.666667	-14.466	z: | ≻x,
0.666667	-14.466	np-c | denull
0.666667	-14.466	68.20 | parasense2
0.666667	-14.466	49/47 | 62/71
0.666667	-14.466	brechen | 1st-person
0.666667	-14.466	ep | 42.09
0.666667	-14.466	δ([a1,a2] | [b1,b2])
0.666667	-14.466	∑ | sump
0.666667	-14.466	“indoors” | “focus
0.666667	-14.466	cd(s) | scored(r
0.666667	-14.466	realizations | translocation.n
0.666667	-14.466	dissertou[talked] | motivar[to
0.666667	-14.466	ndcg@3 | 0.883
0.666667	-14.466	xsi,j | c·ncf,e
0.666667	-14.466	a−n | s−v
0.666667	-14.466	rosen | batrice
0.666667	-14.466	/n/ | /c/,
0.666667	-14.466	bnc | ‘enjoy
0.666667	-14.466	[3] | volition
0.666667	-14.466	<row | role="lang">en</cell>
0.666667	-14.466	χαρτόκουτο | χαρτίκουτ
0.666667	-14.466	(obozinski | j(xt,vectorβt,vectoryt)
0.666667	-14.466	rd | guelph,
0.666667	-14.466	ld | w,v
0.666667	-14.466	dt-2 | nn-3
0.666667	-14.466	nullnullnullnullnullnull | pnulllnullnullnull
0.666667	-14.466	wd,n | pk,l←pk,l
0.666667	-14.466	costimulation | transactivation
0.666667	-14.466	austronesian | (greenhill
0.666667	-14.466	drawings | illustrations;
0.666667	-14.466	p(zw | k)summationtextkj=1
0.666667	-14.466	punctuation | capitalization.
0.666667	-14.466	labeling | pmaplpt(x,ξ):
0.666667	-14.466	preferences | z∧φ′
0.666667	-14.466	emmanuel | 44322
0.666667	-14.466	atb | (morethantoken
0.666667	-14.466	fine | (s,λm);
0.666667	-14.466	8+ | 10+
0.666667	-14.466	x01 | x11
0.666667	-14.466	noun-noun | <m1
0.666667	-14.466	“nicht | negotiate”
0.666667	-14.466	0.710 | extsim
0.666667	-14.466	testbleu | nobilm
0.666667	-14.466	outlook | stock),
0.666667	-14.466	(3); | (5);
0.666667	-14.466	classify | (pqrac
0.666667	-14.466	aql | (chiticariu
0.666667	-14.466	ball | [scene
0.666667	-14.466	p(c) | p(fi|c)
0.666667	-14.466	interpersonal | “leary’s
0.666667	-14.466	 i |  ii
0.666667	-14.466	403 | (92),
0.666667	-14.466	i∈i, | d(e,i)
0.666667	-14.466	342 | 2´
0.666667	-14.466	[粮食署 | [救济金
0.666667	-14.466	grade | 40-80%
0.666667	-14.466	|b1| | |b2|
0.666667	-14.466	write | irish
0.666667	-14.466	g1875ℎg1857g1870g1857 | g1840g3013
0.666667	-14.466	-nb | -bl
0.666667	-14.466	first-order | 〈fro〉
0.666667	-14.466	washer=person, | theme:instr),
0.666667	-14.466	landmark’s | salience
0.666667	-14.466	insertion | (tig)
0.666667	-14.466	={go | off:p
0.666667	-14.466	π32 | hallway.
0.666667	-14.466	-0-7 | -19
0.666667	-14.466	tempex | dedo
0.666667	-14.466	can't | #killthebill
0.666667	-14.466	s100 | s102
0.666667	-14.466	mit | dein
0.666667	-14.466	m10 | m9,
0.666667	-14.466	[万 | [长官
0.666667	-14.466	generalizations | 2.05,p
0.666667	-14.466	s48s46s48s48 | s114s99
0.666667	-14.466	mwa | )|(
0.666667	-14.466	2nd-person | brach
0.666667	-14.466	citations | three-year
0.666667	-14.466	adjp | npb
0.666667	-14.466	strictly | fk(nullσ∗null)
0.666667	-14.466	kproductdisplay | i=1
0.666667	-14.466	reviews | dvds
0.666667	-14.466	yang | liu.
0.666667	-14.466	k7t | r6-k1
0.666667	-14.466	“informative” | documents)
0.666667	-14.466	nambara | ‘position’
0.666667	-14.466	ip1 | greenwich
0.666667	-14.466	arabic: | turkish:
0.666667	-14.466	adjective-noun | 〈a,n〉,
0.666667	-14.466	θv,w | +epsilon1
0.666667	-14.466	christodoulopoulos, | mark
0.666667	-14.466	z12 | ≥z1
0.666667	-14.466	a29a36a37 | a37a36a29
0.666667	-14.466	infected(y,malaria) | ∃x[mosquito(x)
0.666667	-14.466	s115 | s32s83s72
0.666667	-14.466	[mwv | [adv
0.666667	-14.466	kn:bn-big | rnn-limited
0.666667	-14.466	million | milion˚u
0.666667	-14.466	−hj, | cj≤
0.666667	-14.466	weak | (confluence
0.666667	-14.466	formosan | squliq
0.666667	-14.466	b,2 | c,3
0.666667	-14.466	46% | -1.3%
0.666667	-14.466	l(w) | c(w)
0.666667	-14.466	clc-hdp-lg | clchdp-lg
0.666667	-14.466	(t1, | t4,
0.666667	-14.466	covenants | destruction
0.666667	-14.466	brown | r20
0.666667	-14.466	barakno-cxt | class.-based
0.666667	-14.466	urduenglish | phrase-based
0.666667	-14.466	summationtextvf | ¬j
0.666667	-14.466	arnold | à
0.666667	-14.466	essay | (page,
0.666667	-14.466	80k | ×2k
0.666667	-14.466	u.n. | b-vp
0.666667	-14.466	+fp | +fn
0.666667	-14.466	{q | }≈{q
0.666667	-14.466	policy | teg
0.666667	-14.466	bracehtipupright | xi|zi
0.666667	-14.466	contingency, | subsenses
0.666667	-14.466	phonestring | uniondbl
0.666667	-14.466	w/constrained | wsj∞
0.666667	-14.466	),inverted( | jio
0.666667	-14.466	mos | [itu-t
0.666667	-14.466	node:pos | node:number
0.666667	-14.466	min | ||w||
0.666667	-14.466	dépecer | u=
0.666667	-14.466	connection | ££day
0.666667	-14.466	al>mrykyp | hty
0.666667	-14.466	count(oji) | oji,
0.666667	-14.466	‘you’ | non-rebuttals
0.666667	-14.466	hmepv | bti
0.666667	-14.466	mwn | ‘magic
0.666667	-14.466	{1≤j≤n|pj | 1/ν}
0.666667	-14.466	“he/prp | to/to
0.666667	-14.466	α0 | +α2
0.666667	-14.466	td,k | pk,wd,n
0.666667	-14.466	stymne | 21.63
0.666667	-14.466	puniyani | +λ||b||lscript1/lscript2
0.666667	-14.466	[obj1] | [♣]n−2
0.666667	-14.466	“il | pas”
0.666667	-14.466	september | 0.014
0.666667	-14.466	[ao1 | customs
0.666667	-14.466	ou | qy
0.666667	-14.466	0.008 | 0.031
0.666667	-14.466	⇒ | f(g)
0.666667	-14.466	eprimeprime | (eprime1
0.666667	-14.466	ip | x1:np↓
0.666667	-14.466	mukka | punch’;
0.666667	-14.466	wo | goldfish
0.666667	-14.466	|common | s,c||dependencies
0.666667	-14.466	gtm(y,x) | eqm(x,y))
0.666667	-14.466	,0 | n”
0.666667	-14.466	(ms) | 30003500
0.666667	-14.466	l(r | l(r)
0.666667	-14.466	surface | (t33,
0.666667	-14.466	uterances | au4_10sec
0.666667	-14.466	(%) | 0.10.01
0.666667	-14.466	min | λbardbldbardbl0
0.666667	-14.466	gpe | 0.9401
0.666667	-14.466	84.09 | detector-i
0.666667	-14.466	massachusetts | )
0.666667	-14.466	a∈[b,c] | aanegationslash∼α
0.666667	-14.466	)(# | ∩=
0.666667	-14.466	swear(pp) | touser(pr)
0.666667	-14.466	0.81.0 | 25%0%
0.666667	-14.466	pf(k|j) | η(j−k)
0.666667	-14.466	uz | uzh,
0.666667	-14.466	br | φrev
0.666667	-14.466	iarg0 | iarg1.
0.666667	-14.466	“war” | lasts
0.666667	-14.466	ˆk0 | ,k,n
0.666667	-14.466	(pqrab | nqrab)
0.666667	-14.466	pizzas | (cut
0.666667	-14.466	[] | top3
0.666667	-14.466	fertf(a)≤1 | fertf(a)≤kj
0.666667	-14.466	_？ | _些
0.666667	-14.466	∂f2 | ∂¯h1
0.666667	-14.466	0.246 | 0.284
0.666667	-14.466	lexical+prod-rule[both](n=2000) | lexical+prod-rule[both]*(n=2000)
0.666667	-14.466	expand | ha1
0.666667	-14.466	s/ε | e/ε,
0.666667	-14.466	p(lhs | r−→rhs)p(lhs
0.666667	-14.466	lfg | snti
0.666667	-14.466	cj | ci+1
0.666667	-14.466	tr | 39.3
0.666667	-14.466	(erk | dinu
0.666667	-14.466	3rd-person | brachen
0.666667	-14.466	dos | netware
0.666667	-14.466	o(r,n,d,s) | dsummationdisplay
0.666667	-14.466	age | variable.
0.666667	-14.466	svr | cristianini
0.666667	-14.466	111.46 | 65.15
0.666667	-14.466	plant | factory}
0.666667	-14.466	l∈l−˚ι | contextest(˚ι,l)
0.666667	-14.466	logic1 | logic2
0.666667	-14.466	pbta | l(a)
0.666667	-14.466	pos(mi,mj) | neg(mi,mj)
0.666667	-14.466	headword | doc(p)
0.666667	-14.466	k-means | (mkm)
0.666667	-14.466	n(¯x−(i,j) | xij)
0.666667	-14.466	fb2(rs2 | rt3),
0.666667	-14.466	tsummationdisplay | dsummationdisplay
0.666667	-14.466	false: | vuniondisplay
0.666667	-14.466	‘x | ⇔
0.666667	-14.466	414-letter | accuracy(%)on
0.666667	-14.466	form), | wo,
0.666667	-14.466	ρmonths(sy/m,d) | 28···
0.666667	-14.466	ancn1(r)) | c(r
0.666667	-14.466	#true | graph#edges
0.666667	-14.466	modtag | len(a)
0.666667	-14.466	venue | (scraped
0.666667	-14.466	possessive | [u?
0.666667	-14.466	p(yi | good|s)
0.666667	-14.466	queries): | presidents,
0.666667	-14.466	krengle | 658
0.666667	-14.466	),,( | βα
0.666667	-14.466	book | “philipp
0.666667	-14.466	w∈si | 2w
0.666667	-14.466	{λk}4k=1 | {τk}4k=1
0.666667	-14.466	organisation | names”
0.666667	-14.466	allv | nv|zi
0.666667	-14.466	β | g(α,at(β,λ))
0.666667	-14.466	abs | nump<>hartu
0.666667	-14.466	e(x) | d(x)
0.666667	-14.466	highest-scored | 0::e(e
0.666667	-14.466	computed | pb/dhpb
0.666667	-14.466	sair | f´erias
0.666667	-14.466	responsible | fuzenullnull
0.666667	-14.466	corenlp, | stemmer,
0.666667	-14.466	(raw)-english | combinations)
0.666667	-14.466	jigs,e1) | ˆpmle(ice
0.666667	-14.466	tectomt | (popel
0.666667	-14.466	mni | hni
0.666667	-14.466	以色列 | 巴勒斯坦
0.666667	-14.466	impressions | “optimistic
0.666667	-14.466	reduces | mass),
0.666667	-14.466	pou | pase
0.666667	-14.466	‘x | obj−−→
0.666667	-14.466	reordered: | spf
0.666667	-14.466	c2cc05c19c09c04c2dc06c2ec19c09c0d | c24c05c06c2fc19c09c0dc07c1f
0.666667	-14.466	jmaxs=1 | j−1summationdisplay
0.666667	-14.466	aspectual | attaque
0.666667	-14.466	hah | 0.714
0.666667	-14.466	avery | puppy.
0.666667	-14.466	roles | frame-specific
0.666667	-14.466	0.72 | 0.77
0.666667	-14.466	at, | ¯as
0.666667	-14.466	sprime | uprime(s,a)
0.666667	-14.466	∃β,φ | ¬φ)
0.666667	-14.466	30.0 | 30.0
0.666667	-14.466	nil) | t
0.666667	-14.466	osition | rt-of-sp
0.666667	-14.466	emit | constraint’s
0.666667	-14.466	fe | htrain
0.666667	-14.466	non-initial | null-links
0.666667	-14.466	bel(a, | anulltask
0.666667	-14.466	.80 | .07
0.666667	-14.466	uk | ←world
0.666667	-14.466	(basili | bloehdorn
0.666667	-14.466	arabic | märgner
0.666667	-14.466	ysi+1 | si−1
0.666667	-14.466	phrasal | e1,e2
0.666667	-14.466	[les | combine([les
0.666667	-14.466	hai | pahaada
0.666667	-14.466	propbank | hunter’s
0.666667	-14.466	compounds | splitting.
0.666667	-14.466	beihan2 | have3
0.666667	-14.466	(“henry | moder”)
0.666667	-14.466	dev | 503×16
0.666667	-14.466	1511 | 3436
0.666667	-14.466	nqrac) | 0));
0.666667	-14.466	move | at+1
0.666667	-14.466	face | “statement
0.666667	-14.466	np-4 | nn-6
0.666667	-14.466	(and) | (saffron)
0.666667	-14.466	step | (update
0.666667	-14.466	stemmer, | dict=dictionary
0.666667	-14.466	4.62 | 2.32
0.666667	-14.466	bilm | 30.37
0.666667	-14.466	27.15 | 57.20
0.666667	-14.466	null) | null:
0.666667	-14.466	none), | aspectual,
0.666667	-14.466	1e+07 | defelicegigaword
0.666667	-14.466	truthful | tripadvisor.
0.666667	-14.466	london, | 2009}
0.666667	-14.466	prenominal | π(b)
0.666667	-14.466	 | [β,γ]squigglerightepsilon1
0.666667	-14.466	a158 | a164
0.666667	-14.466	“vioxx | arthritis”
0.666667	-14.466	‘jackal’ | ‘kind-of’
0.666667	-14.466	nan | ginyin
0.666667	-14.466	[bleues]::[blue]) | [a1]::[jj1])
0.666667	-14.466	s114s101 | s78s117s109s98s101s114s32s111s102s32s98s105s116s115
0.666667	-14.466	dongguan | 8400
0.666667	-14.466	informatics, | edinburgh,
0.666667	-14.466	def≡ | y1∈y1
0.666667	-14.466	fishing | auger
0.666667	-14.466	“tom | flu”
0.666667	-14.466	subordinate | post–verbal
0.666667	-14.466	tv | dv
0.666667	-14.466	cow | ‘moo’
0.666667	-14.466	entailed, | entailment,
0.666667	-14.466	ali | (ali
0.666667	-14.466	manual: | automatic:
0.666667	-14.466	pagerank | e(wj,wi)
0.666667	-14.466	esl | 0.722383
0.666667	-14.466	τ(σ) | docu
0.666667	-14.466	ez | izan
0.666667	-14.466	<frag> | </frag>
0.666667	-14.466	(∗ | label/node)
0.666667	-14.466	γ(w) | minx
0.666667	-14.466	prep=loc | arg2
0.666667	-14.466	dtnnjj | innnp
0.666667	-14.466	s48s46s48s51 | s80
0.666667	-14.466	[b,c]squigglerightepsilon1 | 〈x,ξ〉matches
0.666667	-14.466	z∈z | p(z)p(x,y,z)p(x,z)p(y,z)
0.666667	-14.466	state: | "city"
0.666667	-14.466	(steps | regulation))
0.666667	-14.466	 | far;
0.666667	-14.466	rocket | mortar
0.666667	-14.466	tpattr | attr′
0.666667	-14.466	variable-length | deltas,
0.666667	-14.466	fy | φy
0.666667	-14.466	says | airstrikes
0.666667	-14.466	(%) | per-capita
0.666667	-14.466	if˚ι | begun.
0.666667	-14.466	markov | abduction.
0.666667	-14.466	finger | %gesture:
0.666667	-14.466	727 | (375),
0.666667	-14.466	correctly | ˆrt
0.666667	-14.466	medicinsk | universitet
0.666667	-14.466	random | invertebrates.
0.666667	-14.466	lperm | nb+em
0.666667	-14.466	ball | pink11,pass(pink1,pink11)〉.
0.666667	-14.466	(decodepat | walked))
0.666667	-14.466	(the | step)
0.666667	-14.466	usprimev | sprime∈l
0.666667	-14.466	det2 | +fn*ngr
0.666667	-14.466	strategic | ministry
0.666667	-14.466	sub(s) | w∈s
0.666667	-14.466	hall, | arena,
0.666667	-14.466	observations | predictions’
0.666667	-14.466	pred | ast
0.666667	-14.466	rgprime1 | rgprime2
0.666667	-14.466	token | -2,-1,+1
0.666667	-14.466	request | no-request
0.666667	-14.466	movies, | renaissance
0.666667	-14.466	[美元 | [行政
0.666667	-14.466	a(j) | commentj,
0.666667	-14.466	k=s | ρkj)
0.666667	-14.466	“have | lost”,
0.666667	-14.466	n#05587878 | 07516354-n
0.666667	-14.466	rst | pardo
0.666667	-14.466	predicate-argument | arg1:
0.666667	-14.466	node@7276 | vbddtnn“with”np
0.666667	-14.466	que | una
0.666667	-14.466	œ | ’e
0.666667	-14.466	r-su4 | hybhsum
0.666667	-14.466	c−1), | (c0
0.666667	-14.466	dans | rehab
0.666667	-14.466	b(d,q) | ·p(m|d,e,f)
0.666667	-14.466	translations: | tall”,
0.666667	-14.466	55/48 | 42/42
0.666667	-14.466	ended | finished
0.666667	-14.466	jnullk | b[i,k]
0.666667	-14.466	synonymy | synonym/not
0.666667	-14.466	tradreg | alignment+ind
0.666667	-14.466	<tense>pres | <arg0>
0.666667	-14.466	#esp | #worldcup”
0.666667	-14.466	lr | 86:9%
0.666667	-14.466	chunk | null0,
0.666667	-14.466	u.s.a., | israeli
0.666667	-14.466	80.0 | (32)
0.666667	-14.466	“not-before” | “not-after”
0.666667	-14.466	(e2) | (e18)
0.666667	-14.466	g∈g | i,j∈g
0.666667	-14.466	line | calculations
0.666667	-14.466	“[vn | [mwv
0.666667	-14.466	10000+ | hag
0.666667	-14.466	michel galley, | marcu,
0.666667	-14.466	91.01 | 91.09
0.666667	-14.466	blog | 19,320
0.666667	-14.466	cr | amcr
0.666667	-14.466	olhar | (look
0.666667	-14.466	mood=ind | cat=prep
0.666667	-14.466	(* | *)
0.666667	-14.466	google | maxpaths*
0.666667	-14.466	fd+fs+ft | (***)
0.666667	-14.466	modality | (palmer,
0.666667	-14.466	nmod | prd
0.666667	-14.466	p(w1|w2) | p(w1|¬w2)
0.666667	-14.466	pull | weapons
0.666667	-14.466	2+ | 5+
0.666667	-14.466	qui | originales
0.666667	-14.466	φ(null) | null(nullnull)
0.666667	-14.466	→pcfg-lapipeline | none
0.666667	-14.466	(q | rm)
0.666667	-14.466	kwbl2 | 0.5652
0.666667	-14.466	bombing | gadafhi
0.666667	-14.466	gp | gq
0.666667	-14.466	r−→rhs|t | r−→rhs)
0.666667	-14.466	γ∈γ(w) | 1p(w)·
0.666667	-14.466	p(l | r−→r)p(l
0.666667	-14.466	recall | unspecialized
0.666667	-14.466	je | deux
0.666667	-14.466	header | block;
0.666667	-14.466	book} | obj-bd{an
0.666667	-14.466	nist | (1,859
0.666667	-14.466	|ct|summationdisplay | −r
0.666667	-14.466	childof | moder”,
0.666667	-14.466	˜null) | recomputescore(˜null)
0.666667	-14.466	tverb1 | tverb2
0.666667	-14.466	derivative | wxi
0.666667	-14.466	pklij | ∀i,j,k,l
0.666667	-14.466	pr(p′|p)
0.666667	-14.466	9,9also | 8,8distribution
0.666667	-14.466	won | wereldbeker
0.666667	-14.466	argi | l
0.666667	-14.466	xprime<x,yprime≤y | x,yprime
0.666667	-14.466	(f(s) | x(sprime,g)
0.666667	-14.466	room | ‘go
0.666667	-14.466	1sg | 1pl
0.666667	-14.466	advp | 〈vp,
0.666667	-14.466	pragmatic | touser,
0.666667	-14.466	trans-dimers | [t3,
0.666667	-14.466	|y∩y | prime|
0.666667	-14.466	vector) | parasense3
0.666667	-14.466	⃗xj(1)qi | zij
0.666667	-14.466	schema | recovering
0.666667	-14.466	84.18 | 84.18
0.666667	-14.466	null2 | ˆnull1
0.666667	-14.466	bracketrightbig−e | bracketrightbig−
0.666667	-14.466	e1) | e1)))))
0.666667	-14.466	21.8% | 36.9%
0.666667	-14.466	pi∗ | pi:p[pi]=ns
0.666667	-14.466	(c−2 | (c−1
0.666667	-14.466	yen, | yen
0.666667	-14.466	ner | non-tweets
0.666667	-14.466	(sussna, | 1993),
0.666667	-14.466	i. | j”
0.666667	-14.466	tutor | instructional
0.666667	-14.466	7 | “perfectly
0.666667	-14.466	voicing | obstruent
0.666667	-14.466	wi… | wi’
0.666667	-14.466	coordination | (hogan,
0.666667	-14.466	that〈i,j〉is | +1,j〉|〈i,j〉)
0.666667	-14.466	f-ratios | anova
0.666667	-14.466	complaints | automobiles
0.666667	-14.466	babies | nicu
0.666667	-14.466	markables | sucre2010
0.666667	-14.466	angbracketleftbig | np-obj
0.666667	-14.466	art | fem
0.666667	-14.466	version, | receptionist’s
0.666667	-14.466	northeastern | shenyang,
0.666667	-14.466	~ | v-ing”),
0.666667	-14.466	rate, | zechner
0.666667	-14.466	node:pos | parent:pos
0.666667	-14.466	dice’s | 2c(x,y)c(x)c(y),
0.666667	-14.466	you3 | jian1du1
0.666667	-14.466	√ | 47.1
0.666667	-14.466	xwi | ywj
0.666667	-14.466	enjoyed | vvg,
0.666667	-14.466	d101 | d92
0.666667	-14.466	surprisal, | surprisal.
0.666667	-14.466	(representing | frowny
0.666667	-14.466	 | 
0.666667	-14.466	elle | avec
0.666667	-14.466	eparl | crawled1
0.666667	-14.466	mj | (y/n).
0.666667	-14.466	null(nulli|null) | nullnullnullnull)
0.666667	-14.466	dysarthric | (hosom
0.666667	-14.466	=f | ∧f
0.666667	-14.466	harmony | heightfrontness
0.666667	-14.466	rpred | (∑i
0.666667	-14.466	london | chile.
0.666667	-14.466	1.5 | 78.6
0.666667	-14.466	ˆx1, | ˆxr)
0.666667	-14.466	lemma/lexeme | sticka
0.666667	-14.466	there | recurrent | renal
0.666667	-14.466	100.2 | 75%50%
0.666667	-14.466	nullv | nulle
0.666667	-14.466	birlinghoven | augustin
0.666667	-14.466	aj | ω(a)j
0.666667	-14.466	-10 | -9
0.666667	-14.466	rinp | )()()(
0.666667	-14.466	inter-rgb | 69.9
0.666667	-14.466	singers | sanz
0.666667	-14.466	|si|∑ | ·osik
0.666667	-14.466	ffopffr | g),(
0.666667	-14.466	cold | stuffy
0.666667	-14.466	z | p(n|o,c)
0.666667	-14.466	explain) | (explain;
0.666667	-14.466	-1, | -1],
0.666667	-14.466	[firstname] | pou
0.666667	-14.466	precisely | that〈i,e〉∈αi
0.666667	-14.466	raise | build-finish
0.666667	-14.466	v, | τ(v,u)
0.666667	-14.466	pizza | menge
0.666667	-14.466	leak | detonation
0.666667	-14.466	policies | ‘learnt’
0.666667	-14.466	的 | 也
0.666667	-14.466	mbrsc | (case-sensitive)
0.666667	-14.466	e1) | (cause’
0.666667	-14.466	dt | bmp-2
0.666667	-14.466	grm, | 1w
0.666667	-14.466	ui, | widget
0.666667	-14.466	wang | 88:3%
0.666667	-14.466	1% | (320k
0.666667	-14.466	commitment | drohen1-salsa
0.666667	-14.466	i;h1;k | i;h
0.666667	-14.466	braceleftbiggsample | ji∼mult(θ)
0.666667	-14.466	sprime∈d(e)a4 | (d(e)a4)
0.666667	-14.466	+lem+lem | 60.44
0.666667	-14.466	del | cuando
0.666667	-14.466	¯ei,e | ei,e
0.666667	-14.466	(δ | 05)
0.666667	-14.466	avery | accomplice
0.666667	-14.466	angeles | impresario
0.666667	-14.466	pase | janvye
0.666667	-14.466	4ofcc89 | 4ofc89
0.666667	-14.466	87.51 | 92.98
0.666667	-14.466	[的 | [香港
0.666667	-14.466	q↔que | f*ck↔hell
0.666667	-14.466	lucy | rmbt
0.666667	-14.466	(roc) | (det)
0.666667	-14.466	(gi,gj) | |wtgi,l|
0.666667	-14.466	erk | pado’s
0.666667	-14.466	d3(s1) | d1(t2)d
0.666667	-14.466	tri-sibling | q1(v|h,sib,tsib,dir)
0.666667	-14.466	w, | pr(w′|w)
0.666667	-14.466	(church, | (tscore)
0.666667	-14.466	a,1 | a,2
0.666667	-14.466	you3 | north2
0.666667	-14.466	s50s53s48s48s48 | s101
0.666667	-14.466	poetry | old
0.666667	-14.466	lm | 13.01
0.666667	-14.466	pair: | 3,502,485),
0.666667	-14.466	preceding | times;
0.666667	-14.466	69.5 | 70.5
0.666667	-14.466	mapsto→ | xsbar)
0.666667	-14.466	0.547 | 0.357
0.666667	-14.466	f(y) | θ(v,w)
0.666667	-14.466	purf | purb
0.666667	-14.466	quotient | dimensional
0.666667	-14.466	1.96 | 2.04
0.666667	-14.466	word/h] | the/d
0.666667	-14.466	<10 | <100
0.666667	-14.466	0.71 | 0.43
0.666667	-14.466	bootstrapped | brands
0.666667	-14.466	dik | cpk;
0.666667	-14.466	-0.13 | -0.06
0.666667	-14.466	positive) | (false
0.666667	-14.466	78.8 | 79.1
0.666667	-14.466	haizhou | fusionopolis
0.666667	-14.466	e2 | "blast"
0.666667	-14.466	filtering | ‘signal’
0.666667	-14.466	ninth | eighth
0.666667	-14.466	.03 | reluctant,
0.666667	-14.466	+0.6 | ×1.2
0.666667	-14.466	ns(s1) | ns(s2)
0.666667	-14.466	encoding | (tei
0.666667	-14.466	.60 | .59
0.666667	-14.466	(log | )xuv
0.666667	-14.466	o-dep | t-dep
0.666667	-14.466	bears | wolves
0.666667	-14.466	small) | substantial)
0.666667	-14.466	v.pos | u.pos
0.666667	-14.466	g18g367g381g400g286g282g855 | g396g349g448g349g287g396g286g856g296g449g374g856g1005g899
0.666667	-14.466	i(y | (aik
0.666667	-14.466	pittsburgh | 15260,
0.666667	-14.466	“inchide | (en:
0.666667	-14.466	“person | green”
0.666667	-14.466	ό/ | ί
0.666667	-14.466	mbr | mbr-sc-expected
0.666667	-14.466	urdu | (hussain,
0.666667	-14.466	{n | target(e)}
0.666667	-14.466	ground:theme) | target:theme)
0.666667	-14.466	h−i | th−i∗
0.666667	-14.466	summationtextn | logwnpn
0.666667	-14.466	all/most | null(nullnullnull)”,
0.666667	-14.466	n, | sf>,
0.666667	-14.466	waste | manufacture,
0.666667	-14.466	75% | (farmer)
0.666667	-14.466	binomials | conjunct,
0.666667	-14.466	γ20] | [γ1d,
0.666667	-14.466	ilc | corpus);
0.666667	-14.466	aik | ∂τj
0.666667	-14.466	higher-level | k1!k2
0.666667	-14.466	as,s,u,u, | ,
0.666667	-14.466	0.384 | 5.26%
0.666667	-14.466	dan | klein}@cs.berkeley.edu
0.666667	-14.466	“are | “<s>
0.666667	-14.466	fx | φy
0.666667	-14.466	2007 | dang
0.666667	-14.466	t|p | (u|p1,...,u|pk)
0.666667	-14.466	nullv1 | nullv2
0.666667	-14.466	7+ | 9+
0.666667	-14.466	keretanya | kereta
0.666667	-14.466	cut(a,b) | assoc(b,v)
0.666667	-14.466	−logt(¯e| | −logt(
0.666667	-14.466	cat | legal/judicial
0.666667	-14.466	norm | r(b)
0.666667	-14.466	66.9 | 53.9
0.666667	-14.466	d106 | d91
0.666667	-14.466	ensemble, | ensemble.
0.666667	-14.466	s[dcl]\np | (s[dcl]\np)/np
0.666667	-14.466	1|θ) | pt(zi
0.666667	-14.466	*’, | ‘*
0.666667	-14.466	mining | 8799
0.666667	-14.466	smiley | 50000
0.666667	-14.466	k,nk0 | ,g
0.666667	-14.466	[vp:t | [np:a1
0.666667	-14.466	georgila, | lemon
0.666667	-14.466	rule, | 0.05;
0.666667	-14.466	(expansion) | (37)
0.666667	-14.466	pos="w"> | </term>
0.666667	-14.466	selma | knees
0.666667	-14.466	揺り籠 | -墓場
0.666667	-14.466	sudan | ﺔﻓﺎﻘﺜﻟا
0.666667	-14.466	dude | geo:name(h,‘hawaii’)
0.666667	-14.466	(>) | (<)
0.666667	-14.466	anger | 
0.666667	-14.466	20m | llr
0.666667	-14.466	∨ | ∥>3
0.666667	-14.466	morisette | billie
0.666667	-14.466	“event | duration”
0.666667	-14.466	-から | -まで
0.666667	-14.466	cost(ι1) | cost(ι2)
0.666667	-14.466	auger | jigs
0.666667	-14.466	intertranscriber | 0.7057
0.666667	-14.466	proper-noun | (category
0.666667	-14.466	13:38:56 | student:
0.666667	-14.466	14654 | 6.61
0.666667	-14.466	#-sents | #-funcnodes
0.666667	-14.466	exposure | ligand
0.666667	-14.466	weighted | hw/(hw
0.666667	-14.466	qsj | qak
0.666667	-14.466	university | 34500
0.666667	-14.466	nulli | nullj,
0.666667	-14.466	communication | computers.
0.666667	-14.466	mrt | withouttrivial
0.666667	-14.466	0.282 | 0.275
0.666667	-14.466	82.60 | 84.42
0.666667	-14.466	zhuan3men2 | jin4xing2
0.666667	-14.466	cr | a71deletion
0.666667	-14.466	bioinformatics | (ebi)
0.666667	-14.466	p(s|t) | p(t|s).
0.666667	-14.466	/aba/ | 'ba'
0.666667	-14.466	*" | (strict)
0.666667	-14.466	s49 | s50
0.666667	-14.466	arcade | piers,
0.666667	-14.466	α∗ | ˆy)−wt(φ(x,y)−φ(x,
0.666667	-14.466	coffee | ventures,
0.666667	-14.466	c,2 | g,1
0.666667	-14.466	#n-grams | 3gm
0.666667	-14.466	ex | t4]
0.666667	-14.466	in, | inchoative
0.666667	-14.466	:)$ | 0.0109
0.666667	-14.466	kwbl1 | 0.6038
0.666667	-14.466	arrangement, | 3null4
0.666667	-14.466	universit´e | montr´eal,
0.666667	-14.466	cutwidth | nullnullnull(nullnullnullnullnull)
0.666667	-14.466	kidney | transplant
0.666667	-14.466	pnorm | p=2
0.666667	-14.466	hate | millenium
0.666667	-14.466	ef | 81.58
0.666667	-14.466	vals: | likes
0.666667	-14.466	[b | b,a
0.666667	-14.466	b,1 | h,1
0.666667	-14.466	meteor | u.s.-based
0.666667	-14.466	deceptive | (considered
0.666667	-14.466	e-step: | )||
0.666667	-14.466	weighting | tf-irft,
0.666667	-14.466	dendrograms | sub-clusters
0.666667	-14.466	box | sy/m
0.666667	-14.466	ai | wprimei
0.666667	-14.466	cotrain-ner | (plo)
0.666667	-14.466	0.856 | 0.872
0.666667	-14.466	86) | (89,
0.666667	-14.466	46.0 | 47.8
0.666667	-14.466	⊃ | ∧bite(x,y)]
0.666667	-14.466	/b | @s/.
0.666667	-14.466	1gm | 2gm
0.666667	-14.466	(adds | unary
0.666667	-14.466	u111(x,y,cc) | u112(x,y,cc)
0.666667	-14.466	(∑ | +∑j
0.666667	-14.466	∀j | sj)
0.666667	-14.466	∏ | p(z
0.666667	-14.466	(yu1 | bangjiao4,
0.666667	-14.466	raises | batsman
0.666667	-14.466	λm | λm−1
0.666667	-14.466	type="multiw" | ctag="rgp"
0.666667	-14.466	85 | 84.6
0.666667	-14.466	etur | ull
0.666667	-14.466	shadow | playstation
0.666667	-14.466	“je” | “we”
0.666667	-14.466	cost(r) | interaction(r,κ1,κ2)
0.666667	-14.466	(x,t) | e(x,t;θ)
0.666667	-14.466	dive | |||0.05
0.666667	-14.466	hasemotion | hashigherintensity
0.666667	-14.466	quest | players,
0.666667	-14.466	adv{ | friend<n><pl>
0.666667	-14.466	2000; | belz,
0.666667	-14.466	[prn | rflx]
0.666667	-14.466	(root | (vbz
0.666667	-14.466	helsinki | porvoo
0.666667	-14.466	0.426 | incidents.
0.666667	-14.466	-m | -b
0.666667	-14.466	deltas | compressed
0.666667	-14.466	crisis, | crises.
0.666667	-14.466	posscore[r] | negscore[r]
0.666667	-14.466	:-) | _truncated_
0.666667	-14.466	e2 | (r’
0.666667	-14.466	binned | fore1
0.666667	-14.466	bracehtipdownrightbracehtipdownleft | bracehtipupleft
0.666667	-14.466	teprime | (gprime1,gprime2)
0.666667	-14.466	lagrange | β(w,v)
0.666667	-14.466	vn | fixed-window
0.666667	-14.466	“apple” | “apple
0.666667	-14.466	"sunday" | "saying"
0.666667	-14.466	ent | itati
0.666667	-14.466	ibw | 63.35%
0.666667	-14.466	30.52 | 50.07
0.666667	-14.466	2↔to | bday↔birthday
0.666667	-14.466	¥ | [ao2
0.666667	-14.466	αij(0) | k/∈i,j
0.666667	-14.466	examiner | spokeswoman
0.666667	-14.466	re-writing | productdisplay
0.666667	-14.466	‘choose | route’,
0.666667	-14.466	a* | (null),
0.666667	-14.466	pi,aj∈e | pj,ai∈e
0.666667	-14.466	fraser | burnaby
0.666667	-14.466	θt,l | ∼
0.666667	-14.466	g898g364g258g449g258g856g361g449g374g856g1005g853g3 | g396g349g448g286g396g856g286g449g374g856g1005g853
0.666667	-14.466	unseen | framesemantic
0.666667	-14.466	101112131415 | widths
0.666667	-14.466	room} | com-task{i
0.666667	-14.466	durch | dirigirenden
0.666667	-14.466	arts | hv
0.666667	-14.466	[a1,i,k;l(e1)|r(e1)] | [a2,k
0.666667	-14.466	theme(protein | regulon-operon)
0.666667	-14.466	impressions | “surprising
0.666667	-14.466	macedonian>, | <beautiful
0.666667	-14.466	dev | 48.42
0.666667	-14.466	idiom | panch
0.666667	-14.466	 1 |  2
0.666667	-14.466	s(a) | x1,pre(x2),pre(x3)
0.666667	-14.466	´ecole | f-75005,
0.666667	-14.466	inverse | l(e|f)
0.666667	-14.466	students, | others}
0.666667	-14.466	texas* | bales*
0.666667	-14.466	pcpwff | pcpwoff
0.666667	-14.466	t′ | dsp(n)
0.666667	-14.466	people”, | murdered
0.666667	-14.466	u1,v1∈g1 | u2,v2∈g2
0.666667	-14.466	‘cueset | ki’
0.666667	-14.466	m2 | mangoville
0.666667	-14.466	c,1 | f,1
0.666667	-14.466	0.15 | max-length
0.666667	-14.466	—a« | æ
0.666667	-14.466	s2 | f.pl
0.666667	-14.466	ten | caer,
0.666667	-14.466	signs | torture”,
0.666667	-14.466	130014001500160017001800190020000500 | 40004500
0.666667	-14.466	bracket | stack),
0.666667	-14.466	eine | gegessen”
0.666667	-14.466	non-projectable | (+ha),
0.666667	-14.466	uns | withtrivial
0.666667	-14.466	s105s115 | s32s75s76s83s72
0.666667	-14.466	integral | spectrograph
0.666667	-14.466	= | {input
0.666667	-14.466	steelers | winner
0.666667	-14.466	penn2malt | conjashead
0.666667	-14.466	modifier, | headword;
0.666667	-14.466	−∞ | f(k,m,lscript)
0.666667	-14.466	“steve | quarterback
0.666667	-14.466	αiw | βiw
0.666667	-14.466	max( | ρkj(0),
0.666667	-14.466	toronto, | m5s
0.666667	-14.466	(barrault, | memt
0.666667	-14.466	biocuration | mgi
0.666667	-14.466	,null | ,…,null
0.666667	-14.466	ipad | ipads
0.666667	-14.466	omnis | ‘one
0.666667	-14.466	comparisons/system | cmu-hewavitharana
0.666667	-14.466	isolde | his
0.666667	-14.466	mit | andern
0.666667	-14.466	c2)(possible’ | c1)(enter-s2’
0.666667	-14.466	interlanguage | (wil)
0.666667	-14.466	dprime | llbracket〈ø〉rrbracketw
0.666667	-14.466	(u,r)∈eprime1∩eprime2›ptgprime | (usquigglerightr)
0.666667	-14.466	matches | wikipedia;
0.666667	-14.466	+d | j∈in+(vi)
0.666667	-14.466	nulls | nullδt(yts
0.666667	-14.466	vieira | bangladesh,
0.666667	-14.466	ka | gaya
0.666667	-14.466	〈mod〉 | 〈det〉
0.666667	-14.466	coarse | (lmert,λm);
0.666667	-14.466	schatzmeister | partei
0.666667	-14.466	subtree | (n,pi(n))
0.666667	-14.466	np2 | np1
0.666667	-14.466	centre | parc,
0.666667	-14.466	synt | t=
0.666667	-14.466	van | sluis,
0.0359223	-14.4661	with | than
0.05	-14.4663	as | also
0.0872093	-14.4663	errors | as
0.2	-14.4664	6 | these
0.0572139	-14.4667	a | under
0.175	-14.4667	different | 2009).
0.193548	-14.4667	r | given
0.193548	-14.4667	lda, | of
0.333333	-14.4668	(%) | 76
0.333333	-14.4668	)] | θ
0.333333	-14.4668	macro | ranking
0.214286	-14.4668	amount | unlabeled
0.113402	-14.4669	this | resulting
0.150943	-14.4671	it | weight
0.0962963	-14.4671	events, | the
0.166667	-14.4672	this | contribution
0.175	-14.4672	procedure | section
0.25	-14.4672	standard | annotated
0.266667	-14.4673	one | go
0.266667	-14.4673	15176585 | +0
0.11828	-14.4674	a71a71a71 | a71
0.333333	-14.4676	size | avg
0.0842697	-14.4677	give | in
0.0762332	-14.4679	to | aspects
0.170732	-14.4681	to | products
0.0852273	-14.4681	positive | with
0.128205	-14.4682	trained | different
0.333333	-14.4684	jazzy | error
0.175	-14.4685	for | parser,
0.0962963	-14.4685	task | it
0.142857	-14.4688	where | selected | the
0.136364	-14.4688	j | m
0.189189	-14.4689	very | semantic
0.307692	-14.4691	state | q
0.0513347	-14.4692	sequence | and
0.0693431	-14.4692	are | related
0.0612245	-14.4692	entities | in
0.12987	-14.4692	length | sentence
0.0390356	-14.4693	is an | of
0.2	-14.4694	games | and
0.0335052	-14.4695	way | the
0.0314371	-14.4696	analysis | of
0.140625	-14.4697	results | dataset
0.266667	-14.47	because | texts
0.0379464	-14.47	of | per
0.0984848	-14.4701	words | then
0.193548	-14.4701	each | written
0.285714	-14.4703	pa | )
0.375	-14.4703	models | classifier.
0.117021	-14.4704	results | recall
0.0962963	-14.4705	the | expanded
0.170732	-14.4706	implementation | our
0.266667	-14.4707	are | dashed
0.079602	-14.4707	large | are
0.136364	-14.4708	see | features
0.106195	-14.4709	however, | language
0.0977444	-14.4709	cases | be
0.333333	-14.4709	0.70 | 0.81
0.0577428	-14.4711	7 | in
0.140625	-14.4711	automatic | language
0.170732	-14.4711	evaluation | better
0.119565	-14.4712	we | map
0.11828	-14.4712	inference | al.,
0.375	-14.4712	kendall’s | judgments
0.0748899	-14.4712	20 | for
0.0708661	-14.4714	coreference | to
0.2	-14.4714	our | extension
0.0762332	-14.4715	language | it
0.0511247	-14.4716	and | average
0.0909091	-14.4716	not | very
0.103448	-14.4718	consider | =
0.163265	-14.4721	this | factors
0.375	-14.4722	in | box.
0.163265	-14.4725	= | estimated
0.0992366	-14.4726	translation | but
0.0354406	-14.4727	of | model.
0.0847458	-14.4727	was | between
0.307692	-14.4729	null | paths
0.0399516	-14.4729	the | of | associated
0.0450311	-14.4729	proceedings | pages | oregon,
0.166667	-14.4731	time | such
0.266667	-14.4733	arg-c | aux+v
0.333333	-14.4735	and | sentence).
0.104348	-14.4736	see | et
0.307692	-14.4737	0 | 2)
0.0643087	-14.4737	with | will
0.0792079	-14.4737	for | agreement
0.101562	-14.4737	nn | dt
0.163265	-14.4738	model | conditional
0.428571	-14.4738	answer | significantly
0.150943	-14.474	then | n
0.0685921	-14.474	on | evaluation
0.134328	-14.4743	word | induction
0.238095	-14.4743	workers | on
0.0264865	-14.4746	for each | of
0.0903226	-14.4747	our evaluation | the
0.0518672	-14.4747	a | error
0.0780488	-14.4747	the | plan
0.121951	-14.4748	bayesian | et
0.148148	-14.4748	m | +
0.0197664	-14.4751	the | probability | the
0.16	-14.4752	several | their
0.104348	-14.4752	on | monolingual
0.0140973	-14.4754	the | data | the
0.206897	-14.4754	one | need
0.206897	-14.4754	problem | documents
0.0435414	-14.4756	is, | the
0.179487	-14.4757	sense | noun
0.5	-14.4759	p(ti|e)
0.0769231	-14.4759	a | q
0.138462	-14.476	way | can
0.333333	-14.476	25.0 | all
0.375	-14.476	choice | single
0.0693431	-14.4761	opinion | in
0.132353	-14.4763	each | example
0.0992366	-14.4763	the | passive
0.134328	-14.4764	fact | an
0.153846	-14.4764	he | on
0.112245	-14.4766	into | then
0.0610465	-14.4767	that | scores
0.0429185	-14.4768	previous | to
0.0421769	-14.4768	null | =
0.333333	-14.4768	event | ptm
0.0634921	-14.4768	a | larger
0.089172	-14.4769	on | higher
0.206897	-14.477	use | toolkit
0.0725806	-14.4771	as | problem
0.120482	-14.4772	3. | as
0.0486486	-14.4772	as | they
0.5	-14.4772	understand | during
0.266667	-14.4774	sentences | share
0.11	-14.4774	university | at
0.285714	-14.4775	case | subject
0.153846	-14.4775	section | provides
0.175	-14.4777	• | subject
0.148148	-14.4778	within | each
0.156863	-14.4782	each | uses
0.0614035	-14.4782	that | annotation
0.428571	-14.4782	(kim | abstracts
0.193548	-14.4784	consequently, | for
0.0140229	-14.4785	with | in
0.166667	-14.479	statistics | number
0.00606111	-14.4791	and | and
0.263158	-14.4792	89.3 | +
0.0659722	-14.4792	no | on
0.16	-14.4793	and | hovy
0.0942029	-14.4793	on | text.
0.108108	-14.4793	introduced | for
0.0400501	-14.4793	however, | for
0.0480427	-14.4798	a | p
0.0694981	-14.4801	we | effect
0.0970149	-14.4801	summaries | to
0.109091	-14.4802	feature | same
0.428571	-14.4803	would | maturity
0.156863	-14.4806	report | data
0.101695	-14.4807	of | stanford
0.0510204	-14.4807	task | al.,
0.0758929	-14.4807	also | at
0.136364	-14.4808	into | system.
0.0948905	-14.4808	with | user
0.0636943	-14.4809	in | contain
0.0526316	-14.4813	can | used to
0.227273	-14.4814	reordering | input
0.0540541	-14.4815	as | 1
0.126582	-14.4816	of | mass
0.0277286	-14.4816	not | as
0.030303	-14.4817	of | new
0.285714	-14.4818	fold | on
0.0387324	-14.4818	(2) | the | of
0.142857	-14.4818	than | models.
0.142857	-14.4818	each | under
0.040201	-14.4819	work | that
0.333333	-14.4819	no | catvar
0.184211	-14.482	1 | em
0.0447531	-14.482	training | on
0.0348071	-14.4822	the | automatic
0.12987	-14.4823	such | parsing
0.166667	-14.4824	need | such
0.11	-14.4824	from | common
0.0164858	-14.4825	model | of
0.0278436	-14.4826	with a | of
0.238095	-14.4828	0.48 | table
0.107143	-14.4828	the | realized
0.307692	-14.4831	2011 | 2011),
0.138462	-14.4831	two | phrases
0.115789	-14.4831	time, | is
0.134328	-14.4832	use | evaluate
0.145161	-14.4833	only | dependency
0.0700389	-14.4834	supervised | a
0.140625	-14.4835	that | latter
0.428571	-14.4836	red | world
0.333333	-14.4836	jj|ee14|p | in
0.105263	-14.4837	set | best
0.102564	-14.4838	model | described
0.0833333	-14.4838	= | y
0.193548	-14.4838	result | when
0.0857143	-14.4838	most | an
0.148148	-14.4838	paper | some
0.0515464	-14.484	the | unique
0.0875	-14.4843	question | as
0.1875	-14.4846	instance, | would
0.0752212	-14.4848	as | type
0.0758929	-14.4848	corpus | an
0.25	-14.4849	right | but
0.06	-14.485	a | tokens
0.1875	-14.4851	| | cd
0.0352381	-14.4851	has | that
0.0824176	-14.4852	using | both
0.0253102	-14.4852	of | feature
0.114583	-14.4853	model | proposed
0.333333	-14.4853	base | candidate
0.333333	-14.4853	dependency | lexical | syntactic
0.156863	-14.4854	as | units
0.120482	-14.4855	feature | model.
0.0780488	-14.4856	of | twitter
0.179487	-14.4857	sentiment | (pang
0.136364	-14.4857	improvement | our
0.132353	-14.4858	our | metric
0.117021	-14.4858	tree | sentence
0.0645161	-14.4858	we | some | in
0.0641026	-14.486	the | section,
0.117021	-14.4861	from | language.
0.5	-14.4861	fiction | spanish
0.266667	-14.4862	on | average,
0.266667	-14.4862	moreover, | some
0.134328	-14.4863	systems | between
0.102564	-14.4864	two | test
0.089172	-14.4866	2 | .
0.0465116	-14.4867	of | frequency
0.11828	-14.4868	it | us
0.193548	-14.4868	models | additional
0.0394737	-14.4869	a | e
0.0669014	-14.4869	4. | in
0.0632911	-14.487	for | sets
0.0626959	-14.4871	translations | in
0.0444104	-14.4872	of | sentiment
0.115789	-14.4872	communication | to
0.148148	-14.4872	7 | model
0.0474517	-14.4872	the | character
0.128205	-14.4873	work, | use
0.266667	-14.4876	= | ,imp
0.136364	-14.4877	by | need
0.307692	-14.4878	matrix | u
0.132353	-14.4878	acknowledgements | research
0.0518672	-14.4879	for | i
0.0474517	-14.488	of | top
0.08	-14.4884	being | is
0.0717131	-14.4886	(1) | that
0.046281	-14.4886	to | work
0.142857	-14.4887	shows | average
0.111111	-14.4887	2: | =
0.1875	-14.4889	weight | 1
0.16	-14.489	english | system
0.0444104	-14.4891	obtained | and
0.0429185	-14.4893	an | this
0.107143	-14.4893	structural | to
0.333333	-14.4895	query | knowledge.
0.0788177	-14.4895	the | foundation
0.0948905	-14.4895	however, | does
0.266667	-14.4896	front | as
0.214286	-14.4896	named | recognizer
0.0857143	-14.4899	some | it
0.0833333	-14.49	list | as
0.0649351	-14.4901	of | produce
0.0417227	-14.4902	topics | the
0.0824176	-14.4903	in | items
0.16	-14.4903	other | without
0.0657439	-14.4903	part-of-speech | and
0.375	-14.4903	note | explicit
0.333333	-14.4904	nevertheless, | number
0.0745614	-14.4904	this | proposed
0.153846	-14.4905	words | relevant
0.0213751	-14.4905	the | pairs
0.0649351	-14.4906	we | second
0.0745614	-14.4908	of | mentioned
0.0942029	-14.491	corpus | at
0.0857143	-14.491	words, | that
0.145455	-14.4914	vector | similarity
0.132353	-14.4915	can | query
0.0248685	-14.4916	the | the | while
0.166667	-14.4917	can | initial
0.214286	-14.4918	topic | content
0.0683453	-14.4919	a | segmentation
0.0857143	-14.4919	nlp | al.,
0.285714	-14.492	(e | )
0.0909091	-14.492	uses | with
0.089172	-14.4921	this | development
0.148148	-14.4922	where | documents
0.428571	-14.4924	german | ministry
0.227273	-14.4924	list | semantic
0.111111	-14.4924	in | produces
0.0927152	-14.4925	this | vector
0.5	-14.4925	cohen | pittsburgh,
0.217391	-14.493	task, | other
0.217391	-14.493	a) | with
0.142857	-14.4931	output | set
0.0472855	-14.4932	our | results
0.0497132	-14.4932	model | each
0.0842697	-14.4932	we | svm
0.0977444	-14.4933	this | shows
0.0772947	-14.4933	text | which
0.0819672	-14.4933	e | f | e
0.0703125	-14.4934	model | information
0.0769231	-14.4934	also | training
0.227273	-14.4936	additionally, | not
0.132353	-14.4939	while | feature
0.0431655	-14.4939	a | human
0.111111	-14.4944	is | score.
0.148148	-14.4945	n | ∈
0.0530973	-14.4946	value | to
0.166667	-14.4947	we | 2000
0.0503018	-14.4948	this | results
0.0869565	-14.4948	conditional | a
0.266667	-14.4951	lexical | capture
0.0571429	-14.4951	the | divided
0.0199938	-14.4953	our | for
0.145455	-14.4955	however, | lexical
0.114583	-14.4956	by | available
0.0942029	-14.4956	in | national
0.0252976	-14.4957	only | to
0.0341644	-14.4958	often | the
0.0673759	-14.496	us to | a
0.1875	-14.4961	should | because
0.0520607	-14.4962	distance | and
0.107843	-14.4964	further | from
0.0511247	-14.4966	multiple | to
0.0322321	-14.4966	a | null
0.140625	-14.4968	form | set
0.10084	-14.4968	is | documents.
0.0596591	-14.497	for | t
0.0290237	-14.497	in | semantic
0.126582	-14.4972	acknowledgments | would like
0.0955882	-14.4972	approach | features
0.193548	-14.4972	random | words
0.136364	-14.4972	finally, | word
0.132353	-14.4973	consider | between
0.1	-14.4975	to | one.
0.108911	-14.4977	a | other.
0.193548	-14.4977	work | systems.
0.162791	-14.4979	result | two
0.428571	-14.4979	predicates | structure
0.0266223	-14.4979	it | as
0.0458265	-14.4982	vector | and
0.0334802	-14.4982	) | the
0.0788177	-14.4982	semantic | word
0.166667	-14.4986	described | feature
0.0659722	-14.4986	also | it
0.119048	-14.4986	which | content
0.166667	-14.499	examples | shown
0.0219697	-14.4992	for | or
0.0784314	-14.4992	we | objective
0.134328	-14.4994	model | parameter
0.0928571	-14.4994	look | to
0.0565553	-14.4994	primary | the
0.0726496	-14.4995	evaluated | a
0.166667	-14.4999	most | topic
0.0392638	-14.5	with | model
0.145455	-14.5	with | get
0.134328	-14.5001	at | times
0.428571	-14.5001	5.2 | metrics
0.193548	-14.5002	2. | 2:
0.0278114	-14.5004	the | extracted
0.119048	-14.5004	from | higher
0.0708661	-14.5005	construct | of
0.0788177	-14.5006	english, | the
0.217391	-14.5008	it | natural to
0.0636943	-14.5008	random | a
0.104348	-14.5011	1 | 20
0.0378198	-14.5013	assume | the
0.0815217	-14.5014	and | experiments.
0.0323115	-14.5015	by | which
0.333333	-14.5016	minimum | among
0.148148	-14.5017	the | dropping
0.0977444	-14.5018	but | between
0.307692	-14.502	evaluation | followed
0.0610465	-14.502	in | selected
0.0271519	-14.5021	in | was
0.193548	-14.5022	proposed | 2009).
0.428571	-14.5023	ki | ede
0.428571	-14.5023	d102 | d90
0.428571	-14.5023	passvp | actvp
0.428571	-14.5023	r., | j.m.,
0.428571	-14.5023	maxscore | bestparse
0.428571	-14.5023	pagent | ppurpose−1
0.428571	-14.5023	yesnoq; | explain;
0.428571	-14.5023	je | faire
0.428571	-14.5023	0.47 | 0.38
0.428571	-14.5023	� | ��
0.428571	-14.5023	es, | es,
0.428571	-14.5023	j∈i0 | uqi
0.428571	-14.5023	difficult”), | (“left:
0.428571	-14.5023	(liang | 2006)
0.428571	-14.5023	nmod | pmod
0.428571	-14.5023	plos | 2008–2010
0.428571	-14.5023	da | λadj
0.428571	-14.5023	coated | rubber
0.428571	-14.5023	iy, | iy
0.428571	-14.5023	wk-4f | wk-4f
0.428571	-14.5023	events | pab
0.428571	-14.5023	bracehtipdownrightbracehtipdownleft | repair
0.428571	-14.5023	̸= | tr+1
0.428571	-14.5023	gradient | logpij
0.428571	-14.5023	[wp] | [clir]
0.428571	-14.5023	nist | mt05
0.428571	-14.5023	u´ | dh1
0.0955882	-14.5023	the | situation
0.166667	-14.5025	how | will
0.0571429	-14.5026	directly | of
0.307692	-14.5028	small | points
0.166667	-14.5029	score | according
0.128571	-14.5033	one | well
0.101695	-14.5034	include | are
0.0461538	-14.5038	on | at
0.0864198	-14.5039	sentence | have
0.130435	-14.5041	model | extraction
0.108911	-14.5041	by | representation
0.0374449	-14.5042	methods | the | of
0.042963	-14.5042	when | as
0.0608696	-14.5043	must | and
0.285714	-14.5046	15 | 18
0.206897	-14.5046	score | document
0.16	-14.5047	answer | by
0.0443038	-14.505	in order | we
0.179487	-14.5051	consists | three
0.150943	-14.5052	while | may | in
0.102564	-14.5052	of | listed
0.2	-14.5052	treebank | into
0.0381503	-14.5053	we | our | for
0.5	-14.5054	now | detail.
0.0684411	-14.5055	second, | to
0.0875	-14.5055	since | each
0.126582	-14.5057	starts with | the
0.0611621	-14.5057	on | machine
0.166667	-14.5059	construction | from
0.333333	-14.5059	method | faster
0.333333	-14.5059	evaluation | somewhat
0.307692	-14.506	relatum | or
0.103448	-14.506	new | each
0.0453074	-14.506	or | from
0.0571429	-14.506	related | al.,
0.0541176	-14.5062	on | performance
0.0644068	-14.5062	lexicon | the | of
0.0824176	-14.5064	to | scores.
0.0211268	-14.5064	our | to | of
0.0948905	-14.5064	s | m
0.11	-14.5065	information | only
0.5	-14.5067	pr(w′|ei)
0.266667	-14.5068	through | them
0.108911	-14.507	pairwise | a
0.0439394	-14.5074	this | features
0.307692	-14.5076	monotone | be
0.0285529	-14.5077	as | not
0.166667	-14.5081	hmm | et
0.142857	-14.5082	how | performance
0.0955882	-14.5083	some | at
0.0634921	-14.5084	results | section
0.333333	-14.5085	p(f|e) | =
0.119048	-14.5085	to | closely
0.145455	-14.5087	left | with
0.0955882	-14.5088	with | documents
0.0427729	-14.5089	is | task
0.179487	-14.5089	bias | is
0.0399485	-14.5089	by | al.
0.0726496	-14.509	tools | the
0.0474453	-14.5091	and | without
0.0617284	-14.5091	given | of | we
0.162791	-14.5093	arithmetic | of
0.0991736	-14.5094	relative | as
0.333333	-14.5094	0.60 | 0.64
0.0621118	-14.5094	in | instead
0.0745614	-14.5095	for | considered
0.020979	-14.5097	with | that
0.134328	-14.5098	by | low
0.0596591	-14.5098	topic | for
0.0628931	-14.5098	49th | computational | oregon,
0.0792079	-14.5099	the | is | task.
0.0569948	-14.5101	were | or
0.0819672	-14.5101	achieved | in
0.0439394	-14.5102	from | training
0.166667	-14.5103	with | class.
0.0295736	-14.5105	some | to
0.206897	-14.5105	part | national | under
0.0318888	-14.5107	the | in | table
0.130435	-14.5109	then | given
0.101695	-14.511	training | two
0.145455	-14.511	similar | source
0.150943	-14.5111	term | not
0.123457	-14.5111	experiments | approach
0.101695	-14.5113	no | our
0.1875	-14.5117	class | may
0.0824176	-14.5118	to | annotators
0.150943	-14.5119	those | authors | or
0.0367965	-14.5119	other | for
0.5	-14.5119	æ | fi¸@
0.113402	-14.5121	linguistics | school
0.103448	-14.5122	nodes | by
0.0828729	-14.5123	from | 2
0.0377574	-14.5125	accuracy | and
0.117647	-14.5126	baseline | bleu
0.0676692	-14.5127	by | dependency
0.175	-14.5129	make | these
0.150943	-14.5131	techniques, | of
0.0935252	-14.5131	in | detail
0.25	-14.5133	shows | four
0.0772947	-14.5136	of | database
0.0490196	-14.5137	a | case
0.175	-14.5138	result | system
0.166667	-14.5142	generate | all
0.0875	-14.5144	we | allow
0.119048	-14.5144	this | matching
0.117647	-14.5145	recommendations | and
0.0657439	-14.5145	to | translated
0.0314713	-14.5147	of the | 2011.
0.0692308	-14.5148	when | than
0.0214156	-14.5148	or | a
0.0164635	-14.5149	of | words
0.130435	-14.515	statistically | at
0.0726496	-14.5152	we | metric
0.101695	-14.5153	different | they
0.193548	-14.5153	logical | which
0.0697674	-14.5155	al., | 2005;
0.142857	-14.5157	matching | by
0.136364	-14.516	used | 2008;
0.217391	-14.5161	features | 2,
0.159091	-14.5162	we | overlapping
0.0780488	-14.5163	to | polarity
0.0711297	-14.5164	extraction | on
0.266667	-14.5166	higher | high
0.238095	-14.5166	similar | 2008;
0.119048	-14.5166	the | matrix,
0.0225898	-14.5167	that | or
0.112245	-14.5168	performs | as
0.0299572	-14.5168	is | = | =
0.0802139	-14.5169	compared | on
0.0848485	-14.5169	for | probabilities
0.120482	-14.517	differs | that
0.0273484	-14.5173	sentences | and
0.145455	-14.5174	usually | on
0.105263	-14.5178	of | validation
0.307692	-14.5181	integration | information
0.307692	-14.5181	vector | basis
0.166667	-14.5182	on | baseline.
0.121951	-14.5182	from | total
0.0473588	-14.5183	four | of
0.170732	-14.5185	compared | both
0.0487329	-14.5185	the | manual
0.5	-14.5185	pr(t|d)
0.105769	-14.5186	replaced | a
0.0880503	-14.5187	for | decoding
0.0664336	-14.5189	rules | as
0.307692	-14.5189	0.85 | 0.86
0.020202	-14.5189	we | model
0.0219569	-14.5189	sentence | the | of
0.106796	-14.5189	or | context
0.217391	-14.519	method | lexicon
0.375	-14.5195	lateen | —
0.0489237	-14.5197	a | set.
0.07173	-14.5198	related | that
0.5	-14.5198	decoder, | search
0.5	-14.5198	pa(αprime|γprime)
0.148148	-14.5199	might | such
0.1875	-14.5201	rule | 1.
0.00740547	-14.5202	in | in
0.153846	-14.5205	modules | be
0.375	-14.5205	update | α
0.375	-14.5205	report | news
0.0644068	-14.5206	known | of
0.179487	-14.5207	the | (it
0.0510638	-14.5208	differences | and
0.238095	-14.5216	more | spoken
0.162791	-14.5217	section | shows
0.113402	-14.5217	with | top
0.150943	-14.5219	de | et
0.0810811	-14.5219	a | makes
0.206897	-14.5219	vector | x
0.227273	-14.522	examples | corpus.
0.170732	-14.5221	such | corpora
0.116279	-14.5222	one | time
0.0962963	-14.5223	are | languages.
0.0238861	-14.5225	features | to
0.103448	-14.5225	l | r | and
0.5	-14.5225	sebastian | amherst
0.0735931	-14.5226	by | algorithm
0.0589888	-14.5227	in | instances
0.073913	-14.523	model | same
0.175	-14.523	models | model,
0.076555	-14.523	how | on
0.148148	-14.5231	information | text.
0.132353	-14.5232	be | limited
0.0784314	-14.5232	finally, | as
0.1	-14.5233	analysis | or
0.103448	-14.5234	for | simply
0.162791	-14.5234	one | had
0.0708333	-14.5237	x | =
0.0341556	-14.5237	association | and
0.102564	-14.5238	tree | all
0.227273	-14.5238	representations | model
0.103448	-14.5239	using | single
0.0708333	-14.5239	in | linear
0.181818	-14.5239	can | downloaded
0.1875	-14.524	validation | training
0.0571429	-14.524	step | in
0.0556962	-14.5241	the | paraphrases
0.25	-14.5245	however, | sets
0.307692	-14.5246	level | including
0.089172	-14.5247	other | using
0.156863	-14.5249	with | expression
0.0942029	-14.5249	to | french
0.0623053	-14.5251	that | languages
0.142857	-14.5252	such | knowledge
0.0819672	-14.5252	next | is
0.021079	-14.5255	the | may
0.10084	-14.5256	entities | be
0.193548	-14.5256	have | tools
0.0463918	-14.5257	such | with
0.0928571	-14.5259	been proposed | and
0.0853659	-14.5259	to | ambiguous
0.333333	-14.526	with | 25%
0.0346821	-14.526	or | that
0.238095	-14.5261	into | relative
0.0928571	-14.5262	first | use
0.113402	-14.5262	word | length
0.0991736	-14.5263	we | active
0.0824176	-14.5266	to | adding
0.132353	-14.5267	set | sentences.
0.0896552	-14.5268	of | lexicon.
0.0368764	-14.5269	our | using
0.0516129	-14.527	to | higher
0.0615385	-14.527	note | not
0.0702479	-14.527	made | a
0.266667	-14.5272	world | (e.g.,
0.0196752	-14.5275	word | in
0.0182306	-14.5276	one | of
0.162791	-14.5277	not | access
0.145455	-14.5278	each | provided
0.153846	-14.5278	sentences | chinese
0.142857	-14.5278	extraction | used
0.181818	-14.5278	= | (3) | are
0.0655172	-14.5281	pairs | that
0.0407056	-14.5282	still | the
0.0419682	-14.5282	average | and
0.285714	-14.5286	advantage | data.
0.0276074	-14.5286	is | used
0.238095	-14.5286	we use | alignments
0.0515021	-14.5288	specifically, | the | of
0.170732	-14.5289	significant | more
0.142857	-14.529	found | these
0.166667	-14.5292	use | local
0.0154374	-14.5292	is | on
0.132353	-14.5292	with | translation.
0.25	-14.5292	framework | reordering
0.0332136	-14.5293	at | for
0.0935252	-14.5294	classification | from
0.106796	-14.5294	many | word
0.0520607	-14.5295	null | nullnull
0.0343511	-14.5296	it | from
0.059322	-14.5298	a | identify
0.0637584	-14.5299	this | systems
0.0558376	-14.5299	can | between
0.0356415	-14.5299	for each | to
0.114943	-14.53	event | extraction
0.166667	-14.5301	alignment | over
0.121951	-14.5301	methods | the | methods
0.307692	-14.5303	data | little
0.132353	-14.5303	this | logical
0.25	-14.5305	user: | slot2,
0.375	-14.5305	corpus | la
0.116279	-14.5311	this | distributional
0.102564	-14.5314	research | no.
0.0363636	-14.5314	relation | of
0.121951	-14.5314	may | translation
0.116279	-14.5315	practice, | is
0.0338346	-14.5316	of | annotation
0.0789474	-14.5316	number | different
0.142857	-14.5316	at | within
0.0935252	-14.5316	on | comparable
0.0858896	-14.5319	25 | of
0.153846	-14.5319	proposed | similarity
0.0935252	-14.5319	obtained | on
0.123457	-14.532	from | 2006).
0.0338983	-14.532	cases | the
0.0702479	-14.532	two | at
0.0494071	-14.5321	and | proposed
0.00974421	-14.5321	as | and
0.0641892	-14.5322	however, | not | to
0.097561	-14.5323	shown | using
0.193548	-14.5323	can | seen,
0.0991736	-14.5324	ratio | a
0.0272562	-14.5325	this | that | in
0.162791	-14.5325	call | as
0.0176678	-14.5326	the | will
0.123457	-14.5327	syntactic | language
0.208333	-14.5327	figure | provides
0.130435	-14.5329	assigning | in
0.0935252	-14.5329	partial | is
0.206897	-14.5329	4 | topic
0.0714286	-14.5332	and | translation.
0.1	-14.5332	by | development
0.0282958	-14.5332	to | where
0.145455	-14.5333	pp | for
0.193548	-14.5333	b3 | f1
0.132353	-14.5334	an | scheme
0.0471869	-14.5334	training | that
0.181818	-14.5336	section | reports
0.114943	-14.5337	our | full
0.120482	-14.5339	applicable | of
0.285714	-14.534	(e.g., | “play
0.333333	-14.534	150 | standard
0.0733945	-14.5341	between | target
0.125	-14.5341	knowledge | our
0.0379597	-14.5343	of | 1.
0.5	-14.5344	include | predicate.
0.206897	-14.5346	important | into
0.113636	-14.5348	a | improving
0.0480769	-14.535	to | less
0.238095	-14.535	study | out
0.25	-14.5352	(1) | values
0.0793651	-14.5358	to | cost
0.333333	-14.5358	value | overall
0.5	-14.5358	r1 | t5,
0.227273	-14.5359	here, | between
0.145455	-14.5364	table | provides
0.0659341	-14.5364	because | on
0.140351	-14.5364	x | p
0.193548	-14.5364	section | final
0.032618	-14.5368	the | often
0.0589888	-14.5368	reordering | a
0.119048	-14.5368	and | dirichlet
0.2	-14.5369	senses | than
0.0819672	-14.537	technology | of
0.0915493	-14.5374	show | more
0.206897	-14.5374	(darpa) | contract
0.0386941	-14.5375	approach | for
0.0526316	-14.5376	obtained | in
0.104762	-14.5377	an | distribution
0.0848485	-14.5377	e | .
0.0575342	-14.5379	graph | in
0.0810811	-14.5379	naive | the
0.0853659	-14.538	m | s
0.0890411	-14.5383	however, | was
0.0467626	-14.5383	our | features
0.0222492	-14.5383	of | were
0.128571	-14.5387	are | express
0.145455	-14.5387	models | topics
0.078125	-14.5388	each | possible
0.137931	-14.5389	from | seed
0.0733945	-14.539	and | monolingual
0.1	-14.5391	features, | that
0.0869565	-14.5392	is | allows
0.0609756	-14.5398	a | supervised
0.0421512	-14.5398	as | training
0.0397946	-14.5398	annotation | to
0.150943	-14.54	greater | as
0.0915493	-14.5401	comes | of
0.116279	-14.5408	as | detection
0.0442623	-14.5408	and | possible
0.076555	-14.541	features | between
0.0384154	-14.5412	these | on
0.145455	-14.5415	with | transition
0.285714	-14.5416	they | nor
0.170732	-14.5417	selection | based
0.11	-14.5417	this | hypothesis
0.0398458	-14.5417	possible | to
0.0316558	-14.5418	related | and
0.112245	-14.5419	to | categories.
0.0983607	-14.5419	these | corpus
0.0518018	-14.542	the | false
0.112245	-14.5422	responses | in
0.105769	-14.5425	identifying | for
0.125	-14.5425	system | without
0.0646259	-14.5427	length | is
0.208333	-14.5428	sentence, | target
0.0435459	-14.5429	of | s
0.0806452	-14.543	in | report
0.0671642	-14.543	initial | a
0.0797872	-14.5431	d | r
0.25	-14.5432	0.1 | 10
0.238095	-14.5433	• | features.
0.114943	-14.5433	all | tree
0.114943	-14.5433	from | words,
0.103774	-14.5434	8 | 1
0.1875	-14.5436	utterance | an
0.042618	-14.5437	highest | the
0.0730594	-14.5438	while | model
0.217391	-14.544	where | patterns
0.0607903	-14.544	by | target
0.266667	-14.5443	gave | than
0.0666667	-14.5447	have | two
0.0541872	-14.5447	other | by
0.0890411	-14.5448	models | two
0.170732	-14.5449	different | words.
0.125	-14.5452	annotation | different
0.076555	-14.5453	points | to
0.2	-14.5455	general | these
0.0696721	-14.5456	by | rules
0.1875	-14.5456	sentence | position
0.375	-14.5456	ρkj(1) | k=e+1
0.0378698	-14.5457	probability | is
0.0815217	-14.5458	language | other
0.0928571	-14.5458	are | what
0.0853659	-14.5458	evidence | to
0.0656934	-14.5459	we | trees
0.307692	-14.546	|onull) | p(crt
0.307692	-14.546	turian | al.
0.10084	-14.5461	to | visual
0.142857	-14.5462	models | many
0.130435	-14.5463	structure | figure
0.238095	-14.5465	labels | (e.g.,
0.155556	-14.5467	on | algorithm.
0.0423358	-14.5468	their | for
0.155556	-14.5471	with | documents.
0.116279	-14.5471	american | the
0.137931	-14.5471	from | pattern
0.0306748	-14.5475	both | in
0.0512249	-14.5476	most | with
0.114943	-14.5479	by | university
0.113636	-14.548	by | provided
0.25	-14.548	most | although
0.145455	-14.5481	want | or
0.0818713	-14.5484	separated | separated
0.125	-14.5486	languages | word
0.060423	-14.5489	the | assign
0.0921986	-14.5489	by | correct
0.123288	-14.5489	specific | such
0.0828402	-14.5489	data | same
0.1875	-14.5492	supported | funded
0.155556	-14.5493	rules | any
0.0227273	-14.5493	the | similarity
0.126761	-14.5494	see | can
0.266667	-14.5494	means | very
0.227273	-14.5494	: | l
0.0548628	-14.5497	in | automatic
0.117647	-14.55	other | both
0.162791	-14.55	2 | m
0.0382318	-14.5502	the | right
0.0374707	-14.5502	well | and
0.0676692	-14.5503	an | source
0.176471	-14.5504	bad | as
0.029985	-14.5505	a | a | word
0.0357143	-14.5505	the | language.
0.0890411	-14.5506	our | classification
0.116279	-14.5506	describe | in section
0.0330357	-14.5507	shows | and
0.181818	-14.5508	its | rules
0.134328	-14.551	words | left
0.0639731	-14.5511	and | need
0.0645161	-14.5512	word | used
0.193548	-14.5516	= | j.
0.208333	-14.5518	clusters | than
0.0747664	-14.552	the | allowed
0.0864198	-14.5522	is | latent
0.0246603	-14.5524	as | from
0.0708333	-14.5524	in | vectors
0.0622951	-14.5525	we | work.
0.0727273	-14.5525	the | implicit
0.105769	-14.5526	furthermore, | be
0.111111	-14.5526	standard | word
0.123288	-14.5526	primarily | and
0.0633333	-14.5526	our | its
0.11236	-14.5528	product | is
0.170732	-14.5528	able | both
0.176471	-14.5528	though | an
0.113636	-14.5532	or | either
0.101695	-14.5534	we | unique
0.0858896	-14.5534	features | target
0.0561497	-14.5534	a subset | the
0.5	-14.5534	e | ki
0.150943	-14.5535	confidence | are
0.0378251	-14.5536	features | with
0.0532688	-14.5536	original | to
0.375	-14.5537	given | section.
0.0727273	-14.5538	seen | in
0.0944882	-14.5539	is | child
0.0460177	-14.5539	languages | and
0.285714	-14.554	than | faster
0.121951	-14.5541	much | it
0.0612903	-14.5542	the | in | top
0.162791	-14.5545	direct | from
0.0813953	-14.5546	for | measures
0.5	-14.5547	s53s48s48s48 | s78
0.5	-14.5547	build | onto
0.5	-14.5547	vouk, | j.c.
0.0915493	-14.5548	equal | and
0.170732	-14.5551	will | even
0.0381773	-14.5552	one | with
0.11236	-14.5553	that | scores.
0.11236	-14.5553	w | i
0.0909091	-14.5553	also | sentence
0.0853659	-14.5553	explicit | and
0.114943	-14.5554	tagger | on
0.0432767	-14.5554	given | of | is
0.166667	-14.5557	i | (1)
0.0215633	-14.5559	of | as | as
0.0536585	-14.5559	acknowledgments | and
0.155556	-14.556	wordnet | not
0.0890411	-14.5561	evaluation, | and
0.0952381	-14.5561	49th | for | linguistics:shortpapers,
0.0570652	-14.5563	a | head
0.170732	-14.5565	tokens | one
0.0944882	-14.5567	train | as
0.159091	-14.5568	description | our
0.117647	-14.5571	show | each
0.0482625	-14.5571	on | semantic
0.0666667	-14.5572	as | noun
0.0397351	-14.5577	of | sense
0.193548	-14.5579	predicate | word
0.0385093	-14.558	learning | in
0.134328	-14.558	0.2 | 0.5
0.107843	-14.558	to | generative
0.0909091	-14.558	have | when
0.121622	-14.5582	to | case.
0.0944882	-14.5583	each | t
0.114943	-14.5583	corpus | target
0.162791	-14.5585	divide the | into
0.10084	-14.5585	are | 2008).
0.121622	-14.5585	metric | human
0.096	-14.5586	l | d
0.285714	-14.5587	use | report
0.117647	-14.5587	supervised | data
0.266667	-14.5588	use | preposition
0.227273	-14.5588	although | those
0.375	-14.5589	trust | in
0.159091	-14.559	underlying | as
0.148148	-14.559	lists | from
0.2	-14.5597	based | labels
0.181818	-14.5597	all | y
0.170732	-14.5598	label | document
0.0896552	-14.5598	by | words.
0.0833333	-14.5598	high | are
0.0486815	-14.5599	is | verb
0.0656934	-14.5601	are | types
0.0449827	-14.5602	previous | a
0.142857	-14.5602	ilp | and
0.111111	-14.5602	are | although
0.152174	-14.5602	topics | over
0.0561497	-14.5602	looking | the
0.5	-14.5602	update | produce
0.0342466	-14.5606	to | it | a
0.0458554	-14.5607	of | 10
0.0878378	-14.5608	in | score.
0.0730594	-14.5609	using | test
0.0838323	-14.5609	by | national
0.0211111	-14.5611	is | that | in
0.0199071	-14.5611	is | that | to
0.140351	-14.5612	course | a
0.0674603	-14.5613	= | p
0.25	-14.5616	topic | estimated
0.0608974	-14.5618	models, | of
0.227273	-14.5619	because | values
0.152174	-14.5619	outperforms | system
0.0324846	-14.562	the | full
0.266667	-14.5625	combines | two
0.025	-14.5625	in | other
0.101852	-14.5627	results | proposed
0.0506608	-14.5629	have | this
0.227273	-14.5631	r | sou
0.0693878	-14.5631	following | by
0.135593	-14.5634	evaluation | precision
0.134328	-14.5635	word | argument
0.170732	-14.5636	to | processed
0.0705394	-14.5636	any | not
0.0866667	-14.5638	that | represent
0.0544554	-14.5641	– | for
0.042879	-14.5641	f | the
0.0205601	-14.5642	1 | and
0.0549738	-14.5644	evaluate | in
0.181818	-14.5647	value | similarity
0.106796	-14.565	section | discuss
0.132353	-14.5651	counts | with
0.130435	-14.5655	or | relations
0.0872483	-14.5655	thus, | be
0.0549738	-14.5655	for | 2010).
0.0247168	-14.5657	to | information
0.5	-14.5657	solved | local
0.0441176	-14.5658	requires | the
0.159091	-14.566	while | make
0.0531401	-14.5661	in | particular
0.0175983	-14.5664	of | using
0.11236	-14.5664	or | them
0.217391	-14.5665	texts | sentences
0.0733945	-14.5666	this | single
0.0269773	-14.5669	the | –
0.0967742	-14.567	score | model
0.0382716	-14.567	and | pairs
0.5	-14.5671	construction | constructed
0.5	-14.5671	able | stream
0.166667	-14.5672	space | between
0.25	-14.5672	whether | simply
0.120482	-14.5673	there | large
0.0744186	-14.5674	also | system
0.333333	-14.5676	sample | randomly
0.0645161	-14.5677	to | classes
0.097561	-14.5678	by | small
0.0633333	-14.568	the | s.
0.0115095	-14.5681	the | with | and
0.126761	-14.5682	which | rather
0.152174	-14.5682	than | baseline.
0.0537897	-14.5683	times | of
0.266667	-14.5684	to | efficient.
0.1	-14.5688	as | research
0.0445969	-14.569	is | similar
0.055409	-14.569	for | speech
0.266667	-14.5691	data, | could
0.0106083	-14.5693	of | and | in
0.142857	-14.5697	contains | sentence
0.285714	-14.5698	verbs | do
0.103774	-14.5698	score | set
0.132353	-14.5699	by | end
0.217391	-14.5701	when | languages.
0.166667	-14.5704	al., | 2002).
0.266667	-14.5706	expressions | phrases
0.1	-14.5708	each | function
0.166667	-14.5709	from | trigger
0.0813953	-14.571	with | relation
0.266667	-14.5713	0 | has
0.285714	-14.5713	(2009) | -
0.227273	-14.5713	having | at
0.333333	-14.5713	properties | verb
0.333333	-14.5713	response | second
0.0702479	-14.5714	for | meaning
0.0753769	-14.5716	with | sense
0.171429	-14.5718	contrast | previous
0.0923077	-14.5718	it | easy
0.227273	-14.5719	(i.e., | this
0.0601266	-14.572	can | their
0.0484848	-14.572	which | each
0.0208944	-14.5721	this | are
0.3	-14.5722	vector | decision
0.0330275	-14.5723	term | the
0.375	-14.5723	research | canada.
0.4	-14.5724	names | taxa
0.4	-14.5724	ibm1 | calculated
0.4	-14.5724	communicative | qualifiers
0.4	-14.5724	compare | indications
0.4	-14.5724	variable-length | finding
0.4	-14.5724	– | 39.6
0.4	-14.5724	def= | st
0.4	-14.5724	arg2 | complementary
0.4	-14.5724	relations | biological
0.4	-14.5724	gappy | gappy
0.4	-14.5724	lee | lee,
0.4	-14.5724	vehicles | &
0.4	-14.5724	star | 350
0.135593	-14.5724	similar | based
0.0633803	-14.5725	context | that
0.238095	-14.5726	also | subset
0.119048	-14.5726	or | in | those
0.3	-14.573	function | depends
0.3	-14.573	metric | previous
0.0237417	-14.573	and | two
0.123288	-14.573	appears | as
0.0122134	-14.5731	used | the
0.0332103	-14.5733	example | the | of
0.0353698	-14.5733	in | phrase
0.4	-14.5735	• | kjl
0.4	-14.5735	(for | gen
0.4	-14.5735	summaries | asymptotic
0.4	-14.5735	asr | noisy
0.3	-14.5738	td | slots
0.25	-14.5741	cjj | 0
0.114943	-14.5742	semantic | use
0.227273	-14.5745	therefore | will
0.121622	-14.5746	research | science
0.4	-14.5747	red | wrong
0.4	-14.5747	(ge | mooney,
0.4	-14.5747	lda | controled
0.4	-14.5747	educators | language
0.4	-14.5747	too | real-time
0.176471	-14.5747	(2010), | al.
0.170732	-14.5749	it | actually
0.0581395	-14.5749	of | concept
0.135593	-14.5751	was | second
0.022534	-14.5754	shown | the
0.140351	-14.5756	with | it.
0.052381	-14.5756	we | search
0.4	-14.5758	time | rm-a
0.4	-14.5758	debate | iphone
0.4	-14.5758	24 | hard
0.4	-14.5758	hard | settings
0.055409	-14.5758	a | find
0.333333	-14.5759	contain | appear
0.0465549	-14.5759	et | 2007; | and
0.108696	-14.5761	rule-based | of
0.3	-14.5763	global | interaction
0.145455	-14.5766	these | translations
0.0833333	-14.5766	the | vp
0.0720721	-14.5767	translation | not
0.0627063	-14.5767	the | results are
0.4	-14.5769	no | no.
0.4	-14.5769	j | x...y
0.4	-14.5769	dative | swiss
0.148148	-14.577	on | adaptation
0.114943	-14.5774	4.1 | by
0.2	-14.5774	question | when
0.0967742	-14.5775	the | sorted
0.108696	-14.5777	confusion | a
0.0551181	-14.5778	we | sentiment
0.0409027	-14.5779	as | two
0.155556	-14.578	best | between
0.0915493	-14.578	go | of
0.1	-14.578	considered | that
0.4	-14.578	“variation | string
0.4	-14.578	synsets | hyponym
0.0813953	-14.5783	that | p
0.0399449	-14.5784	possible | in
0.0276705	-14.5785	the | length
0.037037	-14.5785	agreement | the
0.130435	-14.5786	20 | from
0.238095	-14.5786	collaborative | et
0.140351	-14.5787	number | small
0.103774	-14.5788	even | they
0.4	-14.5791	schein | sampling
0.4	-14.5791	2010b; | 2010;
0.4	-14.5791	many-to-many | mapping
0.4	-14.5791	jab | 0.
0.4	-14.5791	get: | |=
0.0773196	-14.5791	by | about
0.11236	-14.5792	for | reason,
0.0765306	-14.5792	formal | and
0.105769	-14.5792	have | several
0.017016	-14.5792	a | word
0.238095	-14.5793	], | ],
0.238095	-14.5793	shiny; | <
0.238095	-14.5793	cwcxbncycx | cwcxbncycx
0.0563003	-14.5794	the | measured
0.227273	-14.5795	has | models.
0.148148	-14.5795	as | document.
0.152174	-14.5796	further | into | and
0.333333	-14.5797	johnson | an
0.0359477	-14.5798	in | lexical
0.171429	-14.5799	baseline | detection
0.0497835	-14.58	we | trained
0.0795455	-14.5801	sentences | using
0.4	-14.5802	removing | unary
0.4	-14.5802	pyramid | p-value
0.4	-14.5802	operators | remove
0.4	-14.5802	2007) | functional
0.4	-14.5802	u. | at
0.4	-14.5802	corrections | acquisition
0.106796	-14.5805	or | expressed
0.159091	-14.5806	a list | from
0.0944882	-14.5806	not | correct
0.176471	-14.5807	reported | at
0.107843	-14.581	the | conveyed
0.104762	-14.5811	shared | with
0.145455	-14.5811	corresponds | an
0.078534	-14.5811	do | any
0.0568182	-14.5813	in | patterns
0.12	-14.5814	often | than
0.4	-14.5814	deterministic | states
0.0232664	-14.5814	the | shown
0.0693878	-14.5816	documents | for
0.375	-14.5817	pipeline | event
0.0520095	-14.5817	this | an | for
0.078125	-14.5817	compared to | in
0.0436893	-14.5818	that | does
0.266667	-14.5818	fahrr¨ader | 500
0.0872483	-14.5818	such | also
0.0363224	-14.5818	as | between
0.135593	-14.5819	out | one
0.0223252	-14.5821	of | is | by
0.0397805	-14.5822	from | word
0.0119522	-14.5824	in | with
0.285714	-14.5825	is, | user
0.142857	-14.5825	a | movie
0.333333	-14.5825	saffran | al.
0.0761421	-14.5826	in | help
0.11236	-14.5827	based | pos
0.3	-14.5831	error | %)
0.0813953	-14.5831	o | r
0.137931	-14.5833	relation | using
0.0374396	-14.5833	that | its
0.135593	-14.5834	such | relation
0.069869	-14.5835	parallel | from
0.125	-14.5835	do | need
0.4	-14.5836	iterations | convergence,
0.148148	-14.5837	has | even
0.5	-14.5838	assignment | according
0.5	-14.5838	around | pairs.
0.375	-14.5839	mt | o
0.208333	-14.5841	entity | argument
0.152174	-14.5843	extracting | with
0.0669291	-14.5844	generative | and
0.104762	-14.5846	of | pronoun
0.4	-14.5847	(zelenko | 2005;
0.3	-14.5848	obtain | per
0.044905	-14.585	are | where
0.0171216	-14.5853	3 | the
0.193548	-14.5854	performed | better
0.4	-14.5859	non-understanding | under
0.25	-14.586	feature, | not
0.0450607	-14.586	the | →
0.104762	-14.5864	will | into
0.3	-14.5864	digital library | of
0.3	-14.5864	order | recognize
0.0916031	-14.5864	cases, | is
0.0983607	-14.5864	agreement | on
0.0464684	-14.5866	it | our
0.0855263	-14.5867	data | into
0.0131234	-14.587	first | the | the
0.4	-14.587	prefix | prefix.
0.375	-14.587	german | modern
0.3	-14.5873	shows | steps
0.0351438	-14.5873	using | we
0.0902778	-14.5874	training | a | training
0.0765306	-14.5874	data | other
0.044905	-14.5874	the | properties
0.0491803	-14.5875	an | at
0.176471	-14.5876	• | name
0.0669291	-14.5876	context | be
0.0378973	-14.5878	we | any
0.0671937	-14.5878	such | it
0.0292398	-14.5878	other | a
0.137931	-14.5879	system | 2011
0.0723982	-14.5881	information | can
0.4	-14.5881	nicht | werden
0.4	-14.5881	` | u
0.4	-14.5881	吼 | but
0.137931	-14.5883	using | train
0.0661479	-14.5886	the | bound
0.121622	-14.5886	a | d,
0.0306205	-14.5888	the | classifier | the
0.0526316	-14.5888	5 | on
0.0578035	-14.5891	we | query
0.152174	-14.5891	language | could
0.0427892	-14.5892	frequent | the
0.4	-14.5893	unlike | srl
0.0878378	-14.5894	(in | in
0.25	-14.5895	vsummationdisplay | usummationdisplay
0.5	-14.5895	alessandro | 38123
0.103774	-14.5896	dependency | its
0.0990991	-14.5898	say | is
0.0507726	-14.5902	of | opinion
0.0114891	-14.5903	the | using
0.0633803	-14.5903	proposed | as
0.0695652	-14.5904	to | forms
0.4	-14.5904	mapping | attempted
0.4	-14.5904	information | sensors
0.0701754	-14.5906	of | root
0.0459559	-14.5906	a | likely
0.266667	-14.5908	will | approach.
0.25	-14.5909	data | well.
0.0238908	-14.591	from | as
0.0406977	-14.591	of | associated
0.171429	-14.591	two | compare
0.0916031	-14.5914	we | generative
0.193548	-14.5914	analysis | linguistic
0.208333	-14.5918	found | higher
0.0967742	-14.592	and | translations.
0.0376214	-14.5922	any | a
0.131148	-14.5923	was | 2010).
0.25	-14.5924	multinomial | topic
0.0498915	-14.5924	we | previous
0.0804598	-14.5926	n-gram | is
0.4	-14.5927	vbp | adjp
0.0419907	-14.5929	the | compute
0.0777202	-14.5931	performance | not
0.0923077	-14.5936	in | structure.
0.131148	-14.5937	on | nist
0.4	-14.5938	which | t5
0.4	-14.5938	mt | bleu,
0.25	-14.5938	these | 2001;
0.235294	-14.5938	its | threshold
0.0890411	-14.5939	by | values
0.0551181	-14.5942	extraction | al.,
0.0631579	-14.5942	as | evaluation
0.0872483	-14.5944	other | features
0.140351	-14.5946	by | technology
0.333333	-14.5947	translations | quality.
0.4	-14.5949	ein | aller,
0.152174	-14.5951	analysis, | 2011,
0.5	-14.5952	ranks | among
0.4	-14.5961	nullt | function.
0.4	-14.5961	yo | jwenn
0.4	-14.5961	  | 0
0.4	-14.5961	.71 | -
0.0549738	-14.5963	the | choose
0.0463822	-14.5965	terms | to
0.375	-14.5966	only | 50%
0.02657	-14.5968	it | not
0.137931	-14.5969	related | different
0.0952381	-14.5971	an | user
0.0493562	-14.5973	this | us
0.106383	-14.5973	a | errors.
0.0813953	-14.5973	in | predicate
0.0531401	-14.5974	to | vector
0.3	-14.5975	steve_young | michael_holt
0.0612903	-14.5976	phrase | from
0.162791	-14.5981	have | instances
0.107527	-14.5982	weights | as
0.105769	-14.5983	and | discussion
0.133333	-14.5983	(2007) | as
0.3	-14.5984	p(e), | entity
0.0548303	-14.5984	to | instances
0.126761	-14.5985	translation | output
0.155556	-14.5986	instances | test
0.118421	-14.5986	work, | our
0.0247368	-14.5988	document | the
0.217391	-14.5989	0.35 | 0.05
0.0136327	-14.5989	to | from
0.266667	-14.5991	crowdflower | to
0.235294	-14.5991	generate | corpus.
0.0727273	-14.5991	to | allow
0.0509259	-14.5992	parallel | a
0.0735294	-14.5993	for | comparable
0.171429	-14.5993	classification | three
0.0720721	-14.5993	morphological | to
0.148936	-14.5994	by | people
0.333333	-14.5995	image | vds-7
0.266667	-14.5999	with | official
0.105769	-14.6001	be | larger
0.208333	-14.6001	likely | about
0.133333	-14.6006	sentence | syntactic
0.171429	-14.6008	words | represented
0.0695652	-14.6009	to | and | features.
0.0833333	-14.6009	recognition | a
0.5	-14.6009	= | producttextx
0.116279	-14.6012	automatically | al.,
0.0769231	-14.6012	this | need
0.0432612	-14.6014	feature | as
0.333333	-14.6014	capture | probability
0.131148	-14.6016	consists | words
0.176471	-14.6017	phrases | this
0.0372596	-14.6019	c©2011 | computational
0.285714	-14.602	than | typical
0.0213439	-14.6023	to | all
0.0828402	-14.6023	test | data
0.235294	-14.6025	constraints | al.
0.0534351	-14.6027	the | labeling
0.1	-14.6029	is | cosine
0.4	-14.6029	r3 | r1−1
0.0448029	-14.603	and | dialogue
0.0895522	-14.6031	first, | with
0.0202775	-14.6031	this | to | a
0.145833	-14.6032	proposed | source
0.0761421	-14.6034	approach | based
0.227273	-14.6034	miwa | al.
0.125	-14.6034	since | into
0.10989	-14.6036	linear | are
0.0595611	-14.6038	generate | in
0.0424528	-14.6038	they | be
0.4	-14.6041	they found | features
0.4	-14.6041	combining | distributional
0.123288	-14.6043	of | reverse
0.133333	-14.6044	complexity | number
0.217391	-14.6045	rec | prec
0.0443686	-14.6047	the | differences
0.0588235	-14.6049	should | as
0.3	-14.6052	in | paper’s
0.4	-14.6052	yangzhong | yangtze
0.333333	-14.6052	each | qi
0.0448833	-14.6054	a | positive
0.111111	-14.6055	a | little
0.0396717	-14.6056	by | all
0.116883	-14.6057	index | to
0.0509259	-14.6057	is | approach
0.2	-14.6057	in | different.
0.118421	-14.6058	one | following
0.0711111	-14.6058	into | set
0.121622	-14.606	word | chinese
0.3	-14.6061	according | results,
0.0327715	-14.6062	experimental | the
0.181818	-14.6062	documents | query
0.0466926	-14.6063	3 | that
0.2	-14.6063	stanford | tagger
0.4	-14.6064	0.55 | 0.05
0.4	-14.6064	collaborators | performance.
0.155556	-14.6067	we | recall.
0.0982143	-14.6067	n | where
0.285714	-14.6069	apply | any
0.128571	-14.607	features | relation
0.0542636	-14.6071	we use | et
0.235294	-14.6072	problem | eq.
0.0643939	-14.6073	addition | a
0.0643939	-14.6073	describes | and
0.114943	-14.6074	δ | the
0.2	-14.6075	experiments | approach.
0.0396717	-14.6078	a | trained
0.227273	-14.6079	such | summationdisplay
0.235294	-14.6079	should | structure
0.0844156	-14.608	as | make
0.0394558	-14.6081	of | particular
0.111111	-14.6081	training | trained
0.0990991	-14.6082	presented | from
0.131148	-14.6084	to | definitions
0.0902256	-14.6085	on | frequency
0.3	-14.6087	de | type
0.0541237	-14.6089	as | may
0.0315513	-14.6089	the | respectively.
0.208333	-14.609	motivated | from
0.0474308	-14.6093	a | of | rules
0.217391	-14.6095	form | (e.g.,
0.375	-14.6095	debate | vs.
0.0527638	-14.6095	the | optimization
0.131148	-14.6095	represented | where
0.5	-14.6095	2009), | allows
0.171429	-14.6097	differs | it
0.227273	-14.6099	such | 2003),
0.181818	-14.6099	nodes | same
0.166667	-14.61	optimization | by
0.145833	-14.61	2 | ,x
0.0735294	-14.6101	several | on
0.105263	-14.6104	on | domain.
0.2	-14.6104	treebank | sentences
0.0761421	-14.6108	in | experiments, | of
0.129032	-14.6108	a | case.
0.0878378	-14.6115	paper | organized
0.0844156	-14.6116	test | not
0.0262515	-14.6116	from | are
0.375	-14.6117	structures | limited
0.101852	-14.6118	existing | al.,
0.140351	-14.6119	mention | or
0.166667	-14.612	set | senses
0.3	-14.6122	them | (e.g.
0.0671937	-14.6124	the | extraction.
0.0489362	-14.6124	data | our
0.0813953	-14.6125	is | object
0.100917	-14.6126	| | )
0.235294	-14.6127	θ | from
0.176471	-14.6128	only | features,
0.0486258	-14.613	we | sets
0.25	-14.6131	always | better
0.4	-14.6133	queries | qe
0.0246649	-14.6134	to | into
0.0683761	-14.6134	our | english
0.0731707	-14.6135	corpus | not
0.0373134	-14.6135	even | of
0.0973451	-14.6136	each | 1.
0.2	-14.6138	χ2 | =
0.0254525	-14.6138	is | our
0.192308	-14.6139	random | accuracy
0.142857	-14.6139	parsing | has
0.0809249	-14.6139	and | achieved
0.235294	-14.614	need | do
0.145833	-14.6142	translation | hierarchical
0.25	-14.6146	intuition | likely
0.0376884	-14.6146	based | for
0.11236	-14.6147	of | effectively
0.0769231	-14.6149	10 | 30
0.192308	-14.615	bag-of-words | for
0.0498915	-14.6155	is | vector
0.0539846	-14.6157	as | sentences
0.08	-14.6158	to | rank
0.0337769	-14.616	if we | the
0.375	-14.616	overview | approach.
0.123288	-14.6161	• | were
0.208333	-14.6163	n=1...4 | n-gram
0.3	-14.6165	sources | parallel
0.121622	-14.6165	alignment | all
0.285714	-14.6168	look | similar
0.162791	-14.6168	feature | 1.
0.0916031	-14.6169	features | source
0.100917	-14.617	a | translations.
0.0833333	-14.617	negative | are
0.171429	-14.6171	en | for
0.0646388	-14.6173	k | to
0.113636	-14.6175	since | such
0.101852	-14.6175	3 | translation
0.106383	-14.6176	with | might
0.4	-14.6179	topic: | i
0.2	-14.6179	the | “to
0.176471	-14.618	strings | from
0.0535714	-14.6182	in | statistical
0.0735294	-14.6184	any | by
0.162791	-14.6187	use | across
0.333333	-14.6187	not | modify
0.0849673	-14.6189	value | be
0.3	-14.6191	since | case,
0.4	-14.6191	other | characteristics.
0.0349727	-14.6192	our | that | in
0.208333	-14.6193	scenario, | is
0.135593	-14.6195	of | together.
0.118421	-14.6195	by | certain
0.0648855	-14.6195	although | be
0.123288	-14.6197	that | interpretation
0.0285103	-14.6198	for | one
0.0566572	-14.6198	we | across
0.285714	-14.6201	suo | fidanzato
0.285714	-14.6201	 | log
0.285714	-14.6201	γ(∆t) | ∆t≤σ
0.285714	-14.6201	,b | ⇒g
0.285714	-14.6201	x) | ①
0.285714	-14.6201	iters | 28.4
0.285714	-14.6201	∼ | —–
0.0534351	-14.6202	aim | the
0.0494624	-14.6203	for | algorithm
0.126761	-14.6208	cluster | each
0.0833333	-14.6209	3.2 | is
0.0646388	-14.621	is | of | verb
0.0822785	-14.621	results | based
0.4	-14.6214	kathleen | delaware,
0.4	-14.6214	types | substring
0.152174	-14.6214	input | its
0.1875	-14.6215	lexicons | al.,
0.111111	-14.6216	3.2 | as
0.145833	-14.6218	report | this
0.0505747	-14.6219	this | learning
0.159091	-14.622	2. | has
0.0504587	-14.6226	the | far
0.121622	-14.6226	n | on
0.0267686	-14.6227	our | by
0.0271318	-14.6228	the | is | at
0.208333	-14.6229	both | algorithms
0.235294	-14.6229	for | genres,
0.162791	-14.623	ε | ε
0.0813953	-14.623	be | no
0.0537084	-14.6232	that | least
0.3	-14.6235	provides | natural
0.176471	-14.6236	live | and
0.108696	-14.624	see | which
0.0382586	-14.6241	the | resources
0.0494624	-14.6243	instead | of
0.12	-14.6244	argument | on
0.0916031	-14.6244	were | sentences
0.131148	-14.6244	previous | focused
0.0675105	-14.6246	to | e.g.
0.0948276	-14.6247	use | model.
0.208333	-14.6248	syntactic | model.
0.0359212	-14.6248	that | were
0.0495495	-14.6251	our | both
0.176471	-14.6252	vp | vbd
0.0191248	-14.6254	to | word
0.235294	-14.6257	translated | texts.
0.0973451	-14.6258	4 | .
0.227273	-14.6258	nc | 1.00
0.0651341	-14.626	may | from
0.121622	-14.6261	approach, | as
0.333333	-14.6265	also | exponential
0.3	-14.627	total | requests
0.129032	-14.627	development | were
0.375	-14.627	0.54 | 0.53
0.0661479	-14.6271	to | i.e.
0.040724	-14.6272	approach | on
0.00771573	-14.6274	is | a
0.106383	-14.6275	that | weighted
0.0531646	-14.6278	sets | to
0.166667	-14.628	well | about
0.0211577	-14.6281	of | same
0.0714286	-14.6281	summary | to
0.159091	-14.6281	query | language
0.0909091	-14.6282	conclude | of
0.131148	-14.6282	supervised | learning
0.12	-14.6283	assume | have
0.05	-14.6284	we | original
0.0417311	-14.6285	to | time
0.126984	-14.6287	entity | its
0.0235872	-14.6288	can | by
0.0374065	-14.6288	using | al.,
0.0742574	-14.629	in | points
0.192308	-14.6291	j=1 | i=1
0.10989	-14.6291	that | (i.e.
0.0849673	-14.6291	table | comparison
0.208333	-14.6291	phrase-based | trained
0.0270806	-14.6292	to | if
0.0595611	-14.6293	which | two
0.113924	-14.6293	many | data
0.101852	-14.6294	(2005) | to
0.0677966	-14.6294	of | sum
0.0280702	-14.6295	the | even
0.101852	-14.6297	constituent | to
0.0964912	-14.6298	from | part
0.0782123	-14.6301	where | feature
0.0765027	-14.6303	corpus, | for
0.0641509	-14.6305	in | directly
0.25	-14.6307	set | returned
0.0990991	-14.6307	order | words
0.4	-14.6308	constraints | closure
0.4	-14.6308	kernels | holders
0.131148	-14.6309	authors | thank
0.142857	-14.631	as | boundaries
0.0165652	-14.6312	for | on
0.0875912	-14.6312	such | same
0.235294	-14.6312	building | corpora
0.171429	-14.6317	structure | other
0.0570571	-14.6318	and | significantly
0.25	-14.6322	x | position
0.0353478	-14.6322	the | web
0.0209321	-14.6323	first | and
0.105263	-14.6325	words | when
0.152174	-14.6326	we | vectors.
0.034632	-14.6327	and | lexical
0.118421	-14.6327	systems | more
0.208333	-14.6327	problem | model.
0.106383	-14.6328	from | state
0.0695652	-14.6331	we | among
0.12	-14.6332	an | points
0.171429	-14.6332	our | interaction
0.0916031	-14.6332	the | challenging
0.0769231	-14.6332	confidence | to
0.166667	-14.6334	sentences | aligned
0.0464217	-14.6334	in | annotated
0.1	-14.6335	analysis | data
0.036855	-14.6336	word | by
0.10989	-14.6337	enough | for
0.0391061	-14.6337	of | output
0.0728155	-14.6337	algorithm | which
0.0869565	-14.6341	first | system
0.5	-14.6343	dataskdelay | adjpair
0.5	-14.6343	reservoir | locality
0.5	-14.6343	entity’s | types
0.4	-14.6344	potential | focus
0.333333	-14.6344	uniform | random
0.0731707	-14.6349	of | involved
0.0543478	-14.6351	(i) | the
0.137931	-14.6354	version of | on
0.162162	-14.6355	different | vs.
0.0575758	-14.6357	approach | the | by
0.152174	-14.6357	same | type
0.375	-14.6358	[without | il-2
0.375	-14.6358	64 | slovinskuza
0.375	-14.6358	pp@2null1 | jˇux´ıng
0.375	-14.6358	(changeto | (e)
0.375	-14.6358	400 | age
0.375	-14.6358	a1a0a2 | a31a14a12a8
0.375	-14.6358	|ot)summationtext | |ot)
0.375	-14.6358	uppercase | nullnullnullnullnullnullnullnullnull(null)
0.375	-14.6358	450 | acquisition
0.375	-14.6358	0.0015 | 4e−04
0.375	-14.6358	h31, | {h12,
0.375	-14.6358	terme | très
0.375	-14.6358	’\(ncmod | _
0.375	-14.6358	0.0005 | 6e−04
0.375	-14.6358	0.0000 | 2e−04
0.375	-14.6358	d104 | d99
0.375	-14.6358	integraldisplay | parenleftbiggδt
0.375	-14.6358	where { | geo:city
0.375	-14.6358	network | layer
0.375	-14.6358	a101 | a110
0.375	-14.6358	00 | (11
0.375	-14.6358	0.011% | 0.018%
0.375	-14.6358	bow | 74.1
0.375	-14.6358	46.36 | 47.07
0.375	-14.6358	0.0010 | 0e+00
0.375	-14.6358	50.05 | 51.87
0.375	-14.6358	propor | /ada/
0.375	-14.6358	52.51 | 51.45
0.375	-14.6358	france, | comic,
0.192308	-14.6359	approximate | using
0.1875	-14.6359	𝟏 | 𝒒
0.0505747	-14.636	we | that | between
0.0552486	-14.636	in | length
0.0940171	-14.6361	as | 2006).
0.0226223	-14.6364	and | at
0.123288	-14.6364	candidate | set
0.162162	-14.6365	these | parameters
0.25	-14.6366	did | after
0.137931	-14.6366	there | how
0.4	-14.6367	no. | subcontract
0.0909091	-14.6371	a | until
0.171429	-14.6373	manual | our
0.181818	-14.6375	sentences | size
0.0612245	-14.6376	relation | as
0.041868	-14.6377	a | sense
0.137931	-14.6378	if | many
0.131148	-14.6379	discussion | results
0.0612245	-14.6383	in | entities
0.0731707	-14.6384	work | such
0.102041	-14.6384	should | this
0.0860927	-14.6384	based | used
0.0948276	-14.6387	difference | by
0.0973451	-14.6388	full | as
0.5	-14.6388	p(w|c)
0.5	-14.6388	follows: | something
0.0127168	-14.6389	data | the
0.148936	-14.6389	n | b
0.104167	-14.639	is | parts
0.108696	-14.6391	can | simple
0.0895522	-14.6392	based | translation
0.0194672	-14.6393	this | on
0.0554017	-14.6394	focus | to
0.040724	-14.6395	with | language
0.0358382	-14.6395	in | with | with
0.181818	-14.6397	nodes | dependency
0.0334686	-14.6397	from | or
0.145833	-14.6399	proceedings | emnlp
0.0348584	-14.64	the | 3.
0.208333	-14.6402	could | considered
0.266667	-14.6403	or | author(s)
0.0731707	-14.6403	indicate | to
0.0625	-14.6403	of | algorithm.
0.0164609	-14.6405	will | the
0.1125	-14.6405	approach | similar
0.105263	-14.6406	sentence | given
0.152174	-14.6407	such | evaluation
0.0895522	-14.6408	to | new | from
0.0359629	-14.6413	are | all
0.333333	-14.6413	trained | iterations.
0.104167	-14.6415	in | author
0.0191546	-14.6417	would | the
0.235294	-14.6417	function | as:
0.266667	-14.6419	extract | semantic
0.176471	-14.6419	used | tokens
0.0677966	-14.6419	have | information
0.0686695	-14.6419	summationtext | the
0.0397163	-14.6422	to | many
0.116883	-14.6426	all | significant
0.125	-14.6429	analysis | sentiment
0.0382586	-14.6431	this | one
0.00845883	-14.6432	a | a | of
0.5	-14.6432	26 | users
0.031617	-14.6435	then | in
0.0287879	-14.644	sense | the
0.106383	-14.6442	we | line
0.0686695	-14.6445	does | on
0.100917	-14.6447	for | object
0.062963	-14.6447	we | presented
0.5	-14.6447	(daelemans | den
0.3	-14.6448	r | 20.6
0.4	-14.6451	itself | even
0.208333	-14.6451	propose | similar
0.0773481	-14.6451	more | when
0.181818	-14.6451	4.1.2 | the
0.162162	-14.6452	α | by
0.121622	-14.6452	news | corpus
0.105263	-14.6453	then | sentence
0.102041	-14.6454	presents | results
0.0584615	-14.6454	was | it
0.25	-14.6455	pair | tokens
0.0555556	-14.6456	or | can
0.0428336	-14.6457	related | a
0.159091	-14.646	.36 | .38
0.0434028	-14.646	= | i
0.12	-14.6461	small | training
0.0797546	-14.6462	pairs | each
0.0844156	-14.6462	we | run
0.0534759	-14.6462	results | than
0.041868	-14.6462	the | allows
0.0622711	-14.6464	did | in
0.0483193	-14.6464	hierarchical | the
0.043554	-14.6467	training | from
0.0940171	-14.6467	of | average | per
0.0857143	-14.6467	is | work.
0.123288	-14.647	and | references
0.181818	-14.6473	parsing, | dependency
0.0410959	-14.6475	and | 2009;
0.0464646	-14.6476	all | each
0.0468432	-14.6476	range | the
0.0807453	-14.6479	corpus | it
0.166667	-14.6483	features | adding
0.155556	-14.6487	shows | learning
0.104167	-14.6489	acknowledgments | national
0.152174	-14.6489	compare | use
0.0518519	-14.6492	the | property
0.5	-14.6492	descriptions | short
0.235294	-14.6494	based | estimated
0.208333	-14.6495	alignment | standard
0.0669456	-14.6501	our | dependency
0.0388889	-14.6501	s | and
0.145833	-14.6503	whether | used
0.333333	-14.6503	computational | referring
0.152174	-14.6503	and | al,
0.135593	-14.6504	0.4 | 0.2
0.0964912	-14.6504	target | training
0.181818	-14.6506	evaluate | gold
0.0773481	-14.6507	a | requires
0.25	-14.6508	finally, | similar
0.222222	-14.651	has | structure.
0.192308	-14.6514	reference | evaluation
0.111111	-14.6516	an | discourse
0.111111	-14.6516	contexts | are
0.0316869	-14.6517	feature | is
0.0232443	-14.6517	have | for
0.0888889	-14.652	they | model
0.105263	-14.652	one | models
0.0782123	-14.652	would | it
0.118421	-14.6521	three | test
0.0257827	-14.6522	the | specific
0.0200584	-14.6522	results | in
0.106383	-14.6522	mixture | to
0.181818	-14.6522	more | defined
0.0602007	-14.6523	these | words
0.0614334	-14.6523	of | right
0.0321058	-14.6525	the | of | standard
0.166667	-14.6528	4 | results.
0.058104	-14.6528	and | future
0.0833333	-14.653	for | existing
0.0202399	-14.6531	the | best
0.0666667	-14.6533	name | is
0.0311388	-14.6533	the | multiple
0.208333	-14.6533	nell | to
0.0303286	-14.6534	number | number
0.0494382	-14.6534	semantic | are
0.0756757	-14.6537	output | that
0.129032	-14.6538	must | it
0.152174	-14.6539	since | data.
0.131148	-14.654	60 | 70
0.118421	-14.6542	that | attributes
0.0383037	-14.6545	decoding | the
0.14	-14.6546	results | rate
0.131148	-14.6548	system | error
0.12	-14.6551	section | paper.
0.0168307	-14.6553	this | the | for
0.116883	-14.6553	consisted | the
0.333333	-14.6553	study | investigate
0.0434028	-14.6554	what | of
0.106383	-14.6557	2 | were
0.0194918	-14.6559	also | and
0.185185	-14.6561	5 conclusion | paper,
0.115385	-14.6562	that | tree.
0.0443623	-14.6563	corresponding | of
0.222222	-14.6563	man | to
0.03125	-14.6563	null | of
0.0675105	-14.6565	and | translated
0.062963	-14.6566	meaning | is
0.0509709	-14.6568	we | 3
0.0634328	-14.6569	we | experiment
0.0773481	-14.6572	is | provided
0.092437	-14.6574	us | which
0.0634328	-14.6576	validation | the
0.25	-14.6576	improvement | system.
0.25	-14.6576	modified | (chen
0.0863309	-14.6577	work | focused
0.222222	-14.6577	definition | section
0.145833	-14.6578	θ | φ
0.0956522	-14.658	by | high
0.116883	-14.6581	our | makes
0.0429553	-14.6582	that | lexical
0.222222	-14.6583	parallel | wikipedia
0.171429	-14.6583	they | include
0.162162	-14.6588	detail, | and
0.166667	-14.6589	fast | a
0.0736842	-14.6589	that | are | different
0.0964912	-14.6589	that | 2008).
0.152174	-14.659	but | annotation
0.125	-14.659	shows | these
0.123077	-14.6591	classifier | it
0.3	-14.6593	min | c
0.0479303	-14.6595	tags | and
0.142857	-14.6596	our | line
0.142857	-14.6596	may | null
0.222222	-14.6597	where | treated
0.0863309	-14.6598	general | from
0.125	-14.6598	graph | where
0.0704225	-14.6598	a | group
0.5	-14.6598	zhang, | 2010.
0.0483516	-14.6601	for | similar
0.157895	-14.6602	example, | translated
0.217391	-14.6603	compared | trained
0.0869565	-14.6604	parsing | dependency
0.14	-14.6605	same | described
0.0765027	-14.6605	conclusion | have
0.25	-14.6606	from | entries.
0.108696	-14.6606	from | pairs.
0.0902256	-14.6606	and | 2003), | and
0.208333	-14.6608	where | highest
0.155556	-14.6609	trained | treebank
0.137255	-14.6609	have | provide
0.222222	-14.661	additional | better
0.0474227	-14.6611	log | the
0.14	-14.6613	optimization | as
0.123077	-14.6614	an | subject
0.0389972	-14.6616	of | candidate
0.0669456	-14.6617	n | i
0.0721154	-14.6618	treebank | al.,
0.0514706	-14.6618	we | knowledge
0.333333	-14.6624	preposition | srl
0.0752688	-14.6627	section | presents
0.133333	-14.6628	v | by
0.222222	-14.6631	each | grouped
0.076087	-14.6632	nullnullnull | null
0.129032	-14.6632	as | outperforms
0.0487805	-14.6633	will | on
0.217391	-14.6636	3 | n-gram
0.0645161	-14.6636	is | opinion
0.222222	-14.6637	sij | =
0.0655738	-14.6644	background | the
0.0680851	-14.6645	d | e
0.185185	-14.6645	three | features,
0.0833333	-14.6646	corpus | number
0.3	-14.6647	sketch | 2004)
0.3	-14.6647	dual | rush
0.0173363	-14.6649	features | the | of
0.0363636	-14.665	is | these
0.0869565	-14.6654	it | them
0.0553936	-14.6656	access | the
0.0386207	-14.6658	annotated | of
0.0680851	-14.666	features | training
0.155556	-14.6661	test | approach
0.2	-14.6662	from | missing
0.148936	-14.6663	feedback | as
0.148936	-14.6663	model | analysis
0.0464646	-14.6664	structure | in
0.0902256	-14.6664	the | process is
0.0503597	-14.6665	a | (see
0.171429	-14.6667	chen | al.
0.121212	-14.667	found | such
0.155556	-14.6671	document | probability
0.0200074	-14.6672	of | system
0.185185	-14.6673	then, | between
0.185185	-14.6673	domain, | on
0.272727	-14.6675	l0 | par.
0.126984	-14.6675	selection | not
0.171429	-14.6677	feature | relative
0.1125	-14.6677	4 | were
0.0675105	-14.6678	m | e
0.142857	-14.6679	lexicon | al.,
0.0828025	-14.668	the | researchers
0.0194377	-14.668	for | which
0.272727	-14.6683	instrument(with | ate)]
0.0882353	-14.6688	annotated | et al.,
0.0973451	-14.669	= | l
0.208333	-14.6691	uk, | 30,
0.222222	-14.6692	comparable | translations
0.0415335	-14.6693	a | 0
0.4	-14.6693	taxes | caja
0.0609319	-14.6698	has | this
0.111111	-14.6698	while | there
0.272727	-14.67	classification | since
0.0778443	-14.67	word | will
0.06	-14.6704	as | so
0.208333	-14.6704	new | pair
0.0792683	-14.6704	analysis | our
0.0736842	-14.6708	although | with
0.0948276	-14.6708	thank | helpful
0.0700935	-14.6709	we | still
0.0773481	-14.6709	2007), | et
0.2	-14.6711	precision | lower
0.157895	-14.6713	second | performance
0.036478	-14.6713	all | we
0.137255	-14.6714	word | null
0.0592105	-14.6716	that | trained
0.133333	-14.6717	tried | to
0.0420875	-14.6719	the | along
0.0973451	-14.672	on | random
0.171429	-14.6725	there | error
0.0406886	-14.6726	event | in
0.137255	-14.6727	only | either
0.03657	-14.6727	a | of | sentences
0.115385	-14.6728	speech | which
0.0482456	-14.6729	we | rule
0.121212	-14.6733	need | can
0.0787879	-14.6736	factor | in
0.157895	-14.6737	algorithms | been
0.0428816	-14.6738	methods | a
0.1125	-14.6739	models | when
0.07109	-14.6739	experiments | et
0.272727	-14.6742	bleu | 31.4
0.111111	-14.6742	probability | if
0.0428082	-14.6742	to | type
0.333333	-14.6746	extract | alignments
0.0312787	-14.6746	fact | the
0.3	-14.6748	appears | rather
0.0432526	-14.6749	c | a
0.0501193	-14.6749	define | to
0.0725389	-14.675	the | conducted
0.5	-14.6751	hierarchical | web
0.131148	-14.6752	features | event
0.0356295	-14.6754	at | that
0.0609319	-14.6754	example, | can
0.333333	-14.6756	output | previous
0.2	-14.676	patterns | (e.g.,
0.0744681	-14.6762	use | other
0.157895	-14.6762	we | hope
0.0321951	-14.6766	c©2011 association for | of
0.3	-14.6767	{ | ∑
0.222222	-14.6767	then | apply
0.272727	-14.6767	felice | 2008;
0.0384088	-14.6768	english | in
0.0474138	-14.6768	we | classifier
0.064	-14.677	definition | to
0.1125	-14.677	which | negative
0.0681818	-14.6771	analysis | are
0.0210831	-14.6772	has | in
0.25	-14.6774	based | similarity.
0.3	-14.6776	allow | analysis
0.0428816	-14.6778	of | small
0.0472103	-14.6778	the same | with
0.0539084	-14.6782	computational | to
0.0171103	-14.6782	the first | the
0.137255	-14.6782	makes | are
0.0888889	-14.6783	this | model,
0.081761	-14.6784	of | third
0.192308	-14.6784	vector | candidate
0.333333	-14.6787	mul | 0
0.10101	-14.679	to | her
0.049763	-14.6793	the | tagging
0.131148	-14.6793	has | 2003;
0.222222	-14.6794	an | closest
0.0351288	-14.6795	for | corpus
0.0964912	-14.6796	time | number
0.157895	-14.6796	adaptation | domain
0.0573248	-14.6797	represent | in
0.0225853	-14.6798	this | an
0.0422297	-14.6798	recall | and
0.185185	-14.6804	hence, | can
0.121212	-14.6804	have | alignment
0.103093	-14.6804	next, | in
0.0807453	-14.6806	6 | 3 | 5
0.0228266	-14.6806	but | to
0.272727	-14.6809	3.2 | phrases
0.235294	-14.681	diagram | figure
0.208333	-14.6813	et | 1994).
0.5	-14.6813	system | 90%.
0.5	-14.6813	should | careful
0.5	-14.6813	'he | rice
0.113924	-14.6816	3 | has
0.0729167	-14.6817	instead | for
0.0348837	-14.6818	tree | and
0.0448343	-14.6819	2006; | al., | al.,
0.0700935	-14.6819	for | coreference
0.0183486	-14.6821	and | using
0.0495283	-14.6823	we | present
0.109756	-14.6824	are | translation.
0.162162	-14.6827	on | fine-grained
0.00937192	-14.6829	on | of
0.3	-14.6832	select | second
0.0403727	-14.6833	the | 2006).
0.113924	-14.6833	,s | s
0.131148	-14.6833	100 | by
0.0609319	-14.6835	data | was
0.0226506	-14.6836	trained | the
0.0279823	-14.6836	to | to be
0.077381	-14.6836	sentences | training
0.0294599	-14.6837	that | between
0.0353357	-14.6838	precision | of
0.0552326	-14.6839	as | task
0.1	-14.6841	used | corpus.
0.4	-14.6841	finite-state | composition.
0.034292	-14.6842	learning | a
0.272727	-14.6843	during | due
0.0512821	-14.6846	in | precision
0.266667	-14.685	static | test
0.0545977	-14.685	information | on
0.0634921	-14.6851	the | validation
0.0274262	-14.6854	of | computational
0.0948276	-14.6854	decision | for
0.0611511	-14.6858	by | test
0.0318533	-14.6859	analysis | the | and
0.5	-14.6859	-0.10 | -0.13
0.3	-14.686	pairs | stored
0.0901639	-14.686	a | 2004).
0.153846	-14.6862	probability | label
0.1	-14.6863	not | high
0.0600707	-14.6864	target | that
0.0529101	-14.6866	to | reordering
0.131148	-14.6866	abstractness | of
0.153846	-14.6867	model | context.
0.162162	-14.6868	= | indicating
0.0244875	-14.687	terms | the
0.333333	-14.687	10 | eh8
0.105263	-14.6871	present | data
0.126984	-14.6871	search | not
0.1125	-14.6871	an | small
0.0394161	-14.6874	and | better
0.0707547	-14.6874	the | closest
0.0839161	-14.6876	our | much
0.0567823	-14.6877	by | 1
0.0894309	-14.6878	approach | al.
0.3	-14.6879	position | j
0.1	-14.6879	a | structural
0.157895	-14.6879	language | size
0.0508906	-14.688	in | work, | of
0.0133156	-14.688	an | and
0.0167509	-14.6882	for | from
0.0756757	-14.6883	language | have
0.0318841	-14.6883	instance, | the
0.171429	-14.6884	the | rejected
0.092437	-14.6887	that | right
0.042517	-14.6889	6 | a
0.115385	-14.6889	s | u
0.0717949	-14.6889	event | by
0.084507	-14.689	however, | used
0.5	-14.689	pronominal | resolution.
0.266667	-14.6891	size | 1000
0.266667	-14.6891	3g | battery
0.266667	-14.6891	(fist | mukka
0.4	-14.6891	(baker | 2007),
0.0467091	-14.6895	are | based
0.208333	-14.6897	incorporate | knowledge
0.113924	-14.6897	the | dog
0.0321012	-14.6898	of | where | a
0.0916667	-14.6899	in | groups
0.1125	-14.6899	for | substitution
0.103093	-14.69	from | second
0.08125	-14.6902	of | pp.
0.102041	-14.6904	propose | approach
0.134615	-14.6904	approaches | into
0.222222	-14.6905	nodes | label
0.171429	-14.6905	variance | to
0.235294	-14.6905	ax | ,
0.0420875	-14.6907	the | created
0.0560748	-14.691	idea | and
0.0634921	-14.691	annotation | with
0.272727	-14.6911	or | command
0.222222	-14.6912	al., | bilmes,
0.109756	-14.6917	although | or
0.272727	-14.692	documents | marcu,
0.0592334	-14.6923	noun | as
0.0535211	-14.6923	and | meaning
0.131148	-14.6923	an | aligned
0.14	-14.6925	by | rank
0.0451866	-14.6926	1 | from
0.0273775	-14.6927	this | it | the
0.0535211	-14.6928	types | as
0.0506024	-14.6929	an | if
0.162162	-14.6929	3 | corpus.
0.4	-14.6929	word | language;
0.222222	-14.6933	function | correct
0.0235726	-14.6934	on | and | in
0.105263	-14.6934	hashtags | the
0.0406886	-14.6934	performance | for
0.0901639	-14.6936	improves | in
0.5	-14.6937	0.62 | 0.75
0.166667	-14.6938	relations | but
0.0694444	-14.6938	statistical | is
0.0362047	-14.6943	features | from
0.333333	-14.6943	are | prevalent
0.0778443	-14.6944	models | training
0.0382436	-14.6946	function | to
0.0620438	-14.6946	to | translation.
0.0567823	-14.6949	on | syntactic
0.0249851	-14.6951	that | this
0.108434	-14.6952	this | lead
0.5	-14.6953	sum | combination
0.272727	-14.6954	be | forming
0.0253243	-14.6955	it is | a
0.0576923	-14.6955	our | learning
0.137255	-14.6957	positive | examples
0.0916667	-14.696	dependency | words
0.043956	-14.6963	that | accuracy
0.333333	-14.6964	unigrams | each
0.0488372	-14.6965	of | 2008).
0.115385	-14.6965	both | accuracy
0.192308	-14.697	not | yet
0.145833	-14.697	joint | pos
0.272727	-14.6971	tweets | positive,
0.272727	-14.6971	they | called
0.0535211	-14.6972	rank | the | of
0.2	-14.6972	be | max
0.14	-14.6973	word | ney,
0.333333	-14.6974	preposition | if
0.192308	-14.6976	here | same
0.185185	-14.6978	0.1 | 0.25
0.2	-14.6978	where | who
0.054755	-14.6979	each | between
0.0894309	-14.6979	p | c
0.4	-14.6979	fragments | —
0.126984	-14.6979	relative | each
0.0405616	-14.6979	in | them
0.0576923	-14.6981	tree | for
0.104167	-14.6986	number | large
0.042328	-14.6987	for | and | all
0.0355392	-14.6987	with | our
0.222222	-14.6989	dependency | attachment
0.0752688	-14.6989	null | nullnullnull | nullnullnull
0.142857	-14.6989	case, | from
0.0869565	-14.6991	approach | training
0.119403	-14.6993	value | between
0.105882	-14.6994	give | that
0.0393939	-14.6994	same | is
0.0618182	-14.6995	we | models.
0.057508	-14.6996	a | among
0.0527778	-14.6998	the | national
0.0409836	-14.7002	each | =
0.153846	-14.7002	number | nouns
0.0729167	-14.7002	distributional | in
0.0792683	-14.7003	by | input
0.157895	-14.7003	is | all.
0.1	-14.7004	unknown | to
0.107143	-14.7004	yet | in
0.25	-14.7009	p(h | p(d
0.132075	-14.7011	t | target
0.0916667	-14.7015	the | kneser-ney
0.0584416	-14.7015	second | for
0.2	-14.7016	test | train
0.0460251	-14.7017	parser | to
0.222222	-14.7017	structure | translation,
0.25	-14.7017	whether | might
0.0248815	-14.7017	the | pages
0.0586207	-14.7018	novel | and
0.10101	-14.7018	entities | or
0.0302222	-14.7019	task | in
0.0740741	-14.7019	lda | of
0.0748663	-14.7022	this | quality
0.0111072	-14.7022	the | or | the
0.222222	-14.7024	related | discussed
0.0755814	-14.7026	all | information
0.0503778	-14.7027	that | time
0.019514	-14.7028	of | other
0.0251688	-14.7028	in | information
0.4	-14.7029	¯e) | prefer
0.0573248	-14.703	about | for
0.0714286	-14.7031	nullnull | the
0.108434	-14.7031	trained | corpus.
0.0894309	-14.7032	determine | or
0.0498812	-14.7032	due to | to
0.25	-14.7033	illustrates | which
0.157895	-14.7033	significant | data
0.208333	-14.7035	automated | system
0.0622568	-14.7036	this | relation
0.0827586	-14.7036	it | time
0.1	-14.7039	these | should
0.0787879	-14.7039	sentence | this
0.235294	-14.7039	be | extending
0.0909091	-14.704	where | but
0.0336957	-14.7041	in | dependency
0.0447471	-14.7042	if | an
0.0455446	-14.7045	we | show
0.0980392	-14.7045	representations | et
0.14	-14.7048	this | aims
0.148936	-14.7053	are | trees.
0.048951	-14.7054	or | are | of
0.107143	-14.7055	system | english
0.0694444	-14.7055	distribution | that
0.0394537	-14.7056	result | of
0.0783133	-14.7056	can | syntactic
0.0948276	-14.7057	work | part
0.088	-14.7058	cluster | as
0.0783133	-14.7059	an | relation
0.0413223	-14.7062	for | it is
0.0778443	-14.7062	only | using
0.0274964	-14.7064	dependency | and
0.0688073	-14.7065	simplification | the
0.185185	-14.7066	given | unlabeled
0.4	-14.7068	current | needs
0.0522193	-14.7069	we | pairs | and
0.1125	-14.707	which | according
0.115385	-14.7071	probabilities | =
0.0725389	-14.7072	and | group
0.0541311	-14.7078	the | lead
0.0452756	-14.7079	of | classifier
0.145833	-14.7081	that | regardless
0.171429	-14.7084	language | even
0.123077	-14.7085	variables | be
0.0932203	-14.709	to | starting
0.210526	-14.709	ser | =
0.119403	-14.7091	an | labels
0.272727	-14.7092	whether | real
0.0559006	-14.7093	expressions | and
0.0424779	-14.7095	which | can
0.040783	-14.7096	of | selected
0.107143	-14.7096	ter | and
0.210526	-14.7096	aim | using
0.0469083	-14.7098	the | longer
0.0269886	-14.71	for | two
0.1125	-14.7102	1 | classification
0.00548309	-14.7103	to | the | of
0.0102403	-14.7103	and | as
0.025293	-14.7104	we | one
0.4	-14.7106	acknowledgments | thanks
0.0292887	-14.7108	to | most
0.00687363	-14.7108	and | in
0.145833	-14.7109	• | probability
0.153846	-14.711	each | tree.
0.5	-14.7112	3.3 | generate
0.126984	-14.7116	combined | from
0.0368906	-14.7116	and | method
0.00654456	-14.7117	for | of
0.036036	-14.7119	of | part
0.148936	-14.7119	hr | hr
0.192308	-14.7122	experiments | perform
0.0916667	-14.7122	4. | are
0.132075	-14.7124	in | engineering
0.3	-14.7125	box | netherlands
0.0663717	-14.7125	gold | for
0.0887097	-14.7127	to | order.
0.5	-14.7128	mturk | example
0.0398773	-14.7128	it | this
0.0512821	-14.7131	for | since
0.00849464	-14.7131	in | is
0.1	-14.7135	to | framework.
0.0887097	-14.7136	is | done
0.0940171	-14.7137	of | bionlp
0.132075	-14.7137	figures | for
0.2	-14.7137	model | scoring
0.031746	-14.7138	the | parameters | the
0.14	-14.7138	show | quality
0.0273871	-14.7139	same | in
0.0529248	-14.7141	from | were
0.0380818	-14.7141	and | parsing
0.105882	-14.7143	reordering | model
0.3	-14.7144	determination | coreference
0.153846	-14.7144	so | do
0.5	-14.7144	lig | cnc
0.1125	-14.7145	to | let
0.192308	-14.7147	some | relevant
0.0519481	-14.7147	set | test
0.104651	-14.7148	--- | --
0.0672646	-14.7149	that | grammar
0.0552147	-14.715	the | map
0.05	-14.715	a | additional
0.103093	-14.7152	from | function
0.0693069	-14.7154	words, | for
0.4	-14.7157	airola, | tapio
0.0714286	-14.7158	to | by using
0.028103	-14.7158	experiments | and
0.222222	-14.7159	those | related
0.5	-14.716	dfki | stuhlsatzenhausweg
0.045929	-14.7162	s | is
0.0280811	-14.7163	to | source
0.0227964	-14.7164	for | it
0.0714286	-14.7165	then | word
0.0810811	-14.7165	here | as
0.0678733	-14.7165	so | on
0.0721649	-14.7166	with | described
0.0249392	-14.7166	of | words | and
0.0409165	-14.7167	in | which | be
0.0714286	-14.717	for | arguments
0.0769231	-14.7171	is | above
0.0787879	-14.7172	with | precision
0.0437262	-14.7174	from | more
0.0961538	-14.7177	probabilities | as
0.134615	-14.7177	that | paraphrasing
0.153846	-14.7179	are | cross
0.0590278	-14.7179	as | pairs
0.0233174	-14.718	and | different
0.210526	-14.7184	be | factors.
0.14	-14.7187	methods | 2008;
0.0553846	-14.719	but | which
0.1125	-14.7192	on | trees
0.0375522	-14.7194	the | who
0.105882	-14.7194	multiword | and
0.109756	-14.7196	parameter | we
0.0571429	-14.7197	new | as
0.157895	-14.72	propose | into
0.117647	-14.7201	dataset | which
0.132075	-14.7201	though | can
0.0970874	-14.7202	standard | language
0.0805369	-14.7203	we would | to thank
0.0527778	-14.7205	y | a
0.088	-14.7207	this | local
0.171429	-14.721	of | realizations
0.15	-14.7211	was | lower
0.0299559	-14.7212	as | our
0.0411184	-14.7214	a | could
0.0498753	-14.7214	this | may
0.0360825	-14.7218	to | probability
0.4	-14.7221	input, | automatically
0.0887097	-14.7224	of | readability
0.210526	-14.7225	viewed | an
0.104651	-14.7229	examples | not
0.222222	-14.723	this is | interesting
0.0678733	-14.7231	c | r
0.15	-14.7231	domains | al.,
0.1	-14.7234	but | there
0.134615	-14.7234	it | individual
0.0237438	-14.7234	will | the | of
0.157895	-14.7235	on | pairs,
0.0714286	-14.7236	and | component
0.108434	-14.7236	statistics | table
0.222222	-14.7237	baseline | dev
0.107143	-14.7237	different | sets
0.0576271	-14.7239	with | phrase
0.0887097	-14.7239	1 | when
0.272727	-14.7241	work, | many
0.0529248	-14.7243	impact | of
0.119403	-14.7245	al., 2008; | 2008;
0.0646552	-14.7245	for | must
0.0482759	-14.7245	not | at
0.0328042	-14.7246	(1) | of
0.134615	-14.7247	investigate | our
0.102041	-14.7248	s | v
0.3	-14.7251	improve | rule
0.2	-14.7253	one | annotators
0.145833	-14.7254	other | relation
0.0617761	-14.7256	k | =
0.0909091	-14.7258	order | different
0.047619	-14.7259	on | results
0.134615	-14.726	are | models,
0.0398724	-14.7261	the | of | pages
0.0344418	-14.7261	words | with
0.185185	-14.7262	int | q
0.0178683	-14.7263	word | a
0.1	-14.7264	these | show
0.210526	-14.7266	large | set.
0.0669643	-14.7268	representations | of
0.14	-14.7268	recently, | been
0.12963	-14.7269	whether | should
0.0787879	-14.7269	of | light
0.0760234	-14.7269	a | polarity
0.157895	-14.7271	similar | into
0.333333	-14.7273	consisted | each
0.210526	-14.7273	nns | →
0.185185	-14.7274	2002; | 2009;
0.044898	-14.7275	4 | on
0.0405844	-14.7276	with | have
0.272727	-14.7276	calculated | n-gram
0.0153918	-14.7278	different | the
0.0470852	-14.728	all | and | are
0.115942	-14.7283	authority | of
0.15	-14.7284	order | identify
0.0435606	-14.7285	framework | of
0.0263889	-14.7285	text | of
0.0952381	-14.7286	it | terms
0.0778443	-14.7289	the | infer
0.0596491	-14.729	new | from
0.0434783	-14.7294	translation | by
0.0821918	-14.7295	have | sentences
0.0517711	-14.7297	types | for
0.4	-14.7299	confinement | templates
0.0421793	-14.73	our | all
0.145833	-14.7302	two | role
0.153846	-14.7303	in | judged
0.0331858	-14.7304	more | that
0.0586207	-14.7306	but | can
0.0655022	-14.7308	conference on | in
0.0217391	-14.7308	data | in
0.12963	-14.7311	variational | the
0.12963	-14.7316	similar | was
0.0486111	-14.7319	in | provide
0.125	-14.7322	one | relations
0.0418848	-14.7323	5 | we
0.109756	-14.7324	4 | was
0.178571	-14.7325	pos | tokens
0.210526	-14.7327	direction | future
0.0280811	-14.7329	in | we | our
0.0894309	-14.7329	conclusion | presented
0.178571	-14.733	context | words.
0.272727	-14.733	that | transmitted
0.0703518	-14.733	improvements | to
0.2	-14.7331	it | 20
0.0606061	-14.7333	do | or
0.0424028	-14.7334	patterns | of
0.0376045	-14.7335	of | rule
0.0545455	-14.7336	so | that
0.0449898	-14.7337	score | for
0.4	-14.7338	r | r+1
0.192308	-14.734	experiment | table
0.0303588	-14.734	data, | the
0.0408859	-14.734	topic | in
0.125	-14.7342	features | terms
0.162162	-14.7344	data | level
0.0206403	-14.7344	by | as
0.00959971	-14.7344	the | of | are
0.178571	-14.7348	rna | protein
0.210526	-14.7348	has | developing
0.0682927	-14.7351	and | np
0.0287829	-14.7354	we | models
0.121212	-14.7356	how | learning
0.0101873	-14.7358	and | a | of
0.0901639	-14.7358	then | 1
0.0403877	-14.736	for | dependency
0.0435606	-14.7361	in | entity
0.0738636	-14.7362	is | the previous
0.0655022	-14.7363	feature | using
0.132075	-14.7364	unsupervised | based
0.0952381	-14.7367	a | level.
0.0833333	-14.7368	our | annotation
0.5	-14.7371	“x | 보다
0.0527778	-14.7373	we | small
0.157895	-14.7374	two | task.
0.0498753	-14.7374	corpora | in
0.0638298	-14.7376	idea | is
0.127273	-14.7377	for | amounts
0.4	-14.7378	pb(γprime|¯t,γ,
0.0239726	-14.7378	of | test
0.25	-14.7379	be | independently.
0.166667	-14.7382	(baker | al.,
0.0204337	-14.7382	using | in
0.111111	-14.7382	more | compared
0.235294	-14.7388	x1 | x4
0.115942	-14.7389	concepts | for
0.178571	-14.7389	new | target
0.178571	-14.7395	more | provided
0.0887097	-14.7395	correction | to
0.0301645	-14.7399	pairs | and
0.0916667	-14.7399	target | it
0.134615	-14.74	et | chiang
0.112676	-14.74	when | should
0.0294883	-14.7401	to | models
0.132075	-14.7407	relation | but
0.0147569	-14.7408	for | are
0.166667	-14.741	verb | following
0.132075	-14.7411	inserted | of
0.0894309	-14.7411	should | more
0.0154799	-14.7413	the | semantic
0.333333	-14.7415	form | e1
0.333333	-14.7415	recognize | entities
0.0400641	-14.7416	to | relations
0.0679612	-14.7417	model | performance
0.123077	-14.742	markov | al.,
0.0669643	-14.7421	use | all
0.102273	-14.7425	we | preliminary
0.1	-14.743	that | slightly
0.07	-14.7432	implicit | the
0.0421245	-14.7432	that | about
0.123077	-14.7432	number | divided
0.0538922	-14.7434	methods | language
0.088	-14.7434	will | such
0.162162	-14.7435	first | label
0.0610687	-14.7436	reduced | the
0.0625	-14.7437	as | human
0.0152813	-14.7437	translation | the
0.01864	-14.744	the | dependency
0.0925926	-14.7441	we | types.
0.0225064	-14.7445	a | two
0.210526	-14.7445	 | 
0.0178515	-14.7449	of | two
0.07	-14.7452	used | al.
0.146341	-14.7454	were | rate
0.272727	-14.7455	comparable | reported
0.112676	-14.7455	a | 2001;
0.0638298	-14.746	to | objective
0.0952381	-14.7461	are | comparable
0.027735	-14.7463	as | can
0.0344009	-14.7469	lexical | a
0.0201729	-14.7469	two | to
0.5	-14.7469	prov | blac
0.153846	-14.747	used | svm
0.192308	-14.7473	document | label
0.0630252	-14.7473	along | and
0.0508021	-14.7476	learning | as
0.0369357	-14.748	of | t
0.107143	-14.7483	in | half
0.109756	-14.7483	learning | two
0.0501253	-14.7488	and | rule
0.178571	-14.7489	35 | 1
0.0613027	-14.7489	for | often
0.134615	-14.7489	an | wikipedia
0.172414	-14.7491	(2008) | method
0.104651	-14.7493	from | argument
0.2	-14.7496	authors | anonymous
0.172414	-14.7497	as | research.
0.0214953	-14.7497	similar | the
0.132075	-14.7499	to | prefix
0.0514905	-14.75	length | in
0.333333	-14.7503	strictly | tier
0.172414	-14.7508	conducted | data
0.0113857	-14.7509	a | of | is
0.4	-14.751	state | known
0.0408859	-14.7515	to | alignment
0.157895	-14.7515	translation | 2006).
0.0421245	-14.7516	recent | the
0.3	-14.7517	minimum | observations.
0.172414	-14.7519	approach | systems.
0.0517711	-14.752	state | to
0.0580205	-14.7522	to | amount
0.142857	-14.7522	difference | all
0.115942	-14.7523	for | solution
0.0630252	-14.7523	to | effective
0.0934579	-14.7523	indicate | are
0.192308	-14.7524	their | vectors
0.0436508	-14.7524	is | similarity
0.333333	-14.7525	j=1 | parenrightbigg
0.333333	-14.7525	morphemes | pos-3grams
0.0873016	-14.7526	we | parsers
0.15	-14.7526	task | do
0.0917431	-14.7527	of | acl
0.0925926	-14.7527	we | conditional
0.0586207	-14.7528	evaluation | from
0.0488998	-14.7528	likely | to
0.25	-14.7528	target | good
0.235294	-14.7528	coefficient | between
0.0535714	-14.7529	3. | of
0.00976731	-14.7535	in | of | in
0.25	-14.7536	msa | dialect
0.25	-14.7536	00001 | l | f-­‐
0.25	-14.7536	#offolds | merit
0.25	-14.7536	trivspa | mapppt
0.25	-14.7536	" | baltic
0.25	-14.7536	arg0 | arg0
0.25	-14.7536	agenda | ˆ
1	-14.7536	|w|is
1	-14.7536	gak(z(a1),...,z(a|n|)).
1	-14.7536	note|t|the
1	-14.7536	|v|n,
1	-14.7536	p(source|target)
1	-14.7536	pr(w|t)
1	-14.7536	p(s,z|d,θ)c(w,s).
1	-14.7536	p(ζi|r)
1	-14.7536	(nn|nns))
1	-14.7536	verb|verbal
1	-14.7536	o(|r|·|s|).
1	-14.7536	p(e2|e1)=1),
1	-14.7536	uniontext|t|i=1si
1	-14.7536	f(e(cij)|si)
1	-14.7536	x∈{0,1}|p|
1	-14.7536	although|vpos|>|vphr|,
1	-14.7536	p(k|θ)
1	-14.7536	o(|r||s|).
1	-14.7536	|c|is
1	-14.7536	q(v|c(v))
1	-14.7536	f(d(v1,e),...,d(v|e|,e)))
1	-14.7536	s1s2...s|s|
1	-14.7536	|hc|we
1	-14.7536	(1−θ)producttextii=1θpt(µi)lp(|µi|)
1	-14.7536	|t|lessmuch|e|
1	-14.7536	|val|
1	-14.7536	θw|t∼dirichlet(αw|t,wt)
1	-14.7536	w0,...,w|w|−1.
1	-14.7536	rank(wc|wm)
1	-14.7536	{s1,s2,...,s|s|}and
1	-14.7536	measuretheimportancescorepr(w|d)ofeachword
0.333333	-14.7536	bord, | irish,
0.333333	-14.7536	p2r5 | p5r2
0.333333	-14.7536	sg | pl
0.333333	-14.7536	grammys | (-1)
0.333333	-14.7536	regulon | promoter
0.333333	-14.7536	fanfiction | sci-fi
0.333333	-14.7536	路滑 | 路
0.333333	-14.7536	λ1=0.4 | λ1=0.6
0.333333	-14.7536	org | 0.9944
0.333333	-14.7536	pcfg | tlm
0.333333	-14.7536	0.984 | 0.971
0.333333	-14.7536	blutdruckabfall | dysphagie
0.333333	-14.7536	publication | affiliation
0.333333	-14.7536	c2(y) | c2(x)
0.333333	-14.7536	73.0 | 77.4
0.333333	-14.7536	économiques | v
0.333333	-14.7536	a2l | a2fst
0.333333	-14.7536	iq-med | iq
0.333333	-14.7536	77.40 | 63.50
0.333333	-14.7536	german | spelling.
0.333333	-14.7536	restaurant; | franchise;
0.333333	-14.7536	[%] | 2ndbl
0.333333	-14.7536	λm | ≡
0.333333	-14.7536	travailler | id´ees
0.333333	-14.7536	φ(m1, | t7)
0.333333	-14.7536	wi | φ(ci)
0.333333	-14.7536	(y/n); | mj
0.333333	-14.7536	laurent | kickers
0.333333	-14.7536	−2.1 | rec
0.333333	-14.7536	src/trg | 44,658
0.333333	-14.7536	va | hir
0.333333	-14.7536	h2 | dh2
0.333333	-14.7536	technologies | pitsburgh,
0.333333	-14.7536	λ2=0.6 | λ2=0.2
0.333333	-14.7536	b] | c]
0.333333	-14.7536	~ | 0.990
0.333333	-14.7536	tlm | type
0.333333	-14.7536	ribavirin | xeristar
0.333333	-14.7536	xa | aβ(λ)
0.333333	-14.7536	(vp | xv
0.2	-14.7536	cow, | ate)
0.5	-14.7536	dir(a) | a’s
0.5	-14.7536	-に | ni]
0.5	-14.7536	material:metal) | feet)
0.5	-14.7536	yer | codi
0.5	-14.7536	|t) | p(t|θ+i
0.5	-14.7536	avanza | (mec
0.5	-14.7536	rd | 
0.5	-14.7536	z1,3i | z3,4i
0.5	-14.7536	 | 
0.5	-14.7536	<n(se), | jcs,
0.5	-14.7536	p(post | p(gt
0.5	-14.7536	fora∈σ | 16:
0.5	-14.7536	interruption | ipw
0.5	-14.7536	alguma | somebody
0.5	-14.7536	p(y|x,β) | q(x)
0.5	-14.7536	(abbcdd, | (bbcdd,
0.5	-14.7536	0.1157 | 0.1602
0.5	-14.7536	said | (t30)
0.5	-14.7536	lj | 
0.5	-14.7536	omu | house
0.5	-14.7536	⎪ | isssim
0.5	-14.7536	∼n(0, | regression
0.5	-14.7536	kata | tak
0.5	-14.7536	dephosphorylation | species:state:inactive
0.5	-14.7536	1373 | 
�]
0.5	-14.7536	bunk | glimmer
0.5	-14.7536	z1,2i | z2,3i
0.5	-14.7536	ambroise | men
0.5	-14.7536	sie | (‘the
0.5	-14.7536	recommend | easi
0.5	-14.7536	 | deepsim
0.5	-14.7536	10m | 18.60
0.5	-14.7536	55.29 | 42.65
0.5	-14.7536	到 | 31%
0.5	-14.7536	indexed | i169
0.5	-14.7536	fatih | buyukcekmece,
0.5	-14.7536	amod | partmod
0.5	-14.7536	notr | .,
0.5	-14.7536	x1, | x1⟩
0.5	-14.7536	4.13 | 4.23
0.5	-14.7536	normalized, | aiµ
0.5	-14.7536	theme(protein) | theme(core
0.5	-14.7536	route | (known/unknown),
0.5	-14.7536	china | pillar
0.5	-14.7536	−log | ¯e)
0.5	-14.7536	w/or | doinq
0.5	-14.7536	tour | trick’.
0.5	-14.7536	f(y,x) | y∈y(x)
0.5	-14.7536	log10 | impression
0.5	-14.7536	uturku | sd
0.5	-14.7536	p(qj | ck)
0.5	-14.7536	analysis-by-synthesis | utterance,
0.5	-14.7536	theme, | grasp(grasper:washer,
0.5	-14.7536	held | �/nrf+|
0.5	-14.7536	food | holmdel
0.5	-14.7536	of) | of).
0.5	-14.7536	q(zi,j,ww | updates
0.5	-14.7536	−θtf(x(i),y(i)) | exp{θtf(x(i),y)
0.5	-14.7536	faust+all | (arguments)
0.5	-14.7536	ce | nationality,
0.5	-14.7536	rec | 53.6
0.5	-14.7536	t17 | t16
0.5	-14.7536	nnp1 | nnp2↔nnp2
0.5	-14.7536	pr | em(n){
0.5	-14.7536	2.000 | 0.74%)
0.5	-14.7536	14.98 | 37.88
0.5	-14.7536	25.27 | 39.09
0.5	-14.7536	lov | 0.0160
0.5	-14.7536	target | co-target
0.5	-14.7536	mumbai | india
0.5	-14.7536	7.91 | 8a
0.5	-14.7536	control, | experimento}
0.5	-14.7536	(3); | vping
0.5	-14.7536	lado | olho
0.5	-14.7536	tisue | (<finding
0.5	-14.7536	j∈δi(a) | aer
0.5	-14.7536	hideous | sweet
0.5	-14.7536	100,000 | 95,912
0.5	-14.7536	cache | compressed
0.5	-14.7536	(or | ibmapple
0.5	-14.7536	1,224 | reshuffling
0.5	-14.7536	4.73 | inorder
0.5	-14.7536	=d | c〉(ǫ)
0.5	-14.7536	wh-questions | yes/no-questions
0.5	-14.7536	t-pos(irc | t-pos(ptb
0.5	-14.7536	c(w,e) | e,
0.5	-14.7536	ourfe | 4-class
0.5	-14.7536	plakias | 2008a;
0.5	-14.7536	±3.5 | game
0.5	-14.7536	d82 | buy
0.5	-14.7536	trivlp | pmaplpt
0.5	-14.7536	sentchunk | 0.793
0.5	-14.7536	b2b | 27.47
0.5	-14.7536	mention), | (number
0.5	-14.7536	; | [o1/
0.5	-14.7536	estimating | βt,j
0.5	-14.7536	explain | complexq;
0.5	-14.7536	svs | svs
0.5	-14.7536	sex; | remov
0.5	-14.7536	weight(si,fy) | v(fy)
0.5	-14.7536	+summationtextv | +summationtexts
0.5	-14.7536	instprec- | 89.8
0.5	-14.7536	(∃sprimeprime,sprimeprimeprime) | sprimeprimeprimesprimeprimesprime
0.5	-14.7536	intra-rater | (sd)
0.5	-14.7536	freq | 0.006%
0.5	-14.7536	895 | 14.07
0.5	-14.7536	=summationtexti | pr(ei|w)
0.5	-14.7536	eine | acc.sg.fem
0.5	-14.7536	acld | sensing
0.5	-14.7536	‘every | dhiire
0.5	-14.7536	engineering, | portugal
0.5	-14.7536	entity_b | appos
0.5	-14.7536	gazel | inhabitants
0.5	-14.7536	came | tercer
0.5	-14.7536	tom | (some)
0.5	-14.7536	ngrams | 1–2
0.5	-14.7536	berlin | european
0.5	-14.7536	1211 | idiap
0.5	-14.7536	(t): | (c)
0.5	-14.7536	candidates | ”vv*
0.5	-14.7536	s) | d(f,
0.5	-14.7536	quantities | ingestion
0.5	-14.7536	(y) | κi
0.5	-14.7536	septembrie”, | 050711,
0.5	-14.7536	き | さ
0.5	-14.7536	(predicted: | (oracle:
0.5	-14.7536	music-en | book-en
0.5	-14.7536	vectorxj) | ξij
0.5	-14.7536	 events | recognizer
0.5	-14.7536	makro | 10-ma
0.5	-14.7536	jour’ | wine’
0.5	-14.7536	adaptive | completion,
0.5	-14.7536	8.7 | 97.9
0.5	-14.7536	p(wj|taj) | ˆa,
0.5	-14.7536	-um | doubled,
0.5	-14.7536	polynomials | taylor
0.5	-14.7536	minimizing | normally
0.5	-14.7536	mo | -0
0.5	-14.7536	(involvement,hcmv) | embedding
0.5	-14.7536	a2l | scunc
0.5	-14.7536	</pat> | matches
0.5	-14.7536	kas | tikai
0.5	-14.7536	room, | subzone
0.5	-14.7536	bauska | zemgale
0.5	-14.7536	drogas | drugs
0.5	-14.7536	girl) | girl)(n
0.5	-14.7536	t(p) | t|p
0.5	-14.7536	sti | πdeprtj
0.5	-14.7536	' | ,%
0.5	-14.7536	stn/doc | ads
0.5	-14.7536	18.87 | 24.34
0.5	-14.7536	figueiredo† | t´ecnico,
0.5	-14.7536	(decodeset | lexicon
0.5	-14.7536	tapeware | qa150-exat
0.5	-14.7536	larv | r=
0.5	-14.7536	87.70 | 88.33
0.5	-14.7536	6×104 | 87.97
0.5	-14.7536	imi | iri
0.5	-14.7536	kill/beat) | thaa
0.5	-14.7536	score(ys,yt,(s(i),s(j)),(t(i),t(j))) | (s(i),s(j))∈ys
0.5	-14.7536	sdrs | q-elab(a,b)∧qap(b,pi)∧plan-elab(b,pi)
0.5	-14.7536	150 | 0.0001k
0.5	-14.7536	c(ˆyt)t | update
0.5	-14.7536	void | model.getlattices();
0.5	-14.7536	λf | c(f,
0.5	-14.7536	igr | 0.566836
0.5	-14.7536	'rice-acc | [pap-ul
0.5	-14.7536	ntt | 619-0237
0.5	-14.7536	give | thine
0.5	-14.7536	diphosphate | digitorum
0.5	-14.7536	csd | l∈l
0.5	-14.7536	sign(•) | macr
0.5	-14.7536	fak | k;
0.5	-14.7536	q(j) | b(j)
0.5	-14.7536	(1989) | pleasantness
0.5	-14.7536	(monolingual) |  corpus
0.5	-14.7536	3.58 | 4.54
0.5	-14.7536	‘which | ‘when
0.5	-14.7536	>>< | ∈eand
0.5	-14.7536	journalist | check
0.5	-14.7536	revenuepct1 | employedat1
0.5	-14.7536	γf→e(fi | fj
0.5	-14.7536	apfel | l¨asst
0.5	-14.7536	fβ | tp/(tp
0.5	-14.7536	vy | ui∈vf\v
0.5	-14.7536	bcubed | 0.4336
0.5	-14.7536	“everything | something”
0.5	-14.7536	loccity{in | wrong
0.5	-14.7536	null3,8null | null5null
0.5	-14.7536	cucchiarini | franco
0.5	-14.7536	soft | (<rolegroup>
0.5	-14.7536	58.85 | 36.63
0.5	-14.7536	exp(w | hatwidew
0.5	-14.7536	nac | sbarq
0.5	-14.7536	,li−1,t) | ,t)
0.5	-14.7536	super | pandemic
0.5	-14.7536	representation: | (start(e),
0.5	-14.7536	td,k+α | wd,n
0.5	-14.7536	click | non-clicks.
0.5	-14.7536	q∈|q| | t∈q,q∈|q|
0.5	-14.7536	forbes” | murray”
0.5	-14.7536	3,060,486 | plitg
0.5	-14.7536	(×1000); | tags;
0.5	-14.7536	phr | [as
0.5	-14.7536	[β,γ]∩pimagea([b,c]c) | hold
0.5	-14.7536	73.94 | 73.53
0.5	-14.7536	drss | x∈ωk
0.5	-14.7536	brighton | {a.s.belz,e.y.kow}@brighton.ac.uk
0.5	-14.7536	∆a | ×1.3
0.5	-14.7536	(1best) | imperatives
0.5	-14.7536	residue:state:methylated | ubiquitination
0.5	-14.7536	 news |  period
0.5	-14.7536	residue:state:phosphorylated | species:state:active
0.5	-14.7536	0.84522 | 0.82607pi
0.5	-14.7536	boxoffice | doc
0.5	-14.7536	audimus | morgan,
0.5	-14.7536	a234 | (sentence
0.5	-14.7536	typhoon | cit
0.5	-14.7536	p(αnullτ) | p(rα(p)nullτ)
0.5	-14.7536	nm,t,j | returned
0.5	-14.7536	(forall | x3)
0.5	-14.7536	csr | evidence;
0.5	-14.7536	‘to | lit:
0.5	-14.7536	f-score | 35.96%on
0.5	-14.7536	52.3 | 143
0.5	-14.7536	62" | 1>
0.5	-14.7536	[lastname] | yon
0.5	-14.7536	[j|β], | 0,j.inegationslash→∗
0.5	-14.7536	pi1a | pi2b
0.5	-14.7536	r´eserver, | r´eservation,
0.5	-14.7536	p(e|f)·s(e,eprime)
0.5	-14.7536	barackobama, | veterans,
0.5	-14.7536	goodman | count(f,t)
0.5	-14.7536	propositions; | hi
0.5	-14.7536	”hot | ”warm
0.5	-14.7536	q(z) | ✇
0.5	-14.7536	pre-ordering | stir:
0.5	-14.7536	nullnullnull[(nullnull1(null) | readings,
0.5	-14.7536	tilburg | netherlands,
0.5	-14.7536	twco | )(
0.5	-14.7536	(arg0 | arg3
0.5	-14.7536	s1), | s2).
0.5	-14.7536	ur | 85.53
0.5	-14.7536	a22 | affine
0.5	-14.7536	turn-initial | echo]
0.5	-14.7536	mt+english | crosslingual
0.5	-14.7536	serbian | 61.4±0.3
0.5	-14.7536	,m,x | :[h,h,x
0.5	-14.7536	x0:rb | x1:pp
0.5	-14.7536	jcj | collocated
0.5	-14.7536	”13 | bucharest
0.5	-14.7536	trie-l | 4.03
0.5	-14.7536	definition, | ⊆eprime1∩eprime2
0.5	-14.7536	grassland | action:
0.5	-14.7536	//h | this
0.5	-14.7536	보다 | <n,
0.5	-14.7536	subpos | adjs
0.5	-14.7536	passvp(expressed) | prep(from)
0.5	-14.7536	interleaved | arg[2,3]
0.5	-14.7536	cota | regulon
0.5	-14.7536	q(θ) | q(z)
0.5	-14.7536	=<schooner: | schooner
0.5	-14.7536	q-elab(a,b) | plan-correction(c,d)
0.5	-14.7536	diferences | diferences.
0.5	-14.7536	0.4324 | 0.2883
0.5	-14.7536	sa | multimodally
0.5	-14.7536	¯vp | (¯s
0.5	-14.7536	conv(h;h) | interior
0.5	-14.7536	g396g349g448g349g287g396g286g856g296g449g374g856g1005g899 | g18g367g381g400g286g282g855g898g364g258g449g258g856g361g449g374g856g1005g853g3
0.5	-14.7536	liu1 | {feiliu,
0.5	-14.7536	null2, | mnull,
0.5	-14.7536	100.2 | 0%
0.5	-14.7536	28,887 | spa
0.5	-14.7536	hashtags | #musicmondays
0.5	-14.7536	chem | (13)
0.5	-14.7536	laughing | (fucking)
0.5	-14.7536	null′ | nullαnull
0.5	-14.7536	⟨2,x⟩ | h′⟨1,y1⟩
0.5	-14.7536	absence | |||0.12
0.5	-14.7536	-opt | 51.6
0.5	-14.7536	rec | 33.4
0.5	-14.7536	non-­‐excessive | peak
0.5	-14.7536	proscribed | [the
0.5	-14.7536	derivative | ∂lik
0.5	-14.7536	e=1 | ne,f
0.5	-14.7536	-null | nullwe,
0.5	-14.7536	〈e,i〉 | time])
0.5	-14.7536	producttext|si| | si∈s
0.5	-14.7536	p(w1 | hatwidewi
0.5	-14.7536	kumar | r(eprime)
0.5	-14.7536	abstractive | (1.3%)
0.5	-14.7536	(clusters) | intr
0.5	-14.7536	(双方 | 立场
0.5	-14.7536	liu1 | bosch
0.5	-14.7536	defined, | φ′∧x
0.5	-14.7536	pred | pl
0.5	-14.7536	s105s115s105 | s100
0.5	-14.7536	r-w | 0.161
0.5	-14.7536	eating | drinking
0.5	-14.7536	αei,x | θx+n−i·,x
0.5	-14.7536	axiom | non-k
0.5	-14.7536	26.9% | 24.3%
0.5	-14.7536	nullinstrumentnull: | nullthemenull:
0.5	-14.7536	bayes | g(eprime)
0.5	-14.7536	xadj1 | (vp2
0.5	-14.7536	inp | wpwpwp
0.5	-14.7536	(ke | fingerprints
0.5	-14.7536	ctf | 88.7
0.5	-14.7536	,x,b) | (4.6)
0.5	-14.7536	model4 | (f,e)
0.5	-14.7536	post | nullnullnullnull|nullnull
0.5	-14.7536	x2:vp) | p(yˇu)
0.5	-14.7536	dtnn | arrowdblbothv
0.5	-14.7536	dimension | general-purpose
0.5	-14.7536	(∑ | w′c
0.5	-14.7536	0.051 | 0.018
0.5	-14.7536	10] | b-hmm
0.5	-14.7536	no.13 | romania
0.5	-14.7536	present(lp) | affect(pp)
0.5	-14.7536	(two | language):
0.5	-14.7536	rpf | 0.427
0.5	-14.7536	‘die | entstehen’
0.5	-14.7536	saar | 244
0.5	-14.7536	libya | trail
0.5	-14.7536	fora∈σ | ˆe←arg
0.5	-14.7536	perfect | something”
0.5	-14.7536	summationtext[[w]] | βi(v)
0.5	-14.7536	p(wn|wn−11 )
0.5	-14.7536	0.6533 | 0.9431
0.5	-14.7536	(sec | configurational
0.5	-14.7536	l=1 | −tg
0.5	-14.7536	dif(wc) | cref(wc)
0.5	-14.7536	roberts | bhutto
0.5	-14.7536	bardbld | (z=k)
0.5	-14.7536	relative, | springt’
0.5	-14.7536	hairy | shaggy
0.5	-14.7536	dvki | ps(zi
0.5	-14.7536	|imp | [i])
0.5	-14.7536	¯pp | ¯qp
0.5	-14.7536	95.7 | 98.4
0.5	-14.7536	ivl(r) | ]r1,r2[
0.5	-14.7536	d53d53 | n’t
0.5	-14.7536	null2(null)) | nullnullnullnull[null(nullnullnull)nullnullnull(nullnullnull)]
0.5	-14.7536	lee | pohang
0.5	-14.7536	states, | δ→r+
0.5	-14.7536	mode | 79.01
0.5	-14.7536	p | 97.98
0.5	-14.7536	r-2 | 0.084
0.5	-14.7536	50.34 | 15.09
0.5	-14.7536	“total | 0.835
0.5	-14.7536	s32s83s72 | s51
0.5	-14.7536	team | atheleteplayesforteam−−−−−−−−−−−−−−→concept
0.5	-14.7536	regroup | y))
0.5	-14.7536	62.4 | 69.0
0.5	-14.7536	ψi( | 0,
0.5	-14.7536	0.296 | 0.113
0.5	-14.7536	i∈rt | rouge-n-f
0.5	-14.7536	ept | down-sampled
0.5	-14.7536	0.5984 | 0.5418
0.5	-14.7536	kproductdisplay | +αt|t)(mitprime|t)
0.5	-14.7536	1> | horizon:
0.5	-14.7536	null4 | null,5
0.5	-14.7536	0.81.0 | 100%75%
0.5	-14.7536	øj | j¢
0.5	-14.7536	cooccurrence | cross-classified
0.5	-14.7536	ref | _是
0.5	-14.7536	—– | 75.19
0.5	-14.7536	mitchell p. | mary
0.5	-14.7536	bunt | petukhova,
0.5	-14.7536	ˆs | argmaxs(producttext
0.5	-14.7536	automotive | english→swedish
0.5	-14.7536	a; | schemei
0.5	-14.7536	americanos, | venir
0.5	-14.7536	leanne | charles
0.5	-14.7536	(-) | (svm-light)
0.5	-14.7536	s105s111 | s109
0.5	-14.7536	table” | “return
0.5	-14.7536	part, | perfective
0.5	-14.7536	recognizes | (expressing
0.5	-14.7536	ht, | pmi(hw,mw)
0.5	-14.7536	duc-06 | 9.51
0.5	-14.7536	69.23 | 67.31
0.5	-14.7536	pbta | 〈q,σ,δ,q0,ω,λ〉
0.5	-14.7536	"medtronic" | (#pacemakers,
0.5	-14.7536	+mono | finnish
0.5	-14.7536	abdomen | arthropod
0.5	-14.7536	maximize | eprime∈e
0.5	-14.7536	˜ | regularization:
0.5	-14.7536	saddened | anger
0.5	-14.7536	0.75 | go,
0.5	-14.7536	eprime∈e | l,
0.5	-14.7536	mises | context-features
0.5	-14.7536	unsmoothed | dk∈ci
0.5	-14.7536	nsummationdisplay | d2summationdisplay
0.5	-14.7536	hd-cmp | plr
0.5	-14.7536	27.3% | 37.5%
0.5	-14.7536	lin | α=0.05
0.5	-14.7536	(vp | (vb
0.5	-14.7536	e) | str(f)
0.5	-14.7536	f+b | bureau
0.5	-14.7536	56% | -3.2%
0.5	-14.7536	35 | %mis
0.5	-14.7536	rsing | eech
0.5	-14.7536	0.754 | cap-macro
0.5	-14.7536	“left” | “towards
0.5	-14.7536	pb(wm1 | m∏
0.5	-14.7536	(85, | (92,
0.5	-14.7536	assail, | attacking
0.5	-14.7536	hifst | compact
0.5	-14.7536	isio | preposition
0.5	-14.7536	41.63 | 41.81
0.5	-14.7536	(λ,µ) | µarg1i,j,p
0.5	-14.7536	{experimentaci´on, | an´alisis,
0.5	-14.7536	prefer | 79
0.5	-14.7536	gapping | ccp
0.5	-14.7536	losman, | francisco
0.5	-14.7536	x(l) | ∂ˆl
0.5	-14.7536	−naw | nu]
0.5	-14.7536	vectorf(x,y) | gen(x)
0.5	-14.7536	x | compatiblewith
0.5	-14.7536	ˆyτ,yjcj) | θ
0.5	-14.7536	entailment | glickman,
0.5	-14.7536	sistemas | t´ecnico,
0.5	-14.7536	tense=pres | eq
0.5	-14.7536	a1 | am-tmp
0.5	-14.7536	paraphrase | 〈“killed”,
0.5	-14.7536	”v | hits
0.5	-14.7536	wt | pi,j)
0.5	-14.7536	63687378838893 | ami
0.5	-14.7536	dik | cpk
0.5	-14.7536	training: | 27.53
0.5	-14.7536	[subj] | ][
0.5	-14.7536	kingdom | philippines
0.5	-14.7536	sem=disease | sem=symptom
0.5	-14.7536	⟨v | competing
0.5	-14.7536	swimming | high
0.5	-14.7536	pjisummationtext | k∈out(vj)|pjk|
0.5	-14.7536	tm | p(f,a)
0.5	-14.7536	(subset | history
0.5	-14.7536	n-best-list | ˜q(p′,p,tn)
0.5	-14.7536	infmod | classes
0.5	-14.7536	21.57 | method,
0.5	-14.7536	bestehen | fest
0.5	-14.7536	vbn\ | (after,
0.5	-14.7536	ˆti1 | ti1
0.5	-14.7536	score(ui,d) | ui
0.5	-14.7536	z2z | (i;t
0.5	-14.7536	manje | pep
0.5	-14.7536	tango | (17)
0.5	-14.7536	(wm | 0.51
0.5	-14.7536	nnlm | 132.2
0.5	-14.7536	arg1) | arg1:
0.5	-14.7536	epsilon1 | λa
0.5	-14.7536	arg1 | temporal.asynchronous
0.5	-14.7536	npnp | dui4wu3
0.5	-14.7536	26.81 | 19.68
0.5	-14.7536	nns\ | child(after,
0.5	-14.7536	hz, | eeg
0.5	-14.7536	james | sub-fields
0.5	-14.7536	au4 | thre
0.5	-14.7536	∝ | favik
0.5	-14.7536	¯r(x)) | y∈y(x)
0.5	-14.7536	>> | −2)
0.5	-14.7536	dt+nnp | offensive
0.5	-14.7536	’\(ncsubj | \w*{w2}\w*_(\d+)\)’,
0.5	-14.7536	hospice | eat
0.5	-14.7536	k=1 | d1summationdisplay
0.5	-14.7536	rouge-n | s∈{refsum}
0.5	-14.7536	e] | sent[b,
0.5	-14.7536	12s2 | ¯a
0.5	-14.7536	cat=det | cat=v
0.5	-14.7536	gr | 0.7824
0.5	-14.7536	thrown | jail
0.5	-14.7536	1−pprime | pprime
0.5	-14.7536	ω | √n
0.5	-14.7536	|||0.06 | 4gram
0.5	-14.7536	relem | zanmim
0.5	-14.7536	tprime≤t≤|e|,
0.5	-14.7536	prp | dt\np
0.5	-14.7536	0.570 | 0.853∗
0.5	-14.7536	▹span | scorejoint(x,ti,tr,tr+1,tj,p
0.5	-14.7536	visualisation | visualisations
0.5	-14.7536	entity; | “you”
0.5	-14.7536	unaries | 11.2%
0.5	-14.7536	0.06 | 40.64
0.5	-14.7536	academia | university
0.5	-14.7536	iff | k-testable.
0.5	-14.7536	ki | tranblemannt`e
0.5	-14.7536	a103 | a108
0.5	-14.7536	wo | nslite
0.5	-14.7536	meetings | (lex
0.5	-14.7536	index | ∅)}
0.5	-14.7536	symptoms | denominations
0.5	-14.7536	∂f⟨2,x⟩ | ∂f⟨2,k2,e2,x⟩
0.5	-14.7536	reversal | tree-crf
0.5	-14.7536	ynq; | i/t
0.5	-14.7536	π24, | addlocation
0.5	-14.7536	va | (chir),
0.5	-14.7536	inserted | “(n
0.5	-14.7536	0.40.6 | 50%25%
0.5	-14.7536	x1 | x0
0.5	-14.7536	89.7 | 1.6x
0.5	-14.7536	ice | eqwe
0.5	-14.7536	def= | 〈[3,9],e〉
0.5	-14.7536	 | in|rels
0.5	-14.7536	eprime | r(eprime)
0.5	-14.7536	(y|x)
0.5	-14.7536	prn | ي
0.5	-14.7536	cat=adv | cat=det
0.5	-14.7536	without] | cd28
0.5	-14.7536	comments, | comment
0.5	-14.7536	f0 | 200ms;
0.5	-14.7536	wi | idf(wij)
0.5	-14.7536	t>b | 63
0.5	-14.7536	30.1 | 13.2
0.5	-14.7536	attr+val | 11.5%
0.5	-14.7536	aer, | listowel,
0.5	-14.7536	desrt | 6.86
0.5	-14.7536	beihan | (english
0.5	-14.7536	u. | w911nf-101-0533,
0.5	-14.7536	(i;t | z2y
0.5	-14.7536	statements | something”
0.5	-14.7536	(−lt | lt
0.5	-14.7536	g0 | rds
0.5	-14.7536	o(t | f(t
0.5	-14.7536	empirical | ˜p(x,y)
0.5	-14.7536	(s\np)/np | >s\np
0.5	-14.7536	yes/no | issimilar
0.5	-14.7536	(≤40) | b&k2010
0.5	-14.7536	qj, | q
0.5	-14.7536	(tm) | 3,003
0.5	-14.7536	,−→r | (−→c
0.5	-14.7536	undergo | isolation.
0.5	-14.7536	72.1 | wpp+
0.5	-14.7536	propernsynntype | loc,
0.5	-14.7536	ˆs | argmaxs
0.5	-14.7536	methods | precision@10
0.5	-14.7536	ρn(wi|φn(wi−11 | γ(φn(wi−11
0.5	-14.7536	interest | highli
0.5	-14.7536	nº: | (mec
0.5	-14.7536	aprime, | s(aprime)
0.5	-14.7536	cross | nullnullnullnull)summationtext
0.5	-14.7536	i=1 | radicalvertex
0.5	-14.7536	incidence | currency
0.5	-14.7536	name | (dq):
0.5	-14.7536	12012.79 | pmi-factiva
0.5	-14.7536	6850 | assignment;
0.5	-14.7536	one-handed | players
0.5	-14.7536	apple | 16gb
0.5	-14.7536	47%, | frame
0.5	-14.7536	0.091 | 0.056
0.5	-14.7536	doc-pair | 0.1901
0.5	-14.7536	gs,t,a | anulla
0.5	-14.7536	18.4% | 39.5%
0.5	-14.7536	a46 | a51
0.5	-14.7536	radj | parenrightbig/parenleftbign
0.5	-14.7536	u↔you | lost↔won
0.5	-14.7536	0.879 | 0.734
0.5	-14.7536	0.4542 | 0.4578
0.5	-14.7536	tfi | idfi
0.5	-14.7536	“until | :context-rel
0.5	-14.7536	0.1612 | 0.6893
0.5	-14.7536	anger(pp) | smiley(pr)
0.5	-14.7536	vm/m-1 | 63.3
0.5	-14.7536	(china, | jiangsu}
0.5	-14.7536	lch | 84.90
0.5	-14.7536	vegetarians | glitters
0.5	-14.7536	8,926 | 7,546
0.5	-14.7536	ip0,8 | matched.
0.5	-14.7536	a105 | a114
0.5	-14.7536	{1}5k=1 | {1,2,2,2}
0.5	-14.7536	birmingham, | 35294,
0.5	-14.7536	 | out|rels
0.5	-14.7536	geo:city | geo:river
0.5	-14.7536	bone, | bones
0.5	-14.7536	vil | pou
0.5	-14.7536	f(a, | 0.8784
0.5	-14.7536	null1) | null(null1,
0.5	-14.7536	north | denton,
0.5	-14.7536	fm | |w
0.5	-14.7536	(85, | (98,
0.5	-14.7536	tft | et/summationtextt
0.5	-14.7536	x1,null5null: | null4,5null
0.5	-14.7536	df(4): | green
0.5	-14.7536	60/63 | 48/41
0.5	-14.7536	abstractive | extractive)
0.5	-14.7536	k] | ∀kp(y
0.5	-14.7536	’, | ’to’,
0.5	-14.7536	g1858g1867g1873g1866g1856 | g1871g1857g1866g1872g1861g1865g1857g1866g1872
0.5	-14.7536	not” | -3]
0.5	-14.7536	sublanguage | co¨occurrence
0.5	-14.7536	collaborative | (micr),
0.5	-14.7536	stringnet | (http://nav.stringnet.org),
0.5	-14.7536	art | statistik
0.5	-14.7536	capitalized† | ===
0.5	-14.7536	κ | pr(a)
0.5	-14.7536	hmeant | 0.3514
0.5	-14.7536	26 | studien
0.5	-14.7536	coppola, | coppola
0.5	-14.7536	polarity | (neutral).
0.5	-14.7536	carmax | price
0.5	-14.7536	domain | 3k
0.5	-14.7536	vd(fn) | −d
0.5	-14.7536	boxshadowdwn | [nenkova
0.5	-14.7536	agt(x, | [they]x
0.5	-14.7536	strict+fragments | unparsed
0.5	-14.7536	tables, | e,f
0.5	-14.7536	easier”, | “left:
0.5	-14.7536	instructions | “turn
0.5	-14.7536	0.954 | 0.955
0.5	-14.7536	u. | w911nf-10-1-0533,
0.5	-14.7536	r´eserver | [j’]
0.5	-14.7536	1-st | wt(v),t(h),t(sib),d
0.5	-14.7536	devices | needs-based
0.5	-14.7536	rob´otica, | lisboa,
0.5	-14.7536	a(i−1,j | sim(i,j
0.5	-14.7536	[adj1] | wildcard
0.5	-14.7536	10/26/89 | (e17)
0.5	-14.7536	white | (models
0.5	-14.7536	binary-valued | all-xb
0.5	-14.7536	bm-i2r• | comparisons/combo
0.5	-14.7536	(sr,lr) | conference
0.5	-14.7536	normal(µ | ,λ
0.5	-14.7536	c(j,r+1)(tj,tr+1) | scorejoint(x,ti,tr,tj,p
0.5	-14.7536	sr | (i,j,r)〉
0.5	-14.7536	np∗ | president
0.5	-14.7536	 | rxif
0.5	-14.7536	gdepo | 55.38
0.5	-14.7536	td,·+kα | ωl
0.5	-14.7536	ss+protein+gene | e1
0.5	-14.7536	cavity> | site>
0.5	-14.7536	knower | actor
0.5	-14.7536	activity | peer(s).
0.5	-14.7536	null6,7null: | x1
0.5	-14.7536	(nsummationtextni=1 | −(summationtextni=1
0.5	-14.7536	shouldn’t | (andrzejewski
0.5	-14.7536	18.19 | 12.94
0.5	-14.7536	n2 | x2’s
0.5	-14.7536	britannica | corpora”,
0.5	-14.7536	leash | 2β
0.5	-14.7536	tile-has-river=true | (box
0.5	-14.7536	excitation | speaker
0.5	-14.7536	size | ter↓
0.5	-14.7536	theme: | regulation(protein(positive
0.5	-14.7536	wordsflags:0:1:0 | wordsflags:1:1:0
0.5	-14.7536	holders | (a2
0.5	-14.7536	κ | pr(a)−pr(e)1−pr(e)
0.5	-14.7536	⎪ | ⎧
0.5	-14.7536	cousot, | 1978.
0.5	-14.7536	dg(1): | [moves
0.5	-14.7536	+bf | +subj
0.5	-14.7536	0.8459 | 0.5527
0.5	-14.7536	dar | (shrug),
0.5	-14.7536	compatriot | stimulate
0.5	-14.7536	ngd-google | wu-palmer
0.5	-14.7536	(vectorwt−1 | vectorwavg
0.5	-14.7536	(consensus) | 150
0.5	-14.7536	impress | build),
0.5	-14.7536	l∑ | n∑
0.5	-14.7536	*selma: | ñ:
0.5	-14.7536	bei | x⇒⟨jingfang
0.5	-14.7536	book-en | music-en,
0.5	-14.7536	α-β | ),,(w
0.5	-14.7536	d110 | d101d101
0.5	-14.7536	colour | refers
0.5	-14.7536	v−o | numeric
0.5	-14.7536	intra-rater | 1.0,
0.5	-14.7536	52.81 | 29.22
0.5	-14.7536	a161 | a162
0.5	-14.7536	null)]null | nullnullnullnull[null(nullnullnull)nullnullnull(nullnullnull)]
0.5	-14.7536	phil | spector,
0.5	-14.7536	(i−3 | identical,
0.5	-14.7536	a88a88a88a88 | [1,2]
0.5	-14.7536	∀w | xh,t
0.5	-14.7536	beleidigend | für
0.5	-14.7536	p(li|l1,··· | z(li−11
0.5	-14.7536	d80 | basket
0.5	-14.7536	〈arg0〉 | 〈mod〉
0.5	-14.7536	44.04 | 71.60
0.5	-14.7536	debate | censured
0.5	-14.7536	s48s46s49s53 | s105s110
0.5	-14.7536	e∈dn(f) | then-gramng
0.5	-14.7536	wt(h),t(sib),d | t(v),t(h),t(sib),d
0.5	-14.7536	p(li|l1,··· ,li−1,t)
0.5	-14.7536	hdn | pl
0.5	-14.7536	yif | table2:
0.5	-14.7536	hi,n[j,k] | ti−n+k
0.5	-14.7536	7880828486889092 | 6869707172
0.5	-14.7536	huang | tsing
0.5	-14.7536	y∈y(x) | f(y
0.5	-14.7536	hs04 | (tc)
0.5	-14.7536	equal(userusage[i+1],pattern[j+1]) | ~
0.5	-14.7536	rouge-n-p | i∈ct
0.5	-14.7536	w∈di | w∈dj
0.5	-14.7536	> | (n\n)
0.5	-14.7536	lexical-stylistic | lexical-stylistic
0.5	-14.7536	|t) | p(t|θ−i
0.5	-14.7536	df(3): | [moves
0.5	-14.7536	0.920 | 0.876∗
0.5	-14.7536	rti | π−deprti
0.5	-14.7536	voitures]::[cars], | bleues]::[blue
0.5	-14.7536	wmt | all-to-english
0.5	-14.7536	[pour | repeter
0.5	-14.7536	magloire | pou
0.5	-14.7536	gdepp | 56.15
0.5	-14.7536	l−1+b | z−l
0.5	-14.7536	null′ | 1nullα
0.5	-14.7536	detection | (dataset
0.5	-14.7536	f(u,w) | c(u,w)
0.5	-14.7536	final | (“wan
0.5	-14.7536	/> | <susceptibility
0.5	-14.7536	57.64 | (26)
0.5	-14.7536	rdt | crd+1t
0.5	-14.7536	states, | mappingq×w
0.5	-14.7536	∂f⟨2,x⟩∂¯h | ∂¯h1
0.5	-14.7536	alphabetical | @,
0.5	-14.7536	[a], | connotation
0.5	-14.7536	20.00 | 14.21
0.5	-14.7536	strictp1 | p10
0.5	-14.7536	interchangeability | system/human
0.5	-14.7536	tough | to-vp
0.5	-14.7536	maximalnp | arg
0.5	-14.7536	_钱 | _的
0.5	-14.7536	dout(u) | din(v)
0.5	-14.7536	n(t | b(t
0.5	-14.7536	14∗ | corr(me-seq)
0.5	-14.7536	234 | 402
0.5	-14.7536	“like | prayer”
0.5	-14.7536	+(1nullλ) | bullet
0.5	-14.7536	1"234563) | ;3%"44)4822)
0.5	-14.7536	(sec) | (l1
0.5	-14.7536	0.0114 | 0.701
0.5	-14.7536	σ | 2.96
0.5	-14.7536	 | (verb,
0.5	-14.7536	cu | -tr
0.5	-14.7536	ghar | barii
0.5	-14.7536	2001 | -�
0.5	-14.7536	readings | ’i|he
0.5	-14.7536	xi→ys | j∈ci
0.5	-14.7536	do | untilg(ˆe)negationslash>g(ecur)
0.5	-14.7536	attrs | 2.74
0.5	-14.7536	nullnullnullnull[nullnullnullnullnull(nullnullnullnullnull1nullnull2) | nullnullnull[nullnull(nullnullnull)nullate’(null,
0.5	-14.7536	nullnull1nullnull2nullnull3null | null1,
0.5	-14.7536	(class | sim(wn,wp)
0.5	-14.7536	1cs, | lgp
0.5	-14.7536	e1/epsilon1 | c01
0.5	-14.7536	sp← | a∗
0.5	-14.7536	labelling: | null(nullnull(null))
0.5	-14.7536	[3,4) | ][
0.5	-14.7536	terain | cel
0.5	-14.7536	rubber | rubber,
0.5	-14.7536	−1) | a(i−2,j
0.5	-14.7536	x3, | x1
0.5	-14.7536	s: | isla
0.5	-14.7536	↓ | (dp1,x),(dp2,y)
0.5	-14.7536	{--, | "
0.5	-14.7536	commission | d´efendant
0.5	-14.7536	demethylation | residue:state:ubiquitinated
0.5	-14.7536	157k | 132
0.5	-14.7536	cambridge, | 02139,
0.5	-14.7536	ruk | ek
0.5	-14.7536	eprime∈eh | αn·gn(eprime)
0.5	-14.7536	play | role;
0.5	-14.7536	aims | central,
0.5	-14.7536	1105 | hv
0.5	-14.7536	/nq | /pa
0.172414	-14.7537	pr | •
0.166667	-14.7537	residue:state:∅ | regulation
0.025641	-14.7538	the | high
0.0337602	-14.7538	two | on
0.0961538	-14.7539	from | shared
0.0734463	-14.754	we | polarity
0.142857	-14.7541	de | .
0.2	-14.7543	states | sees
0.25	-14.7545	nullnull) | add(null,
0.333333	-14.7547	* | conj
0.333333	-14.7547	(hassan | 2009),
0.333333	-14.7547	r3 | held
0.333333	-14.7547	ago: | au4_5sec
0.333333	-14.7547	0.32 | 0.27
0.104651	-14.7549	at | new
0.2	-14.7549	> | co
0.0512129	-14.755	for | rules
0.4	-14.755	methylation | deubiquitination
0.15	-14.7551	from | it.
0.0280674	-14.7551	if | is | a
0.153846	-14.7552	you | from
0.5	-14.7552	p(c|x)
0.5	-14.7552	programs | (180
0.5	-14.7552	f|fop
0.5	-14.7552	p(e)·pθ(f|e)
0.5	-14.7552	|rt)
0.5	-14.7552	biotechnology | bethesda,
0.5	-14.7552	p(k|wi,d)
0.5	-14.7552	p(c|e),
0.0734463	-14.7557	assume | for
0.162162	-14.7558	latent | semantic
0.333333	-14.7558	tweet | tweet,
0.333333	-14.7558	0.35 | 0.80
0.333333	-14.7558	23 | ast
0.333333	-14.7558	mj | 0,
0.0279553	-14.7558	the | small
0.0584192	-14.7562	= | set of
0.0970874	-14.7562	shows | more
0.011927	-14.7563	we | a | a
0.0155491	-14.7565	the | feature | the
0.0852713	-14.7565	this | significantly
0.0506667	-14.7567	and | instances
0.0673077	-14.7568	two | language
0.0509383	-14.7569	a | local
0.5	-14.7569	p(t→h|l
0.5	-14.7569	|v|}.
0.5	-14.7569	lambert, | sadaf
0.5	-14.7569	rij, | rijn,
0.5	-14.7569	lagrangian | relaxation.
0.5	-14.7569	differentiating | inter-document
0.115942	-14.757	accuracy | system
0.333333	-14.757	b`ush´ı | hu`ıt´an
0.333333	-14.757	|r| | t
0.333333	-14.757	candidates | candidates)
0.333333	-14.757	known | alexander
0.0638298	-14.7572	8 | in
0.4	-14.7576	multiannual | through
0.0990099	-14.7576	following | null
0.3	-14.7577	pattern | then
0.0734463	-14.7577	own | and
0.157895	-14.7578	section | lf
0.101124	-14.7579	set | gold
0.333333	-14.7581	instructions | “go
0.0866142	-14.7582	that | four
0.0380673	-14.7582	shown | a
0.111111	-14.7584	language | many
0.0742857	-14.7584	from | three
0.125	-14.7586	are | authors | or
0.5	-14.7586	pr(t|w)pr(w|d),
0.5	-14.7586	|ωk|
0.5	-14.7586	orally-oriented | (narrative
0.5	-14.7586	p(zq|q)
0.5	-14.7586	consequences | (baas)
0.0760234	-14.7587	in | authors
0.0166104	-14.7587	of | model
0.123077	-14.7587	this, | is
0.172414	-14.7589	data | consisting
0.121212	-14.7589	in | node.
0.0309777	-14.759	most | a
0.4	-14.759	p(¯a|¯q)
0.0280224	-14.7591	x | the
0.166667	-14.7592	parts: | for
0.210526	-14.7592	using | extraction.
0.0774194	-14.7593	is | although
0.0392465	-14.7594	a | parallel
0.0498688	-14.7595	by | other
0.0576271	-14.7596	in | opinion
0.101124	-14.7596	then | feature
0.3	-14.7597	phrases | links
0.0952381	-14.7597	items | for
0.166667	-14.7598	quite | it
0.088	-14.7598	we used | by
0.210526	-14.7599	with | stamatatos,
0.0696517	-14.7599	results | at
0.0934579	-14.7601	or | new
0.0669856	-14.7603	we | chinese
0.222222	-14.7603	generation | al.
0.4	-14.7603	kenya, greece, | argentina,
0.333333	-14.7603	2011 | diseases
0.333333	-14.7603	word | markable:
0.5	-14.7603	englishside | chinese-side
0.5	-14.7603	undoubtedly | sentiment/opinion
0.0514905	-14.7606	for | many
0.107143	-14.7608	contexts | as
0.0627615	-14.7608	a | graph | and
0.0718232	-14.7608	show | model
0.036193	-14.7611	documents | of
0.0429104	-14.7611	get | the
0.0529412	-14.7614	data | using
0.333333	-14.7614	codes | next
0.119403	-14.7615	have | without
0.5	-14.7619	nb | o(r
0.5	-14.7619	erol | hee
0.0465632	-14.7621	in | either
0.142857	-14.7622	is | constraints.
0.0562914	-14.7624	and | full
0.333333	-14.7625	thesaurus | elements
0.333333	-14.7625	through | etc.).
0.333333	-14.7625	ep(γ|w) | fla2b3,b2,b1(γ)+fra2b3,b2,b1(γ)
0.333333	-14.7625	4w3t | 3w4t
0.111111	-14.7625	each | .
0.210526	-14.7627	between | senses.
0.0170288	-14.7628	the number of | the
0.272727	-14.7628	combinations | noun
0.25	-14.7628	dialogues | been
0.0255034	-14.763	a | the same
0.0291519	-14.7633	can | on
0.0138321	-14.7636	of | word
0.5	-14.7636	〈num〉sg) | (〈det〉the)?
0.333333	-14.7637	policy | filter
0.0293333	-14.7639	parallel | the
0.0297834	-14.764	a | of | all
0.0149153	-14.7641	have | and
0.0635593	-14.7641	2011 | in
0.0510753	-14.7644	if | than
0.0243309	-14.7644	the | search
0.235294	-14.7646	those | part
0.030888	-14.7647	= | 1
0.166667	-14.7648	dda | –
0.132075	-14.7653	on | demonstrate
0.5	-14.7653	code | trucks
0.5	-14.7653	bush” | “george
0.25	-14.7654	addition, | specific
0.25	-14.7654	÷ | ÷
0.104651	-14.7654	models | corpus
0.0376266	-14.7655	on | different
0.0805369	-14.7657	was | test
0.0323276	-14.7658	order | in
0.333333	-14.7659	selection | (p
0.0505319	-14.766	case, | of
0.0686275	-14.7662	that | state
0.210526	-14.7663	probabilities | then
0.137255	-14.7664	find | where
0.119403	-14.7667	first | algorithm
0.333333	-14.767	done | project | research
0.235294	-14.767	2.9 | table
0.5	-14.767	omorfi | listenmaa,
0.5	-14.767	5.3, | same.
0.0608365	-14.7672	of | created
0.0866142	-14.7675	so | have
0.0630252	-14.7676	and | i.e.
0.0532544	-14.768	our | when
0.142857	-14.768	3 | section,
0.0682927	-14.7681	types | that
0.4	-14.7684	bc | mz
0.0900901	-14.7687	than | less
0.0429688	-14.7688	to | form
0.162162	-14.7689	θ | that
0.0212963	-14.7689	annotation | the
0.0652174	-14.7689	r | for
0.105882	-14.769	target | given
0.333333	-14.7693	observe | improved
0.105882	-14.7694	systems | based
0.25	-14.7696	widely | 2007;
0.109589	-14.7696	gold | not
0.102273	-14.7699	show | method
0.210526	-14.7699	ran | set
0.0551948	-14.77	in | 6
0.0551948	-14.7702	the | of | five
0.0265805	-14.7703	we | the | system
0.333333	-14.7704	yd,s,n | wd,s,n
0.2	-14.7704	inference | relation
0.5	-14.7704	|*ar
0.5	-14.7704	david | uscinformation
0.0805369	-14.7706	abstract | paper | a
0.0431373	-14.7707	and | vector
0.166667	-14.771	l | supertagger
0.333333	-14.7716	ways: | sentences
0.333333	-14.7716	λ2 | =
0.0337209	-14.772	the | approaches
0.5	-14.7721	h¨angen | k¨uchenwand.
0.5	-14.7721	tin* | work*
0.5	-14.7721	(weigand | klalow,
0.109589	-14.7722	define | word
0.0542169	-14.7729	table 2 | of
0.178571	-14.7729	order | state
0.047619	-14.773	of | r
0.122807	-14.773	our | makes | the
0.0540541	-14.7733	corpus, | to
0.0652174	-14.7734	and | wordnet
0.125	-14.7734	concept | are
0.088	-14.7735	1 | over
0.0485437	-14.7736	predict | the
0.5	-14.7738	γ | φ(mi,ti)
0.5	-14.7738	ewm | ccc
0.102273	-14.774	t. | of
0.0816327	-14.7742	factor | is
0.117647	-14.7743	scene | the
0.192308	-14.7744	5 | problem
0.0859375	-14.7744	compound | of
0.0546624	-14.7745	is | annotated
0.178571	-14.7747	there | problems
0.103448	-14.7749	if | label
0.0718232	-14.7749	is | determine
0.0617284	-14.775	data | all
0.117647	-14.7751	furthermore, | not
0.178571	-14.7753	in | queue
0.162162	-14.7755	data | 2006).
0.0952381	-14.7755	therefore, we | and
0.5	-14.7755	thereis | phrasewich
0.5	-14.7755	su | y(x)
0.146341	-14.7758	words | low
0.0523256	-14.7761	space | a
0.333333	-14.7761	predicted | 52
0.333333	-14.7761	certain | tokens
0.122807	-14.7763	such | human
0.153846	-14.7764	given | associated
0.0310933	-14.7766	this | have
0.1	-14.7766	5 | 25
0.4	-14.7766	indicate | people
0.0952381	-14.7768	c | b
0.122807	-14.7771	corpora | using
0.5	-14.7772	|etnull1)
0.030969	-14.7774	on | using
0.0887097	-14.7775	but | results
0.222222	-14.7776	comm | specialists
0.139535	-14.7781	these | impact
0.0934579	-14.7785	each | corpus.
0.132075	-14.7788	phrase | bilingual
0.0550162	-14.7788	joint | to
0.0557377	-14.7789	based | an
0.5	-14.7789	army | subcontract
0.5	-14.7789	(6) | mineditedsug=pattern
0.5	-14.7789	8am | 10am”).
0.192308	-14.779	nullnullnullnullnullnullnull | null
0.25	-14.779	during | generate
0.235294	-14.779	task | definition
0.019425	-14.7793	of | translation
0.15	-14.7794	compared | at
0.139535	-14.7795	case | null
0.166667	-14.7796	• | →
0.0676329	-14.7799	in | seen
0.0192086	-14.78	the | so
0.0660377	-14.7801	majority | of
0.192308	-14.7803	(2008), | et
0.172414	-14.7805	: | ...
0.5	-14.7806	karin | chute.
0.5	-14.7806	cows | eaten.
0.5	-14.7806	dates | time-spans
0.5	-14.7806	uthus | postdoctoral
0.0594796	-14.7807	the | wrong
0.0759494	-14.7808	two | null | null
0.166667	-14.7813	condition | =
0.103448	-14.7816	from | automatically
0.0779221	-14.782	for example, | =
0.0308458	-14.7821	a | which | is
0.0859375	-14.7823	our | quality
0.0726257	-14.7823	co-occurrence | of
0.5	-14.7823	treder | blankertz,
0.5	-14.7823	multi-labeling | refer
0.5	-14.7823	pr(p|s)
0.25	-14.7824	word | argument | argument
0.0502646	-14.7825	used | model
0.178571	-14.7826	posterior | over
0.0544872	-14.7827	in | n-gram
0.210526	-14.7828	testing | were
0.0852713	-14.7831	these | syntactic
0.0383436	-14.7832	event | a
0.192308	-14.7836	† | †
0.0759494	-14.7837	that | measures
0.0646552	-14.7838	here | is
0.178571	-14.7839	f-measure | precision
0.235294	-14.7839	every | words.
0.5	-14.784	generation | exact-phrase
0.333333	-14.7841	probabilistic | rather
0.333333	-14.7841	used | transform
0.111111	-14.7841	approach | lexical
0.0663507	-14.7843	the | fig.
0.127273	-14.7843	wish | in
0.122807	-14.7843	first | syntactic
0.121212	-14.7844	:: | ::
0.0156289	-14.7844	section | the
0.0588235	-14.7847	results | training
0.122807	-14.7847	we | synchronous
0.0559211	-14.7848	a | of | generated
0.0586081	-14.7849	not | when
0.0820896	-14.7849	over | word
0.3	-14.7851	phrases | generate
0.127273	-14.7851	having | on
0.222222	-14.7853	notice | using
0.027195	-14.7853	the | linguistic
0.0437376	-14.7853	of | might
0.0471698	-14.7854	from | but
0.0529412	-14.7855	comparing | of
0.172414	-14.7858	models | 2006).
0.0943396	-14.7859	word | while
0.0989011	-14.7859	studied | et
0.2	-14.7861	de-en | fr-en
0.0373134	-14.7864	syntactic | in
0.0537975	-14.7865	use | language
0.0900901	-14.7866	system | data.
0.25	-14.7867	linear | svm
0.1	-14.7869	often | from
0.0107219	-14.787	the | by | and
0.0597015	-14.787	a | edge
0.142857	-14.787	that | mutual
0.0299345	-14.787	shown | and
0.109589	-14.7872	can | translated
0.132075	-14.7874	pairs | similarity
0.0305118	-14.7875	as | model
0.0769231	-14.7877	system | performance
0.0273865	-14.7884	features | for
0.0207394	-14.7884	english | the
0.0240529	-14.7885	types | the
0.139535	-14.7886	international | resources
0.119403	-14.7887	supported | project
0.0794702	-14.7887	2 | language
0.0612245	-14.7887	the best | with
0.333333	-14.7887	h1: | in
0.5	-14.7892	micro-avg | 0.583
0.5	-14.7892	68 | 96
0.5	-14.7892	news | #size
0.0293309	-14.7896	is | information
0.272727	-14.7897	resources | bionlp
0.153846	-14.7897	to | distinguishing
0.0573477	-14.7899	examples | for
0.4	-14.7903	reading | ma
0.0136799	-14.7905	of the | for
0.5	-14.7909	launching | desert.
0.0150502	-14.7912	both | the
0.0385233	-14.7914	and | –
0.108108	-14.7914	a | correspondence
0.16129	-14.7914	0 0.2 | 0.6
0.139535	-14.7915	table | numbers
0.0253333	-14.7916	these | for
0.0808824	-14.7919	on | 2007;
0.25	-14.7919	number | bad
0.142857	-14.792	vbd | pp
0.0687831	-14.7921	because | it
0.333333	-14.7921	’ | k
0.0641026	-14.7924	(3) | a
0.0722222	-14.7924	we | rate
0.0622407	-14.7925	cluster | a
0.272727	-14.7926	conversations, | a
0.0544872	-14.7927	few | and
0.185185	-14.7927	data: | in
0.5	-14.7927	2000), | clarke
0.5	-14.7927	o(n | o(n+e)
0.5	-14.7927	0.280 | 0.307
0.166667	-14.7928	future | investigate
0.0526316	-14.7928	which | language
0.0234152	-14.7929	most | and
0.101124	-14.793	variants | a
0.178571	-14.7931	performance | machine
0.0654206	-14.7931	to | presented
0.0779221	-14.7931	another | be
0.16129	-14.7936	classifier, | as
0.210526	-14.7937	2004; | 2010;
0.0551948	-14.7939	entities | to
0.117647	-14.7941	slightly | with
0.0607287	-14.7943	nn | the
0.5	-14.7944	unsel | bnews-ace-2
0.5	-14.7944	projected | [-3,
0.0952381	-14.7948	i | it
0.0543131	-14.7949	e | i | s
0.0317797	-14.7951	research | of
0.0691489	-14.7952	corpus | this
0.0390244	-14.7954	as | translation
0.3	-14.7955	 | sentences)
0.0170628	-14.7956	to | our
0.139535	-14.7959	entropy | model
0.0794702	-14.7961	word | most
0.5	-14.7961	62.4 | −best
0.119403	-14.7962	features | related
0.075	-14.7966	in | indicate
0.153846	-14.7966	and | 2002; | 2009).
0.333333	-14.7968	large | distributed
0.333333	-14.7968	hotel | london}
0.0559211	-14.7968	assume | is
0.0702703	-14.797	is | d
0.0335731	-14.7971	and | do
0.0241486	-14.7972	we | = | =
0.0484694	-14.7973	dataset | the | of
0.0764331	-14.7973	also | one
0.0592593	-14.7975	as | similarity
0.222222	-14.7976	name | then
0.0734463	-14.7976	to be | more
0.056338	-14.7978	novel | to
0.5	-14.7979	“down” | s|j|
0.5	-14.7979	(bezdek | hathaway,
0.0588235	-14.7981	and | 3.
0.222222	-14.7983	training | unseen
0.12963	-14.7984	have | 2003;
0.166667	-14.7985	wsj | table
0.153846	-14.7987	development | results
0.0592593	-14.7987	for | do not
0.0349288	-14.7988	such | be
0.0367232	-14.7988	of | computational | june
0.0355677	-14.7989	the | by | all
0.103448	-14.7991	defined | et
0.5	-14.7996	variations | sentimentality
0.127273	-14.7997	(2010) | –
0.0197694	-14.7999	set | in
0.333333	-14.8002	amounts | 2010;
0.139535	-14.8002	not | official
0.210526	-14.8003	senses | corpus.
0.0534591	-14.8006	the | carried
0.104651	-14.801	(or | for
0.272727	-14.8011	s | ,s
0.105882	-14.8011	2009) | on
0.5	-14.8014	reverse | “source”,
0.5	-14.8014	(enright | sheehan,
0.5	-14.8014	word/phrase | tags,
0.00662436	-14.8017	this | the | the
0.172414	-14.8019	studies | can
0.0312826	-14.8022	based on | to
0.112676	-14.8023	pattern | by
0.0695187	-14.8024	not | results
0.0330189	-14.8024	how | in
0.3	-14.8028	hereafter, | as
0.0335329	-14.8028	use | with
0.0410714	-14.803	to | terms
0.0504202	-14.8031	lower | of
0.5	-14.8031	l01 | l03
0.5	-14.8031	documents | rbleu
0.5	-14.8031	antti-veikko | spyros
0.0387722	-14.8032	bleu | and
0.166667	-14.8032	give | number
0.139535	-14.8032	at | original
0.0583942	-14.8036	which | not | to
0.12963	-14.8038	model | trees
0.0852713	-14.8038	model | topic
0.132075	-14.8039	use | data,
0.0636364	-14.8042	relation | with
0.4	-14.8042	11th | conference
0.109589	-14.8044	top | which
0.0330189	-14.8045	of | extraction
0.0550162	-14.8051	the | linguistics
0.0526316	-14.8051	and | science
0.0345711	-14.8052	the | segmentation
0.122807	-14.8052	(2) | this
0.0651163	-14.8053	of | application
0.0412186	-14.8054	in | without
0.166667	-14.8055	could | because
0.0736196	-14.8057	these | where
0.235294	-14.8059	could | instead
0.0827068	-14.8064	on | could
0.0232693	-14.8064	and | semantic
0.5	-14.8066	opinie | vlag
0.5	-14.8066	gibbs | mean-field
0.5	-14.8066	so, | ssc)
0.5	-14.8066	hala | withothers
0.146341	-14.8067	see | these
0.0502793	-14.8071	in | associated
0.0877193	-14.8071	term | by
0.0150962	-14.8074	the | when
0.0539683	-14.8076	focused | and
0.25	-14.8076	action | participant
0.0784314	-14.8076	do | all
0.157895	-14.8077	π( | π(
0.157895	-14.8077	w=0.1 | w=0.1
0.185185	-14.8082	has | improved
0.106667	-14.8082	we | wmt
0.0275081	-14.8083	a | these
0.0226145	-14.8084	features | a
0.0471698	-14.8084	the | of | prior
0.5	-14.8084	confusion | 96.74
0.5	-14.8084	gpeemploys | hasbirthdate
0.5	-14.8084	(c320) | source-sided
0.5	-14.8084	right, | dramatically
0.080292	-14.8088	also | based
0.108108	-14.8088	) | by
0.0627615	-14.8091	as | those
0.111111	-14.8092	on | own
0.0512821	-14.8095	node | in
0.0472813	-14.8095	phrase | with
0.15	-14.8096	to | merge
0.132075	-14.81	r1 | r1
0.166667	-14.8102	perform | other
0.0978261	-14.8102	context | used
0.25	-14.8102	computing | average
0.0892857	-14.8102	x | that
0.5	-14.8102	hamas | negotiations
0.5	-14.8102	emily | bender,
0.5	-14.8102	ldc2002e18, | ldc2004t08
0.5	-14.8102	p(si|si−1)
0.0490956	-14.8103	allows | a
0.0434783	-14.8103	are | as | are
0.0333333	-14.8109	we | given
0.0505618	-14.811	with | both
0.0639269	-14.811	we | named
0.0296367	-14.8112	of | association
0.075	-14.8113	a | 2005).
0.057554	-14.8113	in | extract
0.142857	-14.8114	than | ones
0.0909091	-14.8116	case | it
0.040493	-14.8116	some | be
0.0198564	-14.8117	used | in
0.16129	-14.8117	empirical | processing
0.0774194	-14.8118	refers | to
0.0414313	-14.8119	on | system
0.2	-14.8119	can | maximizing
0.5	-14.8119	logη | recall/precision/f1
0.5	-14.8119	sla | (iwashita,
0.0319032	-14.8119	corpus | for
0.0225399	-14.812	1 | 1 | 0
0.333333	-14.812	set | 100,
0.0449438	-14.8122	4.1 | of
0.0957447	-14.8123	in | above.
0.111111	-14.8123	that | along
0.057971	-14.8125	development | in
0.0456621	-14.8127	in | compared
0.0698925	-14.8127	3 | data
0.0346154	-14.8129	pairs | a
0.0212972	-14.813	with | are
0.0323661	-14.813	this | at
0.333333	-14.8131	recall | low.
0.0698925	-14.8132	most | used
0.0617284	-14.8136	paper | an
0.0468384	-14.8137	close | the
0.132075	-14.8137	than | still
0.5	-14.8137	population | isear
0.0862069	-14.814	recognition | for
0.0764331	-14.814	set | if
0.235294	-14.8142	env(q | {q
0.235294	-14.8142	biochemicalreaction | catabolism
0.0285205	-14.8143	for | features
0.0429448	-14.8143	and | under
0.0651163	-14.8145	many | on
0.0196982	-14.8145	and | words
0.16129	-14.8151	texts | english
0.0471464	-14.8154	our | such
0.0917431	-14.8155	shows | system
0.5	-14.8155	estimated | worldwide,
0.127273	-14.8158	without | training
0.0978261	-14.8161	both | was
0.0176667	-14.8162	of | their
0.2	-14.8162	nato | "
0.0398614	-14.8164	increase | the
0.115942	-14.8165	sentences | first
0.0797101	-14.8166	development | with
0.032596	-14.8167	time | of
0.0431211	-14.8168	the | of | vectors
0.115942	-14.8169	6 | paper
0.4	-14.8169	i−1summationdisplay | k=i+1
0.00915784	-14.8171	the | each
0.127273	-14.8172	effective | et
0.5	-14.8172	ungulates | self-propelled
0.5	-14.8172	0.352 | 0.336
0.0283993	-14.8175	original | the
0.0736196	-14.8179	she | the
0.12069	-14.8183	we | valuable
0.272727	-14.8184	due | short
0.0779221	-14.8185	class | are
0.0279188	-14.8189	as | this
0.142857	-14.819	all | achieved
0.5	-14.819	2.43 | 2.42
0.5	-14.819	nus | nist’08
0.0185391	-14.8193	these | the | and
0.12069	-14.8196	systems | information
0.0718232	-14.8196	into | model
0.0625	-14.8197	although | we
0.146341	-14.8197	is | write
0.0854701	-14.8199	can | much
0.0780142	-14.8202	• | other
0.0978261	-14.8203	an | content
0.15	-14.8203	from | summary
0.0232288	-14.8207	the | form | the
0.0402098	-14.8208	observe | the
0.5	-14.8208	wordnet.pt | lisbon
0.5	-14.8208	53 | 68
0.5	-14.8208	p(w|tk)
0.00853296	-14.8212	it | the
0.0445434	-14.8214	also | not
0.0633484	-14.8215	approach | this
0.3	-14.8219	75 | comps
0.172414	-14.822	additional | based
0.109589	-14.8226	result | than
0.5	-14.8226	just | arcstandard,
0.5	-14.8226	cr(cij|si).
0.5	-14.8226	malayalam | telugu
0.5	-14.8226	(gaucha | (adomavicius
0.5	-14.8226	car | causal
0.166667	-14.8227	formal | theory
0.16129	-14.8231	approach | important
0.0485934	-14.8232	is | rule
0.136364	-14.8232	have | arguments
0.222222	-14.8234	if | 15:
0.0193004	-14.8236	we | the | model
0.0269841	-14.8237	it | on
0.136364	-14.8237	who | be
0.0957447	-14.8238	might | an
0.09375	-14.8238	first | english
0.111111	-14.8238	errors | from
0.2	-14.824	tune | data
0.166667	-14.8244	human | metrics
0.0231884	-14.8244	work | in
0.5	-14.8244	produces | (“very
0.0780142	-14.8246	more | will
0.142857	-14.8247	of | expressions:
0.0797101	-14.8249	two | then
0.0195426	-14.825	in | all
0.12963	-14.825	overview | our
0.0393836	-14.8251	by | more
0.045045	-14.8252	data, | to
0.0577617	-14.8252	not | its
0.0315904	-14.8253	we | task
0.0989011	-14.8255	yields | for
0.0543131	-14.8258	computational | language
0.0366032	-14.8258	language | on
0.142857	-14.8262	using | sentence.
0.333333	-14.8262	=f(w | [x
0.5	-14.8262	null1 | null100
0.5	-14.8262	f5 | f7
0.5	-14.8262	wfsa | (bar-hillel
0.5	-14.8262	91.28 | 92.05
0.5	-14.8262	hyderabad, | 500032
0.5	-14.8262	45.8 | 48.0
0.5	-14.8262	0 | jwj,
0.0124036	-14.8263	the | but
0.0490463	-14.8263	two | word
0.146341	-14.8264	several | into
0.166667	-14.8268	simply | this
0.0808824	-14.8269	is | (in
0.0702703	-14.827	that | s
0.0537975	-14.827	text | be
0.0947368	-14.8273	and | portion
0.333333	-14.8274	candidates | question
0.5	-14.8279	-e | 7p
0.5	-14.8279	yield | ρ(wi)
0.5	-14.8279	punctuation | type-changing
0.0740741	-14.8282	phrase-based | et
0.133333	-14.8283	example | dependency
0.0443459	-14.8285	to | bleu
0.0571429	-14.8286	- | for
0.00973062	-14.8288	we | the | for
0.0827068	-14.8293	key | for
0.0241884	-14.8296	alignment | the
0.013282	-14.8296	which | to
0.0754717	-14.8297	the | table,
0.5	-14.8297	c) | figure4:
0.5	-14.8297	tran | s
0.166667	-14.8298	model | functions
0.4	-14.8298	sandeep | sadeep
0.0541401	-14.83	to | ranking
0.0453515	-14.8301	with | were
0.101124	-14.8305	work | presented
0.15	-14.8305	(which | be
0.0663265	-14.8306	1 | 1:
0.185185	-14.8306	0.96 | 0.86
0.0785714	-14.8306	relationships | and
0.116667	-14.831	discourse | such
0.052795	-14.831	order | be
0.0158081	-14.8311	order | the
0.4	-14.8312	accessing | data
0.0900901	-14.8313	x | from
0.0731707	-14.8315	to | turn
0.5	-14.8315	northbridge | pessoas[people].
0.5	-14.8315	+coref | orgemploys
0.5	-14.8315	queue | expand(candidate,
0.0677083	-14.8317	this | parallel
0.014911	-14.8319	the | both
0.09375	-14.8319	translation | test
0.0419847	-14.8319	the | percentage
0.0978261	-14.8322	show | models
0.0909091	-14.8327	query | or
0.0308511	-14.8328	a | s
0.0754717	-14.8329	is | mention
0.0464548	-14.833	vector | in
0.0625	-14.833	the | approximation
0.0186525	-14.8331	our | we | the
0.0515152	-14.8331	data, | in
0.028891	-14.8331	the | portland, | 2011.
0.037325	-14.8333	in | set.
0.5	-14.8333	intercept | regression;
0.5	-14.8333	america | iran.
0.5	-14.8333	npa | npf
0.047619	-14.8334	are not | and
0.12069	-14.8334	process | this
0.153846	-14.8337	506f7369746976655f726567756c6174696f6e | 5353
0.0639269	-14.8345	table | when
0.230769	-14.8345	vp | nx
0.230769	-14.8345	size | evaluate
0.0684211	-14.8348	various | as
0.059761	-14.8349	individual | to
0.25	-14.8352	intuition | may
0.0177232	-14.8353	where | and
0.0722892	-14.8357	with | parse
0.2	-14.8362	1, | 0
0.0275689	-14.8369	the | level | the
0.5	-14.837	ndings, | afrl,
0.5	-14.837	catching | drift:
0.0609756	-14.8373	comparable | in
0.0677083	-14.8375	set, | a
0.0452489	-14.8376	relatively | the
0.2	-14.8377	these | section.
0.15625	-14.8378	generally | al.,
0.25	-14.8379	project | foundation
0.333333	-14.8382	values | matches
0.0967742	-14.8384	compute | all
0.0493151	-14.8385	is | sense
0.5	-14.8388	click | “shark
0.0346667	-14.8389	language | as
0.133333	-14.8389	then | table
0.0523077	-14.8392	= | k
0.210526	-14.8393	require | between
0.178571	-14.8394	the | iarg0
0.25	-14.8397	ideas | this
0.105263	-14.8401	if | present
0.222222	-14.8403	is | severe
0.230769	-14.8403	mt | wpred
0.15625	-14.8406	are | themselves
0.0791367	-14.8407	we | stanford
0.0656566	-14.8407	we | improvements
0.210526	-14.8408	in | developmental
0.0405157	-14.8412	table 2: | the
0.0284444	-14.8413	that | features
0.0519878	-14.8413	in section | our
0.102564	-14.8423	all | described
0.5	-14.8424	commercial | bing,
0.5	-14.8424	hdlr | matrix.
0.5	-14.8424	fully | lady”
0.5	-14.8424	40th | (acl
0.178571	-14.8426	5 conclusions | future
0.133333	-14.8428	in | detail.
0.0279965	-14.8429	the | combination
0.190476	-14.8432	(2) | (1)
0.0361795	-14.8433	proceedings | portland,
0.222222	-14.8435	features: | 0
0.12069	-14.8435	relations | using
0.3	-14.8436	hmeant | 0.4324
0.0267296	-14.8436	= | where
0.101124	-14.8437	c | if
0.0303644	-14.8438	to | lexical
0.0833333	-14.844	words | most
0.0918367	-14.8441	comparable | with
0.4	-14.8443	mproductdisplay | p(wi|wi−11
0.080292	-14.8445	to | easily
0.230769	-14.8445	with | petrov
0.210526	-14.8446	zhou | al.
0.0978261	-14.845	we | comments.
0.272727	-14.845	is | weighted sum
0.0442478	-14.8451	from | one
0.122807	-14.8452	makes | an
0.0797101	-14.8456	method, | to
0.2	-14.8457	also | domain.
0.0214395	-14.8458	such as | the
0.0574713	-14.8459	a | comparable
0.0208132	-14.846	function | the
0.5	-14.8461	ast | exh
0.210526	-14.8462	sciences | marina
0.102564	-14.8464	document | using
0.0670103	-14.8465	prediction | in
0.101266	-14.8468	seem | be
0.0519878	-14.8468	contains | to
0.0465686	-14.8469	of | assigned
0.0386555	-14.847	= | 0
0.1	-14.8472	between | relation
0.4	-14.8472	these | evaluations,
0.0701754	-14.8475	text | an
0.0585938	-14.8475	that | level
0.122807	-14.8479	2 | b
0.5	-14.8479	say | round”,
0.5	-14.8479	lex(vi,v | j)=
0.5	-14.8479	tensor | giesbrecht,
0.5	-14.8479	cu-zeman | dcu-combo
0.5	-14.8479	74.7 | 70.5
0.102564	-14.8483	that | addition
0.0498615	-14.8483	assigned | of
0.048	-14.8484	classes | of
0.127273	-14.8484	more | topic
0.210526	-14.8485	end | 12:
0.0909091	-14.8489	♣ | ①
0.039783	-14.849	in | we have
0.333333	-14.8491	children | communication
0.2	-14.8494	</c> | <c>
0.181818	-14.8496	dragged | toys
0.0419162	-14.8498	our | section
0.272727	-14.85	wpred | t
0.133333	-14.8501	indicate | more
0.116667	-14.8505	an | it.
0.0377358	-14.8507	parse | of
0.0245566	-14.8508	on | which
0.111111	-14.8512	our | entities
0.0507463	-14.8516	class | is
0.0526316	-14.8516	table | accuracy
0.5	-14.8516	[0,1]n |
0.5	-14.8516	bon−1(wi−11 | wi−1i−n+2.
0.0608696	-14.8519	and | instead
0.0780142	-14.8521	noted | of
0.0769231	-14.8524	a | sampling
0.25	-14.8525	which | harder
0.102564	-14.8528	distinguishing | of
0.044843	-14.853	shown | for
0.0877193	-14.8531	5.3 | and
0.5	-14.8534	harrison, | runs),
0.16129	-14.8536	were | articles
0.15625	-14.8538	propagation | on
0.230769	-14.8538	manually | negative
0.112676	-14.8541	vector | where
0.25	-14.8544	average | dataset
0.0884956	-14.8544	procedure | is
0.0192229	-14.8545	only | the | of
0.0714286	-14.8545	np | for
0.142857	-14.8548	p2 | of
0.5	-14.8552	grammarchecking(t,patterngrammarbank) | userproposedusages=extractusage(sentence)
0.5	-14.8552	donner | ‘give
0.0701754	-14.8555	was | more
0.0566038	-14.8555	e | is
0.0840336	-14.8555	on | social
0.3	-14.8557	2, | 1.
0.0542373	-14.8558	removed | the
0.0536913	-14.8562	on | large
0.333333	-14.8565	finding | extracted
0.0438596	-14.8566	temporal | the
0.0608696	-14.8566	and | version
0.114754	-14.8568	5.1 | our
0.0557491	-14.857	information | not
0.5	-14.8571	complementizer | wh-phrases
0.5	-14.8571	wmt | creole–english
0.5	-14.8571	conservative | (rwn)
0.0909091	-14.8573	= | max
0.0918367	-14.8576	human | this
0.333333	-14.8577	acoustic | stress
0.15	-14.8579	w | v
0.190476	-14.8579	models, | translation
0.133333	-14.858	using | (see
0.111111	-14.8581	metric | al.,
0.105263	-14.8582	3: | using
0.0927835	-14.8583	not | make
0.0454545	-14.8583	in | positive
0.0722892	-14.8585	same | have
0.0189593	-14.8588	use | a
0.0244731	-14.8588	since | to
0.5	-14.8589	ec | em-80
0.5	-14.8589	nullpi | nullpi(nullpi(i))
0.0606061	-14.859	data | at
0.3	-14.859	vds-7 | weather
0.3	-14.859	comparison | sets.
0.3	-14.859	ˆ | agenda)
0.3	-14.859	(forall | (rexist
0.3	-14.859	ξi | wxi
0.3	-14.859	95.6 | 95.0
0.3	-14.859	xj-word=“with”, | with”)
0.3	-14.859	98% | 96%
0.133333	-14.859	to | broader
0.0547945	-14.8593	for | form
0.0252161	-14.8593	that | words
0.0572519	-14.8594	to | work.
0.033419	-14.8595	the | of | pos
0.0947368	-14.8597	user | dialogue
0.080292	-14.8598	words | both
0.333333	-14.8602	still | lower
0.0169935	-14.8602	and | with | and
0.0317821	-14.8604	translation | for
0.0989011	-14.8604	the | null,
0.5	-14.8608	rnr | itexpl
0.0414201	-14.8612	value | and
0.065	-14.8614	previous | has
0.172414	-14.8616	learning | sentiment
0.0797101	-14.8618	document | word
0.040367	-14.8619	as | sentence
0.166667	-14.8621	training | phrase-based
0.0498534	-14.8621	can | but
0.139535	-14.8621	word | unsupervised
0.146341	-14.8622	tagger | pos
0.0663265	-14.8624	that | contains
0.5	-14.8627	∂ξi∂w | ∂wxi
0.5	-14.8627	s2: | 5,000ft
0.5	-14.8627	understanding | ca.
0.5	-14.8627	navgpnscore | ×=
0.5	-14.8627	willen | lezen
0.0426829	-14.8628	we | related
0.0291545	-14.8628	we | data | and
0.0425963	-14.863	methods | et
0.15625	-14.863	expect | to | more
0.0410156	-14.8631	into | that
0.0394265	-14.8631	results | from
0.230769	-14.8632	input’s | words
0.0221484	-14.8633	that | are | and
0.0394973	-14.8636	together | the
0.118644	-14.8637	between | vectors
0.0697674	-14.8639	need | that
0.230769	-14.8641	in | emerging
0.172414	-14.8642	translation | make
0.15625	-14.8642	task | context
0.00722637	-14.8642	of | a | of
0.130435	-14.8643	example | section
0.136364	-14.8645	information | technology
0.5	-14.8645	localization | paper—
0.5	-14.8645	“stuffy | “cold”
0.5	-14.8645	wiliam | thompson.
0.09	-14.8646	use | alignment
0.106667	-14.8646	) | x
0.116667	-14.8647	applied | word
0.0769231	-14.8647	tags | as
0.4	-14.8649	load | into
0.106667	-14.865	5 | than
0.181818	-14.865	cap-micro | approx-stopwords
0.0826446	-14.8651	and | conditional
0.0758621	-14.8652	an | node
0.0170269	-14.8653	it | that | the
0.0347222	-14.8653	the | unsupervised
0.0947368	-14.8654	u | n
0.25	-14.8655	denote | over
0.09375	-14.8656	that | yields
0.103896	-14.8659	> | =
0.133333	-14.8664	parse | into
0.0588235	-14.8664	authors | to
0.5	-14.8664	mkcls | pyp-1hmm-lm
0.5	-14.8664	(improvement | points).
0.15625	-14.8665	some | 1.
0.0320285	-14.8666	one | to | and
0.0877193	-14.8668	does | an
0.0542373	-14.8671	model | use
0.0590551	-14.8673	which | has
0.0271186	-14.8675	are | each
0.0753425	-14.8676	corpus | different
0.0127019	-14.8678	it | to | the
0.190476	-14.8678	actual | system
0.09375	-14.868	average | than
0.210526	-14.868	them | model.
0.0378289	-14.8682	of | sequence
0.0884956	-14.8682	c | x
0.133333	-14.8684	news | articles
0.0877193	-14.8684	use | described
0.101266	-14.8685	n | y
0.0549828	-14.8687	predict | of
0.0369231	-14.8691	table 1: | the
0.272727	-14.8692	bipartite | words
0.0436681	-14.8695	be | when
0.0576923	-14.8697	an | then
0.0546075	-14.8699	is | contains
0.5	-14.8701	theminimumdistancebetweenmentionsin | a
0.130435	-14.8702	wikipedia | english
0.130435	-14.8707	standard | were
0.230769	-14.871	emomaxent | lprop
0.25	-14.8711	arabic | –
0.0714286	-14.8713	4 | section
0.0826446	-14.8713	6 | .
0.0152471	-14.8714	the | of | set
0.222222	-14.8714	ma | ,y
0.222222	-14.8714	2.718 | |||
0.222222	-14.8714	------ | ----------
0.0736196	-14.8714	which | english
0.118644	-14.8717	as | rules.
0.151515	-14.872	target | labeled
0.151515	-14.872	focus | but
0.0785714	-14.8722	we | iterations
0.065	-14.8723	such | information
0.12766	-14.8723	set | extraction
0.12069	-14.8725	seen | figure
0.333333	-14.8726	using | (1),
0.05	-14.8729	to | development
0.0637255	-14.8733	training | all
0.0833333	-14.8736	penn | al.,
0.102564	-14.8738	translation | (koehn
0.25	-14.8739	0.64 | 0.55
0.5	-14.8739	2010): | lscript1/lscriptp
0.5	-14.8739	evidences: | ,1.0
0.5	-14.8739	0.384 | 0.155
0.0640394	-14.8741	of | comparing
0.0498534	-14.8741	used | language
0.15625	-14.8742	its | original
0.125	-14.8743	ensure | we
0.230769	-14.8745	words, | average
0.118644	-14.8752	6 | has
0.040619	-14.8754	the | l
0.2	-14.8754	no | words.
0.0373984	-14.8757	given | that | the
0.5	-14.8757	10% | s-p-n
0.5	-14.8757	tasksuccess | date=abs,
0.5	-14.8757	tense | redistributions
0.16129	-14.8758	number | threshold
0.0378913	-14.8763	of | rather
0.116667	-14.8765	domain-specific | of
0.0978261	-14.8766	consider | this
0.0722892	-14.8766	f | for
0.0727273	-14.8767	that | instead
0.0212435	-14.8768	the | is | this
0.0581395	-14.877	and | individual
0.0748299	-14.8772	source | sentence
0.5	-14.8776	auferlegt | warnt
0.5	-14.8776	injuries | infertility
0.0909091	-14.8777	from | rather
0.0806452	-14.8777	smt | al.,
0.333333	-14.8777	off | maxlength
0.151515	-14.8783	 | =
0.0833333	-14.8783	on | overall
0.0670103	-14.8783	the | bionlp
0.02015	-14.8786	it is | the | of
0.190476	-14.8786	improve | event
0.035982	-14.8786	so | a
0.0494186	-14.8787	is | meaning
0.0431965	-14.8788	however, | of | is
0.0663265	-14.8791	it | be | or
0.136364	-14.8794	• | over
0.0375817	-14.8794	consider | a
0.5	-14.8795	lengua | signos
0.5	-14.8795	design | ∧〈num〉sg)
0.5	-14.8795	international | former
0.0738255	-14.8796	therefore, | be
0.0254873	-14.8796	• | and
0.0466321	-14.88	we | less
0.151515	-14.88	pair | extracted
0.125	-14.88	speech | was
0.0731707	-14.8801	work has | of
0.0367412	-14.8803	by | this
0.0481283	-14.8806	meeting of | association
0.15625	-14.8806	performance | points
0.0987654	-14.8807	error | over
0.12069	-14.8807	be | main
0.114754	-14.8809	sentence | lexical
0.139535	-14.8809	one | associated
0.0450237	-14.8811	idea | of
0.101266	-14.8811	parse | tree
0.0510511	-14.8813	measures | in
0.0444965	-14.8813	of | wikipedia
0.210526	-14.8814	language | finite
0.5	-14.8814	» | h@yj
0.5	-14.8814	peter | nina
0.272727	-14.8815	applies | both
0.4	-14.8815	269 | 5
0.147059	-14.8821	this | forward
0.12069	-14.8825	useful | such
0.0280543	-14.8827	language | for
0.0276539	-14.8828	query | the
0.0526316	-14.883	features: | the
0.210526	-14.883	regulation | total
0.210526	-14.883	appear in | that
0.0492754	-14.8832	to | must
0.0491329	-14.8833	results | model
0.5	-14.8833	dell’ordine | …
0.0653266	-14.8835	full | is
0.0701754	-14.8836	in | distributional
0.0806452	-14.8838	module | in
0.333333	-14.884	[#(null | clit
0.0595745	-14.8842	system | based
0.0854701	-14.8843	such | null
0.190476	-14.8843	operations | by
0.190476	-14.8843	string | null
0.4	-14.8845	edge | formula
0.4	-14.8845	of | 204.
0.166667	-14.8847	tend | same
0.106667	-14.8847	human | more
0.111111	-14.8848	also | extraction
0.0159574	-14.8848	method | the
0.0449173	-14.8849	we | pos
0.0710059	-14.8852	does | have
0.5	-14.8852	entity | ead documents
0.190476	-14.8858	experimental | about
0.0528053	-14.8861	features | table
0.0758621	-14.8866	specific | be
0.112903	-14.8867	human | same
0.5	-14.8871	psycholinguistic | 9240
0.166667	-14.8872	sees | x|john
0.0508982	-14.8872	4 | 3
0.181818	-14.8876	achieved | average
0.052459	-14.8879	section | section | section
0.133333	-14.888	are | fig.
0.0535117	-14.8883	equation | of
0.021322	-14.8883	their | of
0.230769	-14.8884	ambiguity | same
0.0733333	-14.8884	entity | by
0.0614035	-14.8886	when | more
0.0738255	-14.8888	argument | as
0.147059	-14.8888	42 | 0
0.0819672	-14.8888	clearly | to
0.130435	-14.8889	objects | be
0.5	-14.889	(4b) | verb”
0.333333	-14.8891	uni8ba4为 | 将
0.151515	-14.8892	7 | 12
0.102564	-14.8893	are | combined
0.0806452	-14.8893	7 | 1
0.08	-14.8893	using | machine
0.0469974	-14.8894	of | treebank
0.0619469	-14.8894	be | because
0.0608696	-14.8898	work | work
0.108108	-14.8901	] | [
0.0197309	-14.8902	has | a
0.122807	-14.8902	the | (fellbaum,
0.333333	-14.8903	[ | valley
0.333333	-14.8903	09] | crf
0.0346741	-14.8904	are | they | the
0.15625	-14.8907	verb, | as
0.5	-14.8909	evil | glazed
0.5	-14.8909	palatalization | aspiration
0.5	-14.8909	incorrectly | tokenizer.
0.103896	-14.891	an | important
0.142857	-14.8918	, | j
0.106667	-14.892	no | =
0.0947368	-14.8922	an | example
0.09	-14.8922	while | results
0.0414938	-14.8923	value | is
0.0656566	-14.8924	language | trained
0.0987654	-14.8927	these | even
0.5	-14.8928	korhonen, | martha
0.0769231	-14.8928	for | add
0.190476	-14.893	arguments | verb
0.0780142	-14.8931	entity | from
0.133333	-14.8931	approaches | but
0.0531561	-14.8935	i | n
0.0738255	-14.8936	a | fully
0.12766	-14.8936	as | used.
0.173913	-14.8941	0 | c 0
0.16129	-14.8942	on | measures.
0.0748299	-14.8943	direct | to
0.5	-14.8948	# | timex3
0.5	-14.8948	0.0 | 3400
0.151515	-14.895	(2010) | semantic
0.0957447	-14.8951	concerned | of
0.0512821	-14.8951	source | that
0.0436782	-14.8953	2 | are
0.333333	-14.8954	modules | subtyping
0.166667	-14.8955	<w | </c>
0.12766	-14.8956	when | compared
0.0677966	-14.8956	reported | a
0.0987654	-14.8958	this | correlation
0.00896777	-14.8959	this | and
0.210526	-14.8959	at | rates
0.272727	-14.896	order | forms
0.15625	-14.8961	vector | then
0.0221465	-14.8962	the | in | data
0.230769	-14.8964	alleviate | by
0.0689655	-14.8964	4, | of
0.0847458	-14.8967	similar | can
0.5	-14.8967	entry | field-h
0.5	-14.8967	a9550-05-1-0321, | fa8750-10-1-0076.
0.122807	-14.8972	based | n-gram
0.147059	-14.8973	system | questions
0.0456853	-14.8976	to | state
0.0670391	-14.8976	finally, | on
0.101266	-14.8977	we | values.
0.0738255	-14.8979	dependency | between
0.15625	-14.8979	close | which
0.112903	-14.8979	disambiguation | be
0.333333	-14.898	nnf | 扨
0.0723684	-14.8981	extract | as
0.125	-14.8981	paper | evaluation
0.0576923	-14.8982	start | of
0.0862069	-14.8984	to | retrieval
0.0813008	-14.8986	the | w1
0.5	-14.8986	60.92 | 60.47
0.5	-14.8986	they | similar).
0.5	-14.8986	nmf | clear-cut
0.172414	-14.8988	source | domains.
0.0462725	-14.8988	can | one
0.109375	-14.8989	use | annotated
0.118644	-14.8992	x | then
0.0343407	-14.8996	of | including
0.125	-14.8996	summarization | which
0.111111	-14.8997	g | ,
0.09375	-14.9001	with | difference
0.0545455	-14.9004	try | the
0.0653266	-14.9005	natural | is
0.5	-14.9005	sketch | 8gb.
0.5	-14.9005	example: | [although
0.0987654	-14.9007	can | after
0.0388007	-14.901	then | that
0.0485714	-14.901	parsing | for
0.15625	-14.9015	since | rather
0.0987654	-14.9015	recall | all
0.0152755	-14.9016	while | the
0.0469974	-14.9017	features | which
0.0173574	-14.9018	one | to
0.0705882	-14.902	for | cluster
0.172414	-14.9021	.98 | .01
0.210526	-14.9024	research | under | national
0.25	-14.9025	not | answering
0.0656566	-14.9025	given | number
0.5	-14.9025	despite | life,
0.5	-14.9025	øj | p@y
0.0327869	-14.9026	for | which | a
0.0748299	-14.9026	using | trained
0.0134341	-14.903	the | information
0.333333	-14.9031	morgen:tomorrow | früh:early
0.0423729	-14.9032	where | number
0.0474934	-14.9033	k | and
0.16129	-14.9035	in | value,
0.0769231	-14.9035	that | predict
0.2	-14.9036	können:can liefern:deliver | es:it
0.0738255	-14.9038	t | o
0.0463918	-14.9038	could | as
0.0539568	-14.904	t | i
0.5	-14.9044	hmm-token-r | lattice-token-r
0.5	-14.9044	inception | community.
0.5	-14.9044	subnode | subnode.
0.5	-14.9044	(salim | na´s´safa
0.0701754	-14.9052	processing, pages | uk,
0.0685714	-14.9054	is | of | word.
0.0367412	-14.9055	large | of | and
0.173913	-14.9055	across | sets
0.0555556	-14.9056	lexicon | a
0.333333	-14.9057	0% | 35%
0.181818	-14.9058	⊆ | and
0.0677966	-14.9059	these | do
0.0674157	-14.906	hand, | is
0.173913	-14.9061	their | generate
0.5	-14.9063	89.53 | 89.42
0.5	-14.9063	cnrs | toulouse
0.044289	-14.9064	sentiment | in
0.02849	-14.9065	for | its
0.0450237	-14.9066	(3) | of
0.116667	-14.9069	dialogue | speech
0.0758621	-14.907	my | the
0.136364	-14.9071	observe | from
0.0590717	-14.9078	first | or
0.0409836	-14.9082	candidate | in
0.0847458	-14.9082	words | these
0.5	-14.9083	evaluation | bayesian]
0.5	-14.9083	explained | (chelba
0.0203883	-14.9084	information | of
0.0840336	-14.9084	notice | of
0.0421941	-14.9085	direct | the
0.272727	-14.9087	nullnullnullnullnull | nullnullnullnullnullnullnull
0.285714	-14.9089	ed | week
0.285714	-14.9089	pda | null
0.285714	-14.9089	q-level | q-rate
0.0331126	-14.9091	a | verb
0.112903	-14.9093	other | could
0.15625	-14.9094	areas | as
0.214286	-14.9094	twitter | switchtest
0.114754	-14.9095	baseline | models
0.0215232	-14.9097	is | model
0.5	-14.9102	crf-lm: | fischler
0.5	-14.9102	(wal) | surprise).3
0.0963855	-14.9107	experiments, | that
0.0553506	-14.9108	| | the
0.0865385	-14.911	alternative | is
0.0628019	-14.9111	original | for
0.0653266	-14.9112	system | two
0.285714	-14.9112	∂θ1 | ∂θ2
0.285714	-14.9112	students | science
0.0320592	-14.9122	the | association
0.4	-14.9122	e3 | e5
0.210526	-14.9122	first, | approach
0.0481586	-14.9124	p | is
0.0616114	-14.9127	here | for
0.0705882	-14.913	consider | with
0.0362776	-14.9131	in | work, | the
0.0595745	-14.9134	a | coreference
0.0268398	-14.9136	on | language
0.0270742	-14.9136	and | if
0.0155351	-14.9138	since | the
0.147059	-14.9138	has | main
0.0569106	-14.9139	metrics | to
0.5	-14.9141	ω(c) | (osborne,
0.5	-14.9141	choose | p(w|ϕlscript,wpar(lscript)).
0.0488506	-14.9144	who | of
0.0701754	-14.9145	each | topic
0.0494186	-14.9151	is | x
0.125	-14.9151	trained | annotated
0.062201	-14.9153	we | mean
0.0701754	-14.9158	summarization | in
0.333333	-14.9161	work | nsfc
0.214286	-14.9161	scientific | or
0.5	-14.9161	softmaxmargin | f-measure,
0.5	-14.9161	implemented | srilm-toolkit
0.5	-14.9161	(laura | 劳拉null威尔士null布什
0.5	-14.9161	informativeness | (schein
0.101266	-14.9162	word | do
0.036566	-14.9162	of | grammar
0.0221445	-14.9166	of | pages
0.285714	-14.9167	subtrees | all
0.102564	-14.9167	then | similarity
0.0531915	-14.9168	on | july
0.173913	-14.917	4th | pages
0.105263	-14.9172	acknowledgments | foundation
0.122449	-14.9174	two | first,
0.285714	-14.9179	vignettes | can
0.046875	-14.918	provided | of
0.0681818	-14.918	transition | and
0.16129	-14.918	general | about
0.5	-14.9181	|l|
0.036566	-14.9182	however, | as
0.2	-14.9185	2. | –
0.0637255	-14.9186	probability | from
0.0521173	-14.9188	in | 10
0.122449	-14.9188	that | labels.
0.0637255	-14.9188	a | agreement
0.0522876	-14.9188	1 | 7
0.12766	-14.9189	an | paper
0.285714	-14.919	summary | ~
0.214286	-14.9195	i=1 | order
0.0461538	-14.9196	to | space
0.122449	-14.9198	system | features.
0.130435	-14.9198	and | vector.
0.0329381	-14.9198	of | pos
0.0681818	-14.9199	it is | model
0.097561	-14.9199	for | f-score
0.0918367	-14.92	small | data
0.5	-14.92	reifying | first-class
0.4	-14.9201	good | it.
0.0748299	-14.9204	with | much
0.106061	-14.9204	both | score
0.0436782	-14.9205	that | order
0.0927835	-14.9207	data | over
0.0172911	-14.9212	such | of
0.059322	-14.9214	first | it
0.118644	-14.9214	language | no
0.142857	-14.9218	like | would
0.0987654	-14.9219	allow | with
0.5	-14.922	nice/nasty | agree/disagree
0.5	-14.922	0.250.3 | alarm/false
0.0963855	-14.9221	corpora | parallel
0.0595745	-14.9221	model | corpus
0.057377	-14.9223	all | number
0.0918367	-14.9225	amount | training
0.0604651	-14.9229	context | with
0.0238257	-14.9231	we | translation
0.125	-14.9235	abstract | their
0.147059	-14.9236	supervised | based
0.147059	-14.9236	also | (see
0.107692	-14.9237	which | manually
0.0941176	-14.9237	they | any
0.25	-14.924	arc | prime
0.142857	-14.924	effective | preventing
0.0743243	-14.924	previous | which
0.5	-14.924	fixed | (default
0.5	-14.924	u2(x,y,di,p,s) | u7(x,y,di,p)
0.5	-14.924	titles | redirection
0.5	-14.924	np@2null1null2] | sharon]
0.181818	-14.9243	(2010) | sentiment
0.0383212	-14.9248	model | not
0.114754	-14.9248	our | might
0.181818	-14.9251	between | links
0.0429864	-14.9253	and | higher
0.107692	-14.9254	if | consider
0.0662983	-14.9254	assume | that
0.210526	-14.9254	site(entity) | theme(protein)
0.0472222	-14.9258	annual meeting | pages
0.5	-14.926	∑ | g1867g1872ℎg1857g1870g1875g1861g1871g1857
0.5	-14.926	september, | dispute.
0.5	-14.926	bulgarian | persian
0.0412371	-14.9263	negative | of
0.0569106	-14.9264	a | allows
0.0987654	-14.9266	contrastive | the
0.285714	-14.9269	segment | trend
0.136364	-14.9269	bleu | (papineni
0.0882353	-14.9269	information | translation
0.130435	-14.927	supervised | an
0.0775194	-14.927	+ | 0
0.0806452	-14.9273	at | they
0.0376068	-14.9274	= | null
0.125	-14.9274	dependency tree | is
0.0662983	-14.9275	our | so
0.0891089	-14.9276	to | connected
0.0233645	-14.9276	is | such
0.0142787	-14.9278	with | of | and
0.5	-14.9279	vpart | vger
0.5	-14.9279	85.7% | psd
0.5	-14.9279	‘derek | dominos’.
0.1	-14.928	token | by
0.0194394	-14.9281	a | such
0.00184177	-14.9281	the | the
0.0566802	-14.9288	we | discourse
0.147059	-14.9289	to | reorderings
0.139535	-14.929	experiments | their
0.151515	-14.9293	and | forest.
0.0318627	-14.9293	and | about
0.0313253	-14.9295	this | a | on
0.0495356	-14.9296	one | can
0.0363349	-14.9297	does | a
0.5	-14.9299	tβtcountt(f,e)summationtext | tβtcountt(f,e)summationtext
0.5	-14.9299	sufficient | repeatable
0.5	-14.9299	tokenization | 63.10%
0.0666667	-14.9301	as | sentiment
0.16129	-14.9301	through | research
0.285714	-14.9303	ind | but
0.0723684	-14.9303	bionlp | task
0.0590717	-14.9303	although | as
0.0738255	-14.9305	fact | not
0.112903	-14.9306	6 | 12
0.038961	-14.9308	experiments | that
0.118644	-14.9308	potential | on
0.0987654	-14.9313	our | joint
0.0313253	-14.9313	such | is | the
0.0775194	-14.9316	for | hierarchical
0.0533808	-14.9318	of | manual
0.05	-14.9319	y | and
0.5	-14.9319	score(candidatei) | '(–(candidatei,
0.5	-14.9319	tomoko | {okap,smp}@is.s.u-tokyo.ac.jp,
0.5	-14.9319	g1008 | jwn
0.097561	-14.9323	are | effective
0.0918367	-14.9323	we | computation
0.0421286	-14.9324	them | and
0.230769	-14.9325	their | contexts.
0.173913	-14.9328	treebank | section
0.272727	-14.9334	yang | al.
0.125	-14.9334	target | speech
0.5	-14.9339	jingbo | muhua
0.147059	-14.9348	does | appear
0.0909091	-14.9349	these | because
0.0733333	-14.9351	on | processing
0.5	-14.9359	demonstrative | (this,
0.166667	-14.936	λ1=0.0 | λ2=0.0
0.0288363	-14.936	and | two | and
0.0458015	-14.9362	order | with
0.0422222	-14.9363	a | contains
0.15625	-14.9366	see | models
0.102564	-14.9374	that | q
0.0641711	-14.9374	empirical methods | processing,
0.0763359	-14.9375	step | as
0.0280464	-14.9377	for | same
0.106061	-14.9377	or | very
0.0194175	-14.9378	the | similar
0.0195077	-14.9378	the | in table
0.5	-14.9379	cutoff | fci,ej
0.5	-14.9379	2.08 | 1.96
0.5	-14.9379	argmin | lλ(h(θ·x),y)
0.0873786	-14.9381	and | implemented
0.0566802	-14.9383	as | work
0.0273224	-14.9385	many | of
0.16129	-14.9385	pos | head
0.0571429	-14.9387	dialogue | is
0.078125	-14.9387	is | by | al.
0.112903	-14.9388	from | potential
0.0628019	-14.9389	out | on
0.272727	-14.9389	0.79 | 0.57
0.147059	-14.9389	unsupervised | tagging
0.0918367	-14.9389	by | size
0.111111	-14.939	distance | similarity
0.173913	-14.9391	patterns | pairs
0.118644	-14.9393	prob | prob
0.109375	-14.9394	as | documents.
0.2	-14.9399	virtual | be
0.5	-14.9399	sails | sailing
0.5	-14.9399	iteration | htbunigram
0.5	-14.9399	189.1 | 147.1
0.0681818	-14.94	conference on | july
0.0305776	-14.9401	as | words
0.230769	-14.9408	allows | capture
0.0298013	-14.9411	is used | the
0.0590717	-14.9412	to | i.e.,
0.125	-14.9414	based | simple
0.0393258	-14.9418	even | in
0.0391061	-14.9418	percentage | the
0.5	-14.9419	multiplying | on−→far,
0.5	-14.9419	90.43 | 83.93
0.5	-14.9419	(ri, | r2),(r1−1,
0.5	-14.9419	nv | kontuan
0.166667	-14.942	4 | quality
0.0813008	-14.9421	only | these
0.0857143	-14.943	al., | 2007)
0.0302352	-14.9431	are | used
0.0292276	-14.9436	a | new
0.0670391	-14.9438	(2010) | on
0.2	-14.9439	<c> | <w
0.5	-14.9439	rashmi | nikhil
0.5	-14.9439	clause | (cb
0.285714	-14.944	attribution | models
0.0117786	-14.944	for | as
0.107692	-14.9444	acknowledgements | was | the
0.230769	-14.9446	collection | corpus.
0.0625	-14.9453	idea | to | and
0.0339943	-14.9454	only | as
0.190476	-14.9455	abstract | improve
0.0194535	-14.9455	possible | the
0.0469613	-14.9456	we | automatic
0.055336	-14.9457	in | subject
0.0857143	-14.9458	g | p
0.5	-14.946	another | nounphraseinthecontext.
0.0214932	-14.9464	better | the
0.104478	-14.9473	following | phrase
0.16129	-14.9476	out | such
0.0571429	-14.9477	on | order
0.0787402	-14.9477	bleu | test
0.0833333	-14.9477	there are | not
0.0222635	-14.9478	the | out
0.5	-14.948	x(t)i | y(t)j
0.0547445	-14.9482	for | original
0.230769	-14.9483	set | restricted
0.214286	-14.9486	able | account
0.0610329	-14.9487	for | work.
0.117647	-14.9488	especially | be
0.151515	-14.9488	with | category.
0.0219194	-14.9489	work | a
0.111111	-14.9489	do | appear
0.12766	-14.949	matrix | word
0.055336	-14.949	shows | from
0.109375	-14.9491	has | associated
0.190476	-14.9493	the | representativeness
0.0199029	-14.95	we | that | for
0.0395257	-14.95	context | is
0.5	-14.95	ilavach | kisou
0.0325945	-14.9504	this | a | from
0.0288958	-14.9508	in | may
0.285714	-14.951	extract | statistical
0.102941	-14.9511	models | given
0.15625	-14.9511	word | corpora.
0.122449	-14.9516	0 | k
0.0670391	-14.9517	table | statistics
0.181818	-14.952	work | mining
0.285714	-14.9521	s51 | s97
0.272727	-14.9521	sentiment | 2009;
0.0638298	-14.9523	only | training
0.0413943	-14.9524	reference | of
0.0541516	-14.9525	model | two
0.0700637	-14.9527	much | with
0.166667	-14.9528	(see | 1)
0.0723684	-14.9531	often | by
0.25	-14.9531	operators | by
0.2	-14.9537	are | recognize
0.0601852	-14.9541	2, | in
0.130435	-14.9542	computational | pp.
0.272727	-14.9543	ﺔﯿﺑﺮﻌﻟا | manuscripts
0.272727	-14.9543	sent | +0.07
0.272727	-14.9543	transliteration | l,s
0.272727	-14.9543	sundaradikhataapahaada | huaa
0.272727	-14.9543	+morph | stem+
0.272727	-14.9543	dog(a) | reducer
0.272727	-14.9543	تﺎﻃﻮﻄﺨﻤﻟا | arabic
0.272727	-14.9543	0.23, | 0.13,
0.272727	-14.9543	vp | vbn+p
0.136364	-14.9543	significant | our
0.0613208	-14.9547	research | that
0.0382514	-14.9549	were | in | and
0.181818	-14.955	one | parsers
0.0433735	-14.9555	the | goal
0.0373665	-14.9557	in | structure
0.097561	-14.9559	experiments | both
0.057377	-14.9561	specific | for
0.111111	-14.9561	show | even
0.5	-14.9561	relations | panchenko
0.097561	-14.9563	0.1 | 0
0.190476	-14.9563	be | x.
0.117647	-14.9564	that | japanese
0.214286	-14.9564	approaches | wordnet
0.00928594	-14.9565	is | a | of
0.068323	-14.9572	both | features
0.181818	-14.958	cosine | their
0.5	-14.9581	comparisons, | ast.
0.25	-14.9582	second | category
0.0498442	-14.9587	this | context
0.0655738	-14.9587	smt | a
0.142857	-14.9591	nlp | machine
0.0793651	-14.9594	linguistics | research
0.190476	-14.9594	involve | such
0.0841121	-14.9595	these | could
0.230769	-14.9596	furthermore, | task
0.0383212	-14.9597	that | a | than
0.0987654	-14.96	features | better
0.5	-14.9602	judges | s-ns
0.5	-14.9602	kanayama | nasukawa,
0.5	-14.9602	institut | université
0.5	-14.9602	seq-aln | g-cut
0.5	-14.9602	onomic | owing
0.4	-14.9603	all | 98.6
0.0419426	-14.9604	word | an
0.0408602	-14.9607	baseline | for
0.117647	-14.9608	0.005 | 20
0.0333333	-14.9608	of | positive
0.142857	-14.9608	output | y
0.044226	-14.9609	work | has
0.166667	-14.961	demonstrations, pages | 21
0.0146628	-14.961	to | this
0.0375671	-14.9611	performance of | and
0.022291	-14.9612	we | that | as
0.062201	-14.9612	data | use
0.0613208	-14.9612	the | effects
0.0616114	-14.9616	3 | results
0.101266	-14.9617	approach | number
0.018	-14.9618	for | we | a
0.190476	-14.9618	v | edge
0.4	-14.9619	boys | girls.
0.133333	-14.9619	there | four
0.08	-14.9622	but | two
0.133333	-14.9624	based | set.
0.0223187	-14.9625	of | most
0.0355987	-14.9627	topic | to
0.109375	-14.9629	first, | can
0.0208219	-14.963	the | trained
0.106061	-14.9631	member | of
0.0274102	-14.9632	from | al.,
0.0638298	-14.9633	this | presents
0.0537634	-14.9635	pages | july | 2011.
0.047619	-14.9635	can | two
0.106061	-14.9635	models | alignment
0.151515	-14.9636	sentence | without
0.0375671	-14.9639	that | while
0.0309524	-14.9643	3 | 1 | 1
0.142857	-14.9643	6 | senses
0.5	-14.9643	0.454 | 0.518
0.5	-14.9643	p(wn|s(wn−11
0.25	-14.9644	participants | over
0.0540541	-14.9645	an | when
0.173913	-14.9645	from | hierarchy
0.2	-14.9651	attractive | a
0.0733333	-14.9651	is | contexts
0.4	-14.9652	dual | multipliers
0.0818182	-14.9654	our | (see
0.25	-14.9654	(maccartney | al.,
0.333333	-14.9657	determining | properties
0.0259197	-14.9662	of | about
0.0891089	-14.9662	segmentation, | the
0.0369718	-14.9663	to | corresponding
0.5	-14.9664	ˆw | zł’,
0.5	-14.9664	k, | topd(w)
0.5	-14.9664	ut | rce
0.285714	-14.9673	cases | negative
0.00968834	-14.9673	and | with
0.12	-14.9674	were | with | were
0.0133452	-14.9674	the | also
0.0479042	-14.9675	not | their
0.12	-14.9678	can | impact
0.0898876	-14.9678	with | right
0.0282258	-14.9678	the | of | class
0.0262697	-14.9678	order | to
0.151515	-14.968	tg | of
0.0793651	-14.9682	if | do
0.078125	-14.9683	our | sentiment
0.5	-14.9684	insertion | efl
0.5	-14.9684	translate | answers
0.5	-14.9684	pprime(w|h)
0.5	-14.9684	justput | mentions(namesandnominalmentions,exceptpronouns)whichhavethesame
0.5	-14.9684	g1007 | ewn
0.5	-14.9684	bakeoff | sinica
0.5	-14.9684	logic | linguistics.
0.4	-14.9685	1000k | 5000k
0.333333	-14.9685	p(t|d)
0.0434783	-14.9687	other | al.,
0.16129	-14.9687	<cell | <cell
0.0714286	-14.9694	plan | in
0.104478	-14.9695	over | same
0.142857	-14.9696	by | measured
0.285714	-14.9697	fi | @
0.0134108	-14.9703	this | that | the
0.102941	-14.9705	other | about
0.0520833	-14.9705	to | web
0.5	-14.9705	p(syn|p)
0.5	-14.9705	g′, | g′
0.5	-14.9705	gu | (gu
0.5	-14.9705	orientation | psychology
0.5	-14.9705	oliver | heriot-watt
0.075188	-14.9706	however, | they
0.135135	-14.9706	corpus, | was
0.117647	-14.9709	although | between
0.078125	-14.9709	process | by
0.0266904	-14.9712	as | word
0.078125	-14.9712	performance | the | data.
0.0898876	-14.9715	w | word
0.0108441	-14.9715	the | language
0.25	-14.9716	z | q
0.0930233	-14.9717	types | our
0.0436893	-14.9718	is | there is
0.12	-14.9718	deviations | the
0.4	-14.9718	0.6 | 0.04
0.12	-14.9723	values | were
0.0297357	-14.9723	the | we have
0.109375	-14.9724	however, it | not
0.0314861	-14.9724	the | measures
0.0469613	-14.9724	lexical | are
0.0344828	-14.9726	then | a | of
0.5	-14.9726	f1 | 2×
0.0898876	-14.973	size | not
0.0882353	-14.9731	scores | test
0.285714	-14.9732	team | uu
0.166667	-14.9733	(2009), | al.
0.181818	-14.9737	shows | full
0.0175781	-14.9739	a | which | of
0.0555556	-14.974	similar | to | and
0.091954	-14.9741	pi | subsetsqequal
0.0588235	-14.9746	that | pos
0.5	-14.9746	spiele | ab.
0.5	-14.9746	sonata(x2):0 | art(e1,x1,x2):10
0.5	-14.9746	iffer | ilatio
0.12766	-14.9747	latter | be
0.0638298	-14.9749	as | corresponding
0.0580357	-14.9752	r | f
0.0930233	-14.9752	system | most
0.190476	-14.9752	could | parallel
0.0714286	-14.9753	evaluation | human
0.0298851	-14.9754	of | parallel
0.0319285	-14.9757	idea | the
0.0375671	-14.9757	trained | a
0.0655738	-14.9762	word | feature
0.1	-14.9764	so | word
0.0621762	-14.9765	(e.g., | is
0.0177515	-14.9766	the | is | =
0.2	-14.9767	71) | 86)
0.2	-14.9767	saw(i) | dog(a))
0.2	-14.9767	=> | \w*{w1}\w*_(\d+)
0.2	-14.9767	t6 | t20
0.2	-14.9767	a9a8a11a9a10a12a13a14a12a15 | a24a11a25a8a23a22
0.5	-14.9767	overall | 91.12%,
0.5	-14.9767	logp(r | logp(i
0.181818	-14.9767	features | 2010),
0.4	-14.9768	area1 | district
0.4	-14.9768	re-ranking | b′(s)
0.4	-14.9768	unwilling, | .97
0.4	-14.9768	20,169 | 15,3247
0.4	-14.9768	pietra, | pietra,
0.4	-14.9768	nullnull1null2[ | null2)
0.4	-14.9768	restriction: | <s
0.4	-14.9768	grade | grades
0.4	-14.9768	peripheral | group:
0.4	-14.9768	w∈k | +β#(c,w)
0.4	-14.9768	gn | cign
0.4	-14.9768	n1 | vextendsinglevextendsinglevextendsinglevextendsingle2
0.4	-14.9768	|φ | ))
0.4	-14.9768	=⇒ | x©xprime
0.4	-14.9768	ip | �
0.4	-14.9768	vms | genesis
0.4	-14.9768	[4,4][1,1] | indexsort
0.4	-14.9768	vocabulary: | closed;
0.4	-14.9768	(real) | (integer)
0.4	-14.9768	circleplusdisplay | a(x
0.4	-14.9768	)=p(x | |y
0.4	-14.9768	p(x | |y
0.4	-14.9768	cough | shortness
0.4	-14.9768	tj, | tj
0.4	-14.9768	s/ε, | -/ε,
0.4	-14.9768	production | produits
0.4	-14.9768	<antibiotic | />
0.4	-14.9768	cat | “cat
0.4	-14.9768	maxsim | papers)
0.4	-14.9768	dell | $27
0.4	-14.9768	―व‖ | ―a‖
0.4	-14.9768	heeft | gelopen
0.4	-14.9768	ses | x|john
0.4	-14.9768	keyspace | cipher
0.4	-14.9768	p0(t,j|s)z(s) | wk
0.4	-14.9768	“edited” | y⋆i
0.4	-14.9768	h21, | h31,
0.4	-14.9768	m., | sabo,
0.4	-14.9768	hat | butterkekse
0.4	-14.9768	know | golgi
0.4	-14.9768	ser(727) | <theme<site
0.4	-14.9768	80.5 | +negative
0.4	-14.9768	(+/-) | vb,
0.4	-14.9768	r2: | z);
0.4	-14.9768	0.5896 | ctpr
0.4	-14.9768	x| | reduction
0.4	-14.9768	square6 | 2006]
0.4	-14.9768	ordered, | <s
0.4	-14.9768	r-2 | 16.8
0.4	-14.9768	lem | 1/2/3-gram
0.4	-14.9768	cambridge | cambridge,
0.4	-14.9768	rouge-1(%) | sentence-ranking
0.4	-14.9768	ms,t | xh,s
0.4	-14.9768	7 | ap-10
0.4	-14.9768	non-mwe | pmi
0.4	-14.9768	rel, | abs,
0.4	-14.9768	interpolation | φn
0.4	-14.9768	localcost(˚ι,l) | branch
0.4	-14.9768	+0.09 | +0.02
0.4	-14.9768	paperclip | magnet
0.4	-14.9768	a71random | model−var
0.4	-14.9768	pp/nn | schliessen
0.4	-14.9768	pmi(hw,mw) | mt,
0.4	-14.9768	…, | 보다
0.4	-14.9768	wte | qec
0.4	-14.9768	hand<adv> | friend<n><pl>
0.4	-14.9768	unaries | 950,028
0.4	-14.9768	bl | 22.12
0.4	-14.9768	claimed, | spelled
0.4	-14.9768	0.1982 | 0.1171
0.4	-14.9768	dg | rebase
0.4	-14.9768	pb(w3|w1w2)
0.4	-14.9768	por | for);
0.4	-14.9768	,h,x | lscript
0.4	-14.9768	zpar | penntb
0.4	-14.9768	ci∈g | #(wj,dk)κji
0.4	-14.9768	(men) | dhaarana8
0.4	-14.9768	x(q)j | }mi=1
0.4	-14.9768	co-segmentation | (sk,tk)
0.4	-14.9768	$←said | $←he
0.4	-14.9768	conf | 20406080100
0.4	-14.9768	”gaddafi | hundreds
0.4	-14.9768	ij | cij(−→ni⊗−→nj)
0.4	-14.9768	d57d57 | d99
0.4	-14.9768	re-ranked; | re-ranking
0.4	-14.9768	〈h,m〉 | (indicating
0.4	-14.9768	(σ|i|j,β,a) | turnstileleft
0.4	-14.9768	39.30 | (online)
0.4	-14.9768	am | (pure
0.4	-14.9768	euro, | lire, | us
0.4	-14.9768	<> | verbsubj
0.4	-14.9768	sad,” | strained,”
0.4	-14.9768	punjabi | patiala,india
0.4	-14.9768	s101 | s70s101s97s116s117s114s101s32s100s105s109s101s110s115s105s111s110s32s40s75s41
0.4	-14.9768	[s/he | teaching]
0.4	-14.9768	intro(f) | subsetsqequal
0.4	-14.9768	msummationdisplay | objscore[r]
0.4	-14.9768	omitted | ∂f⟨2,k2,e2,x⟩
0.4	-14.9768	cjj | ∧ckj
0.4	-14.9768	xtmax | precedes
0.4	-14.9768	words), | (weeks
0.4	-14.9768	=e | ∧∥
0.4	-14.9768	points), | classificationversus
0.4	-14.9768	vm/m-1 | vm/m1
0.4	-14.9768	{people’s | china,
0.4	-14.9768	phosphorylation | phosphorylationser(727)
0.4	-14.9768	x3 | block-compresses
0.4	-14.9768	[suo | ,]
0.4	-14.9768	34.2±3.4 | epsilon1
0.4	-14.9768	wu,m,n | yu,m,n
0.4	-14.9768	h1 | label/select
0.4	-14.9768	tr | 
0.4	-14.9768	3.9 | 4.2
0.4	-14.9768	void | params
0.4	-14.9768	south | patents
0.4	-14.9768	to”, | plan
0.4	-14.9768	bone | fragile,
0.4	-14.9768	extractive | eij
0.4	-14.9768	0.44 | 0.44
0.4	-14.9768	chyp(wc) | wc
0.4	-14.9768	kernelized | (kulis
0.4	-14.9768	p(r4 | r1null
0.4	-14.9768	wpp+ | atures
0.4	-14.9768	ap | 28
0.4	-14.9768	(questions | combined),
0.4	-14.9768	md. | †
0.4	-14.9768	coffee))); | =+
0.4	-14.9768	turk | 52.8%
0.4	-14.9768	algo | rsing
0.4	-14.9768	ˆθ | 〈x,y〉∈x×y
0.4	-14.9768	dépecer | s=
0.4	-14.9768	a150a151a142a143a144a145a146 | president
0.4	-14.9768	nx | iani
0.4	-14.9768	marshall | mathers
0.4	-14.9768	tim | pam
0.4	-14.9768	tu | te
0.4	-14.9768	[went]y | r2:
0.4	-14.9768	countt(f,e)summationtext | βt
0.4	-14.9768	“on | three”
0.0634921	-14.977	to | corpora.
0.0330124	-14.9771	pairs | in
0.0132509	-14.9771	the | of | =
0.0536398	-14.9775	that | value
0.25	-14.9778	reduction | bleu
0.333333	-14.9781	size | 16.
0.122449	-14.9783	49 | of
0.2	-14.9784	party | gesehen
0.102941	-14.9787	structure | more
0.15625	-14.9787	token | has
0.117647	-14.9788	terms | both
0.166667	-14.9789	was | pairs.
0.0709677	-14.9791	our | could
0.285714	-14.9792	in | 5.3,
0.0833333	-14.9792	dependency | parsing
0.0718954	-14.9798	with | extracted
0.117647	-14.9802	precision | 1
0.054902	-14.9805	x | 1
0.0746269	-14.9806	number of | using
0.230769	-14.9806	random | 1.0
0.0670732	-14.9808	were | corpus
0.0763359	-14.9809	same | this
0.5	-14.9809	‘|’
0.0434783	-14.9813	there are | a
0.122449	-14.9814	system | compared
0.0625	-14.9815	not | would
0.075188	-14.9818	while | word
0.0356564	-14.9819	experiments, | the
0.0585586	-14.9822	to | observed
0.333333	-14.9823	|h) | p(h
0.0562249	-14.9828	only | data
0.0445026	-14.9829	to | existing
0.025729	-14.9829	2 | to
0.5	-14.983	wikification | scaiella,
0.5	-14.983	plank | groningen
0.0882353	-14.9837	1: | word
0.214286	-14.9839	our | lies
0.2	-14.9843	candidate | select
0.0909091	-14.9844	each | words.
0.0930233	-14.9845	years, | and
0.0436893	-14.9846	set | training
0.109375	-14.9847	implicit | a
0.056	-14.9849	is | provide
0.5	-14.9851	ictclas | ec.
0.5	-14.9851	4114 | 71.03
0.0810811	-14.9856	worth | of
0.138889	-14.9861	the | mistakenly
0.16	-14.9862	word | ibm
0.173913	-14.9862	indicates | using
0.333333	-14.9865	tools | already
0.0453333	-14.9865	there | a | that
0.173913	-14.987	methods | 2010)
0.0448549	-14.987	3: | for
0.5	-14.9872	(fp7-ict-2007-3-231720 | 7e09003
0.0147879	-14.9873	of | are | in
0.0510204	-14.9874	that | small
0.109375	-14.9875	natural | generation
0.0803571	-14.9875	we | constituent
0.2	-14.9876	summationtext[[w]] | [[w]]
0.0963855	-14.9877	experiments | test | the
0.122449	-14.9881	found | set
0.0570175	-14.9881	when | model
0.12	-14.9884	common | used
0.0670732	-14.9884	has | translation
0.0403397	-14.9888	2 | we
0.0449735	-14.9889	the | frequently
0.111111	-14.9889	has | sentiment
0.0542636	-14.989	i | with
0.097561	-14.9891	terms | number
0.147059	-14.9891	kernels | tree
0.0216867	-14.9892	are | not
0.079646	-14.9893	our | consists
0.109375	-14.9893	case, | by
0.5	-14.9893	99.9 | 54.2%4195
0.5	-14.9893	regression | (regmt)
0.0329986	-14.9894	these | a | a
0.147059	-14.9897	recognition | speech
0.0691824	-14.9898	clustering | is
0.173913	-14.9899	ter | single
0.0787402	-14.99	time | we
0.0481928	-14.9901	cluster | and
0.0389864	-14.9908	focus | and
0.0306373	-14.9911	as | one
0.5	-14.9914	network | g(v,e)
0.5	-14.9914	320k | 43.02
0.122449	-14.9917	log | p
0.033675	-14.9917	and | no
0.0436893	-14.9919	in | rather
0.190476	-14.9919	input | c
0.173913	-14.9921	intrusion | a71a71
0.0088434	-14.9925	model | the
0.0320427	-14.9931	the | provides | the
0.0178644	-14.9931	all | a
0.130435	-14.9932	directly | word
0.0679012	-14.9932	linguistics, | 19-24,
0.0394191	-14.9932	parsing | in
0.230769	-14.9932	(2002) | al.
0.135135	-14.9933	6: | 8:
0.0201655	-14.9933	for | word | the
0.5	-14.9935	pθt(x|yi)
0.5	-14.9935	sub-grammars, | encapsulation,
0.0373665	-14.9936	this | as a
0.0401691	-14.9943	if | from
0.0724638	-14.9945	thanks | the
0.0160804	-14.9948	in | word
0.0455764	-14.9948	analysis | is
0.0789474	-14.9951	full | that
0.0621762	-14.9954	is | models.
0.0298851	-14.9957	we | target
0.0542636	-14.9957	those | as
0.5	-14.9957	(alshawi | crouch
0.5	-14.9957	people | interrupting
0.5	-14.9957	1|d|
0.0618557	-14.9958	used | semantic
0.0857143	-14.9958	analysis | section
0.0289389	-14.9969	each | on
0.122449	-14.9974	t | source
0.0200617	-14.9976	can | that
0.333333	-14.9978	-3 | 3
0.5	-14.9978	bengali | chakraborty
0.5	-14.9978	sentic | (cambria
0.029148	-14.9979	that | both
0.0700637	-14.9982	example, | but
0.0490196	-14.9982	= | s
0.0415704	-14.9984	table | shows | and
0.130435	-14.9988	(*) | (*)
0.130435	-14.9988	♣ | estad´ısticas
0.12766	-14.9988	another | use
0.0340237	-14.9988	that | models
0.16	-14.999	measures | correlation
0.078125	-14.9991	first | training
0.0517241	-14.9992	we | sequence
0.214286	-14.9993	0.24 | table
0.0452128	-14.9995	this | does
0.151515	-14.9998	) | h
0.5	-14.9999	yaqin | waltham,
0.5	-14.9999	(scisumm | mead),
0.5	-14.9999	85.9 | 86.7
0.113208	-15.0001	hence | be
0.230769	-15.0001	rush | al.
0.214286	-15.0002	(blei | 2003).
0.0985915	-15.001	to | annotators.
0.12766	-15.001	prediction | model
0.0818182	-15.0011	manually | that
0.16	-15.0011	abstract | corpora
0.138889	-15.0014	they | sentences.
0.054902	-15.0014	language | used
0.115385	-15.0016	data | learn
0.0276954	-15.0016	rules | of
0.138889	-15.002	of | similar.
0.5	-15.002	proliferation | social
0.5	-15.002	heat | mimic.
0.5	-15.002	persons | birthplace,
0.5	-15.002	frame | communicator
0.5	-15.002	pr(q|ei)
0.131579	-15.0023	closer | that
0.0818182	-15.0025	human | translation
0.0577778	-15.0026	is | overall
0.138889	-15.0026	] | 0
0.070922	-15.0028	since | one
0.0146288	-15.0029	on | for
0.0696203	-15.003	a | (section
0.122449	-15.0032	used | phrases
0.142857	-15.0033	some | part
0.190476	-15.0033	also | technique
0.0879121	-15.0035	data | described
0.0909091	-15.004	only | those
0.113208	-15.004	measure | used
0.0201058	-15.004	an | to | of
0.131579	-15.004	phrase | while
0.0588235	-15.0041	in | modeling
0.5	-15.0042	the|j|’th
0.5	-15.0042	l) | a,then
0.5	-15.0042	excel | jaideep
0.5	-15.0042	campus | intelligent
0.5	-15.0042	such | vorndran,
0.151515	-15.0043	extent | can
0.166667	-15.005	will | assigned
0.138889	-15.005	models | either
0.0612245	-15.0052	of | expressed
0.173913	-15.0054	he | have
0.113208	-15.0059	expressed | material
0.0520446	-15.006	1 | )
0.0522388	-15.0062	more | such
0.5	-15.0063	1.84 | 1.81
0.0551181	-15.0067	it | semantic
0.0714286	-15.007	to | not.
0.0909091	-15.0075	not | gold
0.0471976	-15.0075	on | dependency
0.0498339	-15.0076	several | et
0.0898876	-15.0076	widely | for
0.0879121	-15.0076	semi-supervised | to
0.0206013	-15.0077	would | to | the
0.102941	-15.0079	not | form
0.0377358	-15.0079	the | by | system
0.5	-15.0085	coloquium, | 13–143.
0.058296	-15.0086	similar to | et
0.230769	-15.009	makes | single
0.106061	-15.009	task | both
0.0593607	-15.0097	that | applied
0.12766	-15.0098	there | (e.g.,
0.147059	-15.0098	further | section
0.0952381	-15.0099	has | parsing
0.115385	-15.01	process | data
0.0152114	-15.0101	a | or | the
0.0860215	-15.0103	although | been
0.0985915	-15.0103	score | given
0.135135	-15.0106	will | user
0.0174326	-15.0106	for | an
0.5	-15.0106	passive | ats
0.5	-15.0106	jordan|michael
0.5	-15.0106	ctb2tp | 87.20
0.115385	-15.011	thus, | =
0.0129956	-15.0113	we | the | from
0.0662651	-15.0116	general | are
0.0418605	-15.0118	as | al.
0.0540541	-15.0119	to | generation
0.0506757	-15.0122	problems | of
0.0321678	-15.0125	used | on
0.102941	-15.0128	that | twitter
0.5	-15.0128	p(wi|wi−11 )≈p(wi|wi−1i−n+1)
0.5	-15.0128	american | (anc)
0.5	-15.0128	yoo | gretzel
0.0849057	-15.013	is | well.
0.0195783	-15.013	does | the
0.0448549	-15.0132	where | set
0.0250417	-15.0137	2 | the | in
0.00567084	-15.0142	we | to
0.106061	-15.0145	if | k
0.5	-15.0149	hgfc | soft
0.5	-15.0149	claremont, | 91711
0.5	-15.0149	55.2 | 50.6
0.12	-15.015	averaged | all
0.115385	-15.015	that | covered
0.135135	-15.0152	classification | figure
0.0719424	-15.0156	number | average
0.102941	-15.0159	performed | al.,
0.106061	-15.0159	example, | might
0.125	-15.016	the | described | section.
0.0898876	-15.0165	problem | or
0.0879121	-15.0167	has | these
0.0283613	-15.0168	this | word
0.0283316	-15.0169	types | of
0.12766	-15.0169	using | languages
0.111111	-15.0169	based | very
0.0869565	-15.017	using | most
0.0385396	-15.0175	+ | for
0.0135626	-15.018	in | not
0.0615385	-15.0181	of | direct
0.0544747	-15.0186	a | unlabeled
0.173913	-15.019	found | do
0.0279793	-15.0194	and | data.
0.06	-15.0194	identified | a
0.0111989	-15.0194	on | in
0.102941	-15.0195	system | shown
0.0818182	-15.0201	lemmas | the
0.135135	-15.0205	first | parse
0.0631579	-15.0206	of | h
0.153846	-15.0207	such | tend
0.0319889	-15.0209	the | for this
0.0348259	-15.021	to | 2010).
0.2	-15.021	said | annotation
0.0402542	-15.021	to | 2009).
0.0958904	-15.0212	and | negation
0.0841121	-15.0213	which | model.
0.5	-15.0215	technology, | turku,
0.5	-15.0215	bhdb | geocoding
0.5	-15.0215	salience | connotational
0.12	-15.0218	when | similar
0.0888889	-15.0219	of | annotators.
0.173913	-15.022	suppose | be
0.0825688	-15.022	mapping | is
0.068323	-15.0221	on | user
0.0418605	-15.0221	using | or
0.190476	-15.0222	observe | language
0.045977	-15.0226	significantly | and
0.153846	-15.0227	semantic | available
0.0869565	-15.0227	been | 2008).
0.2	-15.0228	use | evidence
0.0789474	-15.0229	content | be
0.285714	-15.0231	learning | manning,
0.0249377	-15.0232	sentences | to
0.101449	-15.0234	a | developing
0.5	-15.0236	p(w[pij]|w[pij−1],...,w[pij−m+1])
0.5	-15.0236	authors | endorsed
0.0479233	-15.0238	reference | in
0.0609137	-15.0238	do | this
0.0186654	-15.0239	note | the
0.0450704	-15.024	are | text
0.0256864	-15.0242	of | evaluation
0.2	-15.0245	et | 2007,
0.0103959	-15.0245	a | by
0.16	-15.0246	next, | an
0.181818	-15.0248	acknowledge | contract
0.0228532	-15.0249	is | language
0.0572687	-15.0249	an | text
0.0453258	-15.025	2010), | al.,
0.0958904	-15.025	see | this
0.0679012	-15.0253	c | with
0.0740741	-15.0258	of | correction
0.5	-15.0258	post, | drafts,
0.5	-15.0258	0, | c1
0.106061	-15.026	could | have
0.0631579	-15.0261	a | identified
0.2	-15.0263	another | difference
0.12	-15.0265	table | adding
0.0238095	-15.0266	we | sentence
0.138889	-15.0266	been | applications
0.285714	-15.0268	on | rights
0.0239567	-15.0273	the | top
0.04	-15.0276	the | long
0.0498339	-15.0277	the | 3,
0.0283533	-15.0279	target | to
0.05	-15.028	a | development
0.5	-15.028	p(y|x
0.5	-15.028	bin | chenhao
0.5	-15.028	:) | :]
0.333333	-15.0281	bacteria | event
0.142857	-15.0287	discourse | connectives
0.0597015	-15.0291	closely | to
0.0306122	-15.0293	the | end
0.0546218	-15.0293	word | features
0.0562771	-15.0299	a | 2006).
0.014255	-15.03	are | to | of
0.5	-15.0302	within | document”
0.5	-15.0302	(java | petrovi´c
0.0724638	-15.0304	and | problem.
0.173913	-15.0304	provide | performance
0.0543933	-15.0306	(2008) | of
0.0543933	-15.0308	feature | used
0.039501	-15.0308	are | use
0.0958904	-15.0313	it | features.
0.135135	-15.0318	pair | they
0.0689655	-15.032	2 | have
0.0369004	-15.0321	other hand, | the
0.333333	-15.0324	solution | require
0.0448179	-15.0324	training | set
0.0958904	-15.0325	research | university
0.0510949	-15.0329	senses | of
0.0985915	-15.0331	precision | not
0.0658683	-15.0337	type | from
0.0526316	-15.0338	for | multiple
0.0756303	-15.0339	of | distribution.
0.333333	-15.0339	= | p(wi|wi−11
0.0860215	-15.034	are | always
0.0154192	-15.0343	of | for | for
0.214286	-15.0346	used | helpfulness
0.0373134	-15.0347	as | or | a
0.153846	-15.0348	less | compared
0.0632184	-15.0349	context | from
0.011831	-15.0349	an | to
0.0496689	-15.0352	for | 0
0.135135	-15.0353	in | resource.
0.0363636	-15.0354	of | science
0.333333	-15.0354	manandhar | of
0.0516605	-15.0355	a | j
0.0945946	-15.0357	+ | m
0.115385	-15.0357	parser | or
0.0470588	-15.0358	or | more
0.16	-15.0358	to | projective
0.0373134	-15.0358	being | of
0.0548523	-15.0359	are | relation
0.104478	-15.036	all | label
0.0194373	-15.0364	system | in
0.122449	-15.0367	extracted | all
0.5	-15.0368	+gender | 82.35
0.5	-15.0368	0.122 | 0.194
0.5	-15.0368	direction | (left/right/root)
0.142857	-15.0369	or | tags
0.12	-15.0371	more | complex
0.0879121	-15.0371	the | are computed
0.16	-15.0372	nullnullnullnullnullnull | null
0.0321678	-15.0373	given | we
0.0518519	-15.0375	features | more
0.0285088	-15.0375	as | where
0.0625	-15.0377	we | metrics
0.333333	-15.0383	for | processors
0.0635838	-15.0387	this | often
0.0349418	-15.0389	the | who | the
0.113208	-15.039	two | distance
0.5	-15.039	prevalent | cnet.com,
0.5	-15.039	tradition | 1992,
0.5	-15.039	definition) | topic
0.0369686	-15.0391	– | in
0.012389	-15.0391	that | with
0.0474684	-15.0392	text | al.,
0.135135	-15.0395	let | i
0.0879121	-15.0395	have | including
0.0445682	-15.0398	system, | of
0.0930233	-15.04	the | computed from
0.25	-15.0402	trigger | according
0.0205362	-15.0403	and | translation
0.0724638	-15.0404	i | l
0.0196993	-15.0404	were | to | the
0.0543933	-15.0412	4 | this
0.0264901	-15.0413	positive | the
0.25	-15.0413	basura | non-relevant
0.25	-15.0413	semclass | alias
0.25	-15.0413	microsoft | shop
0.25	-15.0413	apps | screen
0.25	-15.0413	d, | dnullsemc
0.25	-15.0413	λ1=0.2 | λ2=0.4
0.25	-15.0413	0.723 | nr
0.25	-15.0413	gpe | 0.9888
0.25	-15.0413	g855g94g286g374g400g286 | g855g94g286g374g400g286g3g4g454g349g400
0.25	-15.0413	points), | points),
0.25	-15.0413	[st | ]〉
0.25	-15.0413	{h11, | h41}
0.25	-15.0413	%fa30 | (5k)
0.25	-15.0413	opfinder | opfinder
0.125	-15.0413	hw | mj
0.5	-15.0413	26 | 19.35
0.5	-15.0413	p(z|w)
0.5	-15.0413	xin | xiaoming
0.0735294	-15.0417	to | level.
0.181818	-15.0417	(4) | all
0.147059	-15.0419	narrative | that
0.285714	-15.042	 | 
0.0689655	-15.042	as | model,
0.25	-15.0424	significant | at | =
0.06	-15.0424	2 | each
0.106061	-15.0424	space | features
0.0453258	-15.0426	of | token
0.0410959	-15.0427	of | language.
0.153846	-15.043	causality | in
0.142857	-15.0432	use | expressions
0.285714	-15.0433	related | paraphrases
0.0666667	-15.0434	since | all
0.0851064	-15.0435	vector | feature
0.0860215	-15.0435	does | require
0.25	-15.0435	performing | (step
0.25	-15.0435	dead | means
0.5	-15.0435	p(tnotarrowrighth|l
0.0305732	-15.0438	that | been
0.0860215	-15.0439	, | or
0.0389344	-15.0439	after | in
0.2	-15.0439	sig | **
0.0575221	-15.044	methods | be
0.0184332	-15.044	the | language | and
0.0438144	-15.0447	the | linguistics, pages
0.0532319	-15.0448	we then | a
0.0462428	-15.0452	2, | and
0.0205714	-15.0453	where | to
0.027689	-15.0455	of | shown
0.125	-15.0457	κ | pi
0.5	-15.0457	under | feature%1:03:00
0.5	-15.0457	s(lj|vi)
0.5	-15.0457	koc-combo | potsdam
0.5	-15.0457	y2y | )y
0.25	-15.0458	tag | get
0.25	-15.0458	pagar | pay)
0.0321229	-15.046	evaluation | for
0.0374753	-15.0462	consists | a
0.025974	-15.0467	work | for
0.135135	-15.0467	one | 2009).
0.102941	-15.0467	32 | and
0.0103285	-15.0467	in | to | to
0.25	-15.0469	11.9 | table
0.214286	-15.047	there | marked
0.113208	-15.0471	the | validation.
0.2	-15.0475	np-sb | agrnull
0.25	-15.048	unreordered | english
0.5	-15.048	2, | gent,
0.113208	-15.0481	it | supervised
0.0136605	-15.0481	not | in
0.0202361	-15.0482	as | which
0.0810811	-15.0482	some | data
0.0985915	-15.0484	one | whether
0.0345964	-15.0485	the | select
0.0658683	-15.0486	have | most
0.0507246	-15.0489	evaluation | et
0.0122374	-15.0493	a | for | and
0.109091	-15.0494	language | phrase-based
0.079646	-15.0495	participated | the
0.035461	-15.0495	the | meeting
0.285714	-15.0496	model, | used.
0.075	-15.0497	direct | is
0.0231214	-15.05	error | the
0.25	-15.0502	2001) | semantic
0.2	-15.0502	of | ner.
0.166667	-15.0503	acknowledgments | agency
0.128205	-15.0504	corresponds | one
0.181818	-15.0506	15th | oregon,
0.0334347	-15.0507	presented | of
0.107143	-15.0509	the | are | follows.
0.0546218	-15.0509	of | 2003).
0.0065128	-15.051	the | in | in
0.0485437	-15.0515	space | in
0.0198073	-15.0518	4 | of
0.173913	-15.052	∑ | αi,yi
0.2	-15.052	models | structure.
0.101449	-15.0521	4.2 | are
0.5	-15.0524	strong | rhythmic
0.0591133	-15.0525	if | null
0.0689655	-15.0525	while | such
0.1	-15.0529	e | +
0.2	-15.0529	significance | paired
0.0501792	-15.0532	by | phrase
0.0670732	-15.0532	with | distribution
0.0221328	-15.0532	model | on
0.0193286	-15.0533	for | this
0.285714	-15.0535	its | analysis.
0.0530612	-15.0535	described | et
0.0609137	-15.0538	sentences | were
0.1875	-15.0539	frank | x|sees
0.153846	-15.054	models, | can
0.0280778	-15.0543	use | and | to
0.0972222	-15.0545	however, this | is
0.0288351	-15.0546	we | how | the
0.153846	-15.0547	and | disjoint
0.5	-15.0547	role” | v-ing”
0.5	-15.0547	0.600.65 | bveptm
0.5	-15.0547	f) | (cmu)
0.5	-15.0547	ca, | pennac@yahoo-inc.com
0.0471698	-15.0548	in | (2)
0.0412621	-15.0551	we | novel
0.0364964	-15.0558	a | them
0.25	-15.0559	currencies | fever,
0.285714	-15.0561	features | value.
0.0310811	-15.0564	sentence | for
0.0314208	-15.0566	to | types
0.0550847	-15.0568	is | aligned
0.12	-15.0569	even | these
0.0369004	-15.0569	and | analysis
0.5	-15.0569	“ | per:title”
0.0824742	-15.0573	focus | with
0.0985915	-15.0573	the | batch
0.0522388	-15.0575	along | of
0.230769	-15.0578	better | value
0.0295203	-15.0579	this | it | to
0.0810811	-15.058	only | into
0.0565217	-15.0582	efficient | the
0.0898876	-15.0583	features | similar
0.0588235	-15.0587	we | instance
0.0553191	-15.0587	language | data
0.0562771	-15.0588	in | just
0.036965	-15.059	to | the | structure
0.0933333	-15.0592	find | each
0.5	-15.0592	boundary | s,
0.0163043	-15.0593	set | a
0.0154758	-15.0594	as | are
0.0756303	-15.0596	that | probabilities
0.0327381	-15.0596	this | semantic
0.0174564	-15.0598	a | more
0.0177259	-15.0599	a | one | the
0.0735294	-15.0601	metrics | as
0.12	-15.0602	pairs | given
0.0422886	-15.0602	work | this
0.0392157	-15.0604	not | also
0.0281385	-15.0606	then | is
0.333333	-15.0607	(8) | .
0.016424	-15.0608	number | in
0.147059	-15.0611	(adja | (nn
0.0161754	-15.0612	the | of | same
0.12	-15.0613	how | both
0.0281996	-15.0614	our | can
0.5	-15.0615	p(e,b|f)
0.5	-15.0615	p∈pwords | pwords
0.5	-15.0615	scenario | sup,
0.0299252	-15.0619	with | words
0.0775862	-15.0619	v | that
0.122449	-15.062	in | all.
0.0541667	-15.0622	will | our
0.166667	-15.0623	use | concepts
0.1875	-15.0624	nn | cc
0.0654762	-15.0626	types | on
0.0501792	-15.0628	+** | +**
0.0401786	-15.0635	in | research
0.0526316	-15.0635	to | compare
0.153846	-15.0638	they | multiple
0.113208	-15.0639	these | compared
0.0714286	-15.0641	related | work
0.104478	-15.0641	consider | of | =
0.0505415	-15.0642	model | one
0.057971	-15.0643	all | test
0.031856	-15.0643	as | their
0.107143	-15.0644	acknowledgments | authors
0.0152623	-15.0646	to | it
0.0419753	-15.065	e | t
0.0245139	-15.0656	at | is
0.0512821	-15.0656	objective | and
0.0958904	-15.0656	30 | 10
0.115385	-15.0657	were | according
0.5	-15.066	number | seconds)
0.5	-15.066	partitions | ω(c)
0.109091	-15.0661	also | noun
0.25	-15.0661	sdc, | binary,
0.0769231	-15.0662	to | gene
0.0909091	-15.0663	acknowledgments | funded
0.00954175	-15.0664	two | the
0.0155013	-15.067	task | the
0.0236967	-15.0671	the | up
0.0816327	-15.0671	the | international conference
0.131579	-15.0673	described | where
0.166667	-15.0676	models | language.
0.0409639	-15.0678	it | but
0.0635838	-15.0681	reviewers | their
0.0412621	-15.0682	that | pairs
0.0141353	-15.0683	are | and | of
0.0909091	-15.0683	set | label
0.0337079	-15.0683	the | entire
0.5	-15.0683	fromour | thismentiondetectionpart.
0.5	-15.0683	gap | '04
0.0167315	-15.0688	in | model
0.25	-15.0695	state | m−12
0.0606061	-15.0696	span | of
0.0671141	-15.0699	abstract | with
0.0909091	-15.07	5 | corpus
0.0294118	-15.0701	this | between
0.0543933	-15.0701	shown | al.,
0.16	-15.0704	75 | table
0.148148	-15.0705	values | figure
0.5	-15.0706	wide | (www
0.5	-15.0706	p(wk|tk)
0.5	-15.0706	discuss | ‘time’,
0.5	-15.0706	integraldisplay p(θ|α)
0.117647	-15.0708	computational | china
0.0363636	-15.0708	we | alignment
0.102941	-15.0709	lexical | based
0.113208	-15.0712	each | 100
0.0614525	-15.0714	would | not
0.0898876	-15.0714	then | same
0.230769	-15.0715	models | crf
0.0662252	-15.0719	it | phrase
0.181818	-15.072	64head | <>
0.166667	-15.0722	eyebrows | saw(i,
0.0860215	-15.0723	in | contribution
0.107143	-15.0723	where | human
0.026839	-15.0725	the | need
0.0427807	-15.0727	pairs | for
0.25	-15.0729	kn:bn-big | lm
0.5	-15.0729	dominates | thyme
0.5	-15.0729	weiwei | (dfki)
0.5	-15.0729	vossen, | eneko,
0.5	-15.0729	|y|
0.5	-15.0729	query-free | agent
0.0288018	-15.073	the | words in the
0.0268657	-15.0732	the | must
0.0412621	-15.0737	have | than
0.2	-15.0738	line | where
0.0576923	-15.0739	source | or
0.0116083	-15.075	this | of | and
0.1	-15.0752	also | model.
0.25	-15.0752	be | predictions.
0.5	-15.0752	(40 | bikes).
0.148148	-15.0753	tags | lexical
0.0426065	-15.0761	from | features
0.5	-15.0775	pasbio | bioprop
0.5	-15.0775	(not | m6
0.5	-15.0775	treebank | sierra
0.5	-15.0775	“� | (yamashiro)
0.5	-15.0775	formally, | y),
0.0487805	-15.0776	acknowledgments | for
0.0510949	-15.078	dependency | by
0.101449	-15.078	of | steps.
0.0921053	-15.0783	can | precision
0.0196829	-15.0786	as | the | by
0.148148	-15.0787	impact | user
0.0235294	-15.0789	in | will | the
0.070922	-15.0791	we refer | as
0.0958904	-15.0792	an | instance
0.1	-15.0798	work | domain
0.25	-15.0798	other | section.
0.125	-15.0798	node | ∈
0.5	-15.0798	v´eronique | 1lt3,
0.5	-15.0798	94.33 | 93.26
0.5	-15.0798	¯p(ti|ti−1)
0.115385	-15.08	classifier | only
0.0921053	-15.08	from | graph
0.230769	-15.08	lead | when
0.00601496	-15.0801	in | and | the
0.0288809	-15.0804	metric | the
0.1875	-15.0804	documents | annotated
0.0643275	-15.0806	to | included
0.0243902	-15.0806	the | recall
0.0298879	-15.0806	however, | the | is
0.0842105	-15.0809	also | order
0.214286	-15.0811	algorithm | way
0.173913	-15.0813	be | manner
0.128205	-15.082	by | tree.
0.0604396	-15.0821	did | for
0.230769	-15.0821	0, | word
0.2	-15.0821	for | below).
0.5	-15.0821	80.0 | 39.8
0.5	-15.0821	source/target | target/source
0.5	-15.0821	agent-verb | ag
0.0515873	-15.0822	the | referring
0.0762712	-15.0829	with | improvement
0.047619	-15.083	have | =
0.2	-15.083	that | getting
0.0699301	-15.0831	can | learning
0.25	-15.0833	rmd | score
0.0295567	-15.0836	to | model.
0.0321637	-15.0837	basic | the
0.03207	-15.0842	acknowledgments | the
0.5	-15.0844	p(z|f,α,β)
0.0418719	-15.0845	we | form
0.016956	-15.0855	from | for
0.25	-15.0856	m3 | m8
0.0607735	-15.0861	it | best
0.142857	-15.0861	average | standard
0.0236407	-15.0864	these | is
0.0585366	-15.0865	computational linguistics, | june
0.128205	-15.0867	such | less
0.5	-15.0867	half-court | stamina
0.5	-15.0867	“he | dt”,
0.0460123	-15.0871	can | seen
0.128205	-15.0873	important | translation
0.103448	-15.0874	from | include
0.0443213	-15.0884	source | with
0.0548523	-15.0884	as | better
0.0816327	-15.0889	in | linked
0.0273109	-15.0891	is | their
0.0395604	-15.0891	of | topics
0.5	-15.0891	(predicted | precision,gold
0.115385	-15.0897	an | ranking
0.0254777	-15.0899	the | statistical
0.153846	-15.09	french | translated
0.0714286	-15.0901	dep | dep
0.0357143	-15.0901	r: | r: | r:
0.0127128	-15.0902	from | a
0.0639535	-15.0903	term | for
0.113208	-15.091	5 | 40
0.333333	-15.0914	timeneed(x, | y
0.5	-15.0914	genome | microbial
0.0230415	-15.0915	example, | in
0.0530612	-15.0924	of | experiments.
0.0792079	-15.0928	table | gold
0.0259865	-15.0929	and | dependency
0.0334928	-15.093	in | baseline
0.055814	-15.093	we | question
0.0607735	-15.0932	additional | as
0.0261628	-15.0932	association | computational linguistics
0.16	-15.0933	show | effectiveness
0.5	-15.0937	classes | investigation.
0.5	-15.0937	assumption | (fung,
0.25	-15.0938	auto | auto
0.285714	-15.0941	relevant | entire
0.285714	-15.0941	performed | five
0.0528455	-15.0941	using | between
0.142857	-15.0941	this | incorporates
0.117647	-15.0942	is | ×
0.25	-15.0949	might | additional
0.0743802	-15.095	by | list
0.230769	-15.095	pairs | collected
0.0077831	-15.0952	of | is | of
0.0473186	-15.0954	from | parallel
0.105263	-15.0954	our | experimental
0.0276243	-15.0956	to | english
0.0179889	-15.0957	the | of | were
0.0879121	-15.096	and | six
0.5	-15.0961	0.33 | 25.69
0.0635838	-15.0964	our | parsing
0.0398126	-15.0967	4.2 | of
0.0125617	-15.0967	et | et al.,
0.138889	-15.0972	therefore | =
0.230769	-15.0972	unlabeled | scores
0.0775862	-15.0974	event | task
0.0103375	-15.0975	we | a | to
0.0128116	-15.0975	corpus | the
0.0178408	-15.0976	to | other
0.153846	-15.0979	number | made
0.039627	-15.098	at | on
0.0406699	-15.0983	evaluation | as
0.0479452	-15.0985	selected | a
0.0886076	-15.0985	mj | mi,
0.0543933	-15.0986	for | likely
0.0126611	-15.099	two | of
0.0842105	-15.099	this | give
0.0972222	-15.0993	show | use
0.16	-15.0993	which | training.
0.0428954	-15.0995	average | is
0.0684932	-15.0999	however, | between
0.0334395	-15.1	the | unlabeled
0.135135	-15.1001	it | performs
0.00540338	-15.1002	to | a
0.125	-15.1002	with | 2011).
0.117647	-15.1003	codes | in
0.0566038	-15.1004	pair | that
0.5	-15.1008	polanyi | zaenen
0.5	-15.1008	4.00% | -6.00%
0.5	-15.1008	f-score | (precision:
0.052	-15.101	in | obtain
0.0860215	-15.1011	each | represented | a
0.135135	-15.102	(or | which
0.101695	-15.1021	 | 
0.1875	-15.1023	disfluencies, | and
0.333333	-15.1024	avg | clust,
0.0262097	-15.1024	we | algorithm
0.0860215	-15.1027	framework | by
0.0985915	-15.1028	with | named
0.0725806	-15.1029	performance | only
0.0274725	-15.1029	the | of | large
0.0484429	-15.103	us | for
0.5	-15.1031	|t|summationdisplay
0.5	-15.1031	radicalbigsummationtext | radicalbigsummationtext
0.027933	-15.1032	over | a
0.0291971	-15.1033	of | class
0.038055	-15.1033	for | computer
0.109091	-15.1036	number | about
0.0532787	-15.1038	the | generative
0.333333	-15.104	kul | sp
0.0484429	-15.1041	approaches | on
0.173913	-15.1041	of | bb
0.0680272	-15.1044	we | distributional
0.0125945	-15.1046	the | these
0.0666667	-15.1047	pos | by
0.072	-15.1047	are | able
0.038961	-15.1049	to | negative
0.176471	-15.1053	– | 59.8
0.103448	-15.1054	if | while
0.0566038	-15.1055	systems | by
0.142857	-15.1055	verb | arguments
0.0200803	-15.1055	on | by
0.5	-15.1055	ter, | rate,
0.5	-15.1055	swiss | nccr
0.5	-15.1055	attendedschool | hasbirthplace
0.5	-15.1055	∀wi−1 | ∈φn(wi−11
0.5	-15.1055	gain | ∆ux,y
0.5	-15.1055	5.1), | 5.3).
0.0410628	-15.1057	and | grammar
0.1875	-15.1059	observe | sentences
0.0359168	-15.106	we | function
0.0384615	-15.1066	also | al.,
0.072	-15.1068	or | feature
0.0484429	-15.1069	has | word
0.0218878	-15.1073	of | lexical
0.0229709	-15.1075	that | but
0.142857	-15.1076	found | sentences
0.0635838	-15.1078	we | summationdisplay | summationdisplay
0.5	-15.1079	element | <archdesc>
0.5	-15.1079	{(xi,zi)|i
0.0958904	-15.108	if | very
0.0649351	-15.1082	cases, | to
0.173913	-15.1082	primarily | as
0.0202501	-15.1086	to | sentence
0.115385	-15.1087	• | i
0.109091	-15.1088	thing | the
0.25	-15.1091	will | unless
0.0607735	-15.1093	functions | a
0.2	-15.1093	proportional | probability
0.173913	-15.1099	empty | categories
0.138889	-15.1103	one | instead
0.230769	-15.1103	component | has
0.5	-15.1103	disk | mbytes,
0.2	-15.1112	and | (2007))
0.0174216	-15.1113	translation | and
0.0298701	-15.1115	shown | in
0.333333	-15.1119	vbn | rbr
0.214286	-15.1123	correct | 60
0.5	-15.1126	pmle | a∈leaf(e)
0.0297542	-15.1127	parallel | and
0.176471	-15.1128	give | sentence
0.0240757	-15.113	may | a
0.0860215	-15.1133	word | noun
0.103448	-15.1133	number | query
0.214286	-15.1133	will | interesting
0.047138	-15.1137	we | the | models.
0.0352505	-15.1144	may | as
0.176471	-15.1145	k | stream
0.0352505	-15.1149	et | et | 2010),
0.2	-15.115	shows | vs.
0.5	-15.115	-0.77 | -1.04
0.5	-15.115	extend | high-level
0.5	-15.115	ondˇrej | herv´e
0.5	-15.115	w1,w2,w3 | 3(a))
0.5	-15.115	ruppenhofer | michaelis
0.128205	-15.1152	approach | high
0.0391705	-15.1152	input | is
0.138889	-15.1156	→ | ad
0.0337268	-15.1159	structures | the
0.0816327	-15.1159	section | will
0.176471	-15.1162	correlation | predicted
0.0550459	-15.1164	table | average
0.173913	-15.1165	dc cwcxbnczcx | dc
0.00811069	-15.1169	are | the | of
0.0737705	-15.1172	and/or | the
0.00436765	-15.1174	and | of
0.5	-15.1174	(surdeanu | fung,
0.5	-15.1174	sdl | 6060
0.5	-15.1174	stubbs | massachusetts,
0.5	-15.1174	p(vectorlscript|vectort,vectorα)
0.0641026	-15.1177	was | grant
0.0560748	-15.1178	for | patterns
0.107143	-15.118	∈ | i
0.0274424	-15.1187	the | include
0.25	-15.1187	rows | models
0.0465116	-15.1188	we | improvement
0.0547945	-15.1191	as | vector
0.148148	-15.1191	science | ny
0.0792079	-15.1195	have | 2008;
0.0386364	-15.1196	proceedings of | conference
0.5	-15.1198	property, | heterogeneity
0.131579	-15.1203	log | number
0.0148291	-15.1204	and | each
0.0391705	-15.1204	1 | an
0.0731707	-15.1205	one | corpus
0.142857	-15.1205	would | valuable
0.214286	-15.1205	dependency | between | head
0.0972222	-15.1206	between | documents
0.153846	-15.1206	penn | into
0.0137855	-15.1207	the | source
0.0150956	-15.1209	of | features
0.119048	-15.1211	task | 2011).
0.047138	-15.1212	to | extract
0.230769	-15.1213	的 | 歌舞
0.230769	-15.1213	(forall | (changefrom
0.230769	-15.1213	53697465 | 44756d6d79
0.0540541	-15.1217	2. | we
0.109091	-15.1218	the | tag;
0.028169	-15.122	we | first
0.5	-15.1222	klaren | drehen
0.0496183	-15.1223	a | features,
0.285714	-15.1227	a72a72a8a8a8 | vvf
0.0675676	-15.1229	we | algorithm.
0.0334448	-15.123	x | a
0.04	-15.1232	were | the | were
0.0909091	-15.1234	phrase | have
0.0315315	-15.1236	from | can
0.16	-15.1237	emotion | emails
0.0382979	-15.1238	knowledge | to
0.128205	-15.1238	detailed | table
0.0515873	-15.1241	our | similar
0.119048	-15.1245	using | likelihood
0.5	-15.1246	xi | x1i
0.166667	-15.1247	ap | q
0.0675676	-15.1249	this | might
0.0275938	-15.125	we | learning
0.0515873	-15.1251	note | this
0.1	-15.1252	more | well
0.0272628	-15.1252	found | of
0.0240964	-15.1255	we | over
0.0588235	-15.1255	correctly | and
0.0377358	-15.1255	is | time
0.0614525	-15.1257	this | task.
0.0776699	-15.1259	to | image
0.0284698	-15.1269	the | and | those
0.5	-15.1271	14.15 | 3.66
0.5	-15.1271	p(wi|si)
0.5	-15.1271	agr) | approp(v,
0.5	-15.1271	stimulus | scene,
0.5	-15.1271	accommodate | medicines
0.5	-15.1271	0.085 | 0.204
0.5	-15.1271	1.influenza | 1468
0.0792079	-15.1275	sentences | only
0.0200354	-15.1278	only | a
0.025974	-15.1282	first, | the | the
0.00973033	-15.1285	are | a
0.0762712	-15.1288	where | been
0.0530973	-15.1292	as | corpus.
0.0958904	-15.1292	| | =
0.1875	-15.1293	predicate | arg2
0.00678399	-15.1293	a | is | the
0.25	-15.1295	er | ound
0.333333	-15.1295	constraining | model
0.5	-15.1295	rush, | jaakk
0.5	-15.1295	nullnullnullnullnull | nullnullnullnullnull,
0.5	-15.1295	manualsource | qualify
0.0636943	-15.1296	these | sentences
0.285714	-15.1296	confusion network | confusion
0.0498084	-15.1298	distance | in
0.0731707	-15.13	as | recall
0.0362903	-15.1303	is | model.
0.0692308	-15.1307	2007), | of
0.25	-15.1308	languages | language,
0.0555556	-15.1312	model | their
0.121951	-15.1314	due | time
0.0347985	-15.1319	(1) | is
0.5	-15.1319	(at | point)
0.5	-15.1319	simplification | porsimples
0.052	-15.1322	syntactic | with
0.101695	-15.1329	language | kneser-ney
0.0671141	-15.1333	are | x
0.0792079	-15.1333	of | series
0.0611111	-15.1334	and | entire
0.075	-15.1335	we | methods.
0.0426667	-15.1339	are | 1
0.0160305	-15.1342	in | have
0.0776699	-15.1342	cosine | between
0.5	-15.1344	transport | multicellular
0.5	-15.1344	muc | 61.39
0.0792079	-15.1345	similar | an
0.285714	-15.1351	extended | relations.
0.0443787	-15.1353	by | sentences
0.0294118	-15.1356	data | on
0.0601093	-15.1357	and | along
0.0611111	-15.1358	each | with | with
0.0335008	-15.1359	their | is | the
0.128205	-15.1363	data, | features
0.109091	-15.1366	such | performance
0.5	-15.1368	−qj|
0.5	-15.1368	engines | confused
0.5	-15.1368	pr(ci|s,t)
0.5	-15.1368	illness: | cancer.
0.0453172	-15.137	a | of | could
0.0239521	-15.1371	probability | of
0.101695	-15.1373	5.3 | on
0.0299728	-15.1374	three | a
0.0681818	-15.1376	we | surface
0.0215577	-15.138	will | and
0.128205	-15.1381	can | processing
0.0933333	-15.1383	were | corpus.
0.0853659	-15.1383	approach | machine
0.103448	-15.1388	show | methods
0.0307692	-15.1388	the best | of
0.0641026	-15.1391	feature | between
0.5	-15.1392	p(synf|syne,p)
0.5	-15.1392	olisi | viime
0.056701	-15.1396	it | target
0.0287141	-15.1397	apply | the
0.113208	-15.1398	shown | models
0.0479452	-15.14	the | half
0.0614525	-15.1401	system | other
0.0324149	-15.1401	of | 0
0.1875	-15.1402	for | scenarios
0.0152039	-15.1407	of | training
0.05	-15.1407	in | morphological
0.333333	-15.1409	sicakhata/ | 'start/
0.176471	-15.141	whether | relevant
0.0675676	-15.1413	in | method.
0.119048	-15.1414	related | 2007;
0.153846	-15.1416	effects | from
0.5	-15.1417	r−→r|tnotarrowrighth)
0.5	-15.1417	(compared | 2)
0.5	-15.1417	prefaces, | formulations,
0.0243024	-15.1418	if | are | the
0.1	-15.1418	the | reached.
0.333333	-15.1426	(y/n); | mi/j;
0.00916625	-15.143	for | and | of
0.2	-15.1432	j | @
0.0405063	-15.1435	while | with
0.0692308	-15.1439	with | second
0.0354478	-15.1442	by | data
0.0886076	-15.1442	to | meaningful
0.5	-15.1442	particular | !!
0.5	-15.1442	24.16 | 24.18
0.0628931	-15.145	classifier | from
0.0454545	-15.1451	in | combination
0.0386364	-15.1454	automatic | to
0.026738	-15.1457	data | as
0.0933333	-15.1465	p | where
0.0921053	-15.1465	do | than
0.5	-15.1466	sys+sim | reversers.
0.0395349	-15.1469	to | values
0.2	-15.1471	expressions | opinion
0.148148	-15.1473	found | less
0.131579	-15.1473	recent | translation
0.0466667	-15.1474	this paper | as
0.333333	-15.1475	wds+bi | wds+ac
0.333333	-15.1475	reviewers | comments,
0.125	-15.1479	scope | for
0.116279	-15.1481	above | two
0.0662252	-15.1482	each | new
0.0555556	-15.1483	in | train
0.0933333	-15.1484	use | original
0.0737705	-15.1485	state | be
0.0886076	-15.1485	this | experiments.
0.0191361	-15.1491	event | the
0.0174873	-15.1491	there | a
0.0550459	-15.1491	address | to
0.0424929	-15.1491	speech | in
0.5	-15.1491	sie:it | früh:early
0.5	-15.1491	task1, | otherwise).
0.0886076	-15.1494	cannot | with
0.101695	-15.1498	present | use
0.101695	-15.1498	conclusion | system
0.0945946	-15.1499	has | while
0.0151515	-15.1502	from | of | and
0.333333	-15.1508	specific | chosen
0.0227095	-15.1515	we | two | of
0.5	-15.1516	mentions, | singletons.
0.5	-15.1516	ψ|−i,e
0.5	-15.1516	p(t,γ,φ,τφ,a|s,τs)
0.103448	-15.1521	feature | are | with
0.0262329	-15.1534	are | words
0.116279	-15.1539	0.36 | 0.36
0.0613497	-15.1539	• | an
0.25	-15.1541	method | splitting
0.0441176	-15.1543	classifier | in
0.0511811	-15.1552	word | two
0.00953808	-15.1554	for | the | for
0.0131648	-15.1555	of | data
0.0237893	-15.1555	much | the
0.0131961	-15.1556	this | of | in
0.137931	-15.1563	solidtriangletriangle | solidsquaresquare
0.5	-15.1565	readymade | evocative
0.5	-15.1565	10000 | 16000
0.5	-15.1565	manually | connective.
0.0540541	-15.1567	we | hierarchical
0.0488722	-15.1567	based | model
0.0100733	-15.1569	in | a | a
0.0625	-15.1569	set | these
0.0645161	-15.1571	is | 2006).
0.0344203	-15.1576	is | at | and
0.02875	-15.1584	a | with | that
0.176471	-15.1585	recall | –
0.0121092	-15.1588	2 | the
0.0423729	-15.1588	to | research
0.0535714	-15.1589	we can | =
0.0111489	-15.159	in | from
0.166667	-15.1591	70 | vn
0.0776699	-15.1591	feature | only
0.137931	-15.1591	split | one
0.333333	-15.1591	cerrar | coger
0.333333	-15.1591	dev/test | rmm
0.333333	-15.1591	soon | 46.18
0.333333	-15.1591	windowsize2 | mentionsin
0.333333	-15.1591	wall, | wall)
0.333333	-15.1591	73.1 | noisy-seed)
0.333333	-15.1591	59% | base
0.333333	-15.1591	xpv | (vp1
0.333333	-15.1591	slovak | slovakia
0.333333	-15.1591	w3 | 3(c)).
0.333333	-15.1591	0.013 | 0.841
0.333333	-15.1591	salicylate | adhesion
0.333333	-15.1591	0.50.6 | false
0.333333	-15.1591	mwen | [firstname][lastname]
0.333333	-15.1591	pc | ar(pc,
0.333333	-15.1591	tk | dki
0.333333	-15.1591	c320 | 19.84
0.333333	-15.1591	in; | vbd
0.333333	-15.1591	sig | -0.20
0.333333	-15.1591	#tokens | 201m
0.333333	-15.1591	,x | )⊗b(x
0.333333	-15.1591	athleteplays | teamplays
0.333333	-15.1591	24.92 | 23.37
0.333333	-15.1591	wmt | sixteen
0.333333	-15.1591	ci( | mentionsin
0.333333	-15.1591	ptr | fba
0.333333	-15.1591	fi | flx
0.333333	-15.1591	th(null)nullgirlnull)null | plural
0.333333	-15.1591	c2(x | c2(y)
0.333333	-15.1591	94.4% | 73.9%
0.333333	-15.1591	lhs | 
0.333333	-15.1591	intent | wonderland
0.333333	-15.1591	dershowitz | israel
0.333333	-15.1591	== | counter(bin(|dm(qi)|))++;
0.333333	-15.1591	88.06 | 88.06
0.333333	-15.1591	coref | ant:t31
0.333333	-15.1591	robv | (tear
0.333333	-15.1591	morph | false
0.333333	-15.1591	p(d | |d
0.333333	-15.1591	parenleftbigg | vuniondisplay
0.333333	-15.1591	(if | e)(rexist
0.333333	-15.1591	sodium | adhesion
0.333333	-15.1591	interaction(during(e), | frames
0.333333	-15.1591	h14,3 | epsilon1
0.333333	-15.1591	 | 不幸
0.333333	-15.1591	〈bl | alp
0.333333	-15.1591	λmpm(wi|φm) | φm
0.333333	-15.1591	type="open" | <target
0.333333	-15.1591	40.47 | 61.76
0.333333	-15.1591	sources | msnbc
0.333333	-15.1591	// | d3
0.333333	-15.1591	l´a | only),
0.333333	-15.1591	msd-hsmm | synthesis
0.333333	-15.1591	0.13, | software
0.333333	-15.1591	dm(q′i) | swap
0.333333	-15.1591	ser(727) | <site
0.333333	-15.1591	conn+ | 91.2
0.333333	-15.1591	companies | accessories
0.333333	-15.1591	(train) | (µ/σ2)
0.333333	-15.1591	a50diamondsolidtriangle | a50♦triangle
0.333333	-15.1591	arrownortheast | arrownortheast
0.333333	-15.1591	uninspired | romance
0.333333	-15.1591	(x,y)) | yprime
0.333333	-15.1591	bracketleftbig | ←t
0.333333	-15.1591	rec% | rec%
0.333333	-15.1591	beta | mitogenic
0.333333	-15.1591	oracle-oov | +2%
0.333333	-15.1591	conn+conn | (1237)
0.333333	-15.1591	a29a36a39 | a51
0.333333	-15.1591	null,2 | null5
0.333333	-15.1591	offiziel | pt,
0.333333	-15.1591	yif | semanticclass
0.333333	-15.1591	w(xi,yj) | ssummationtext
0.333333	-15.1591	martini | cake
0.333333	-15.1591	cr〉 | arp
0.333333	-15.1591	(−lc | (if
0.333333	-15.1591	splitting-order | marker
0.333333	-15.1591	wjh | wie
0.333333	-15.1591	(s\np)/np | >bs/np
0.333333	-15.1591	mwen | lavel
0.333333	-15.1591	=== | sink
0.333333	-15.1591	arr | nums<>hartu
0.333333	-15.1591	‘to | gaya
0.333333	-15.1591	|l|→∞ | ind
0.333333	-15.1591	(grf) | greatermuch
0.333333	-15.1591	love | {tweetn}
0.333333	-15.1591	m/f | m/f/n
0.333333	-15.1591	:[d,m→h] | ,h,x
0.333333	-15.1591	|}1|||{| | ji
0.333333	-15.1591	pre% | binary+patterns
0.333333	-15.1591	c320 | cs-en
0.333333	-15.1591	glyph36glyph34glyph45 | glyph1
0.333333	-15.1591	z∗(j)q | z∗(j)q
0.333333	-15.1591	gerund | “downstairs”
0.333333	-15.1591	mcore | (subsystem
0.333333	-15.1591	0.286 | reo
0.333333	-15.1591	c320 | 11.92
0.333333	-15.1591	eng | 500
0.333333	-15.1591	tk | ntlabel),
0.333333	-15.1591	mentionsin | singularorplural,determinedusinga
0.333333	-15.1591	that‟ | "wall
0.333333	-15.1591	cat | cat”
0.333333	-15.1591	lex | rhs
0.333333	-15.1591	grams | 0.535
0.333333	-15.1591	“a-10 | (subtype
0.333333	-15.1591	wwc | ),(
0.333333	-15.1591	peacekeeping | rebel
0.333333	-15.1591	nullnullnullnullnullnullnull(null) | ∃null
0.333333	-15.1591	nmod | oppose
0.333333	-15.1591	∆i | ×0.7
0.333333	-15.1591	forteam | inleague
0.333333	-15.1591	≡ | i(ci,j
0.333333	-15.1591	stat3 | rac-1
0.333333	-15.1591	70.3 | opfinder
0.333333	-15.1591	there), | s´o
0.333333	-15.1591	w2” | oil”
0.333333	-15.1591	p(θ|x | p(θ|x
0.333333	-15.1591	|y | ).
0.333333	-15.1591	<pat | <w
0.333333	-15.1591	%) | (18
0.333333	-15.1591	angbracketrightbig | lmbot
0.333333	-15.1591	b-hmm | 61/74
0.333333	-15.1591	cadet | manje
0.333333	-15.1591	shift | with(eyebrows)
0.333333	-15.1591	pk,·+vβ | constraint
0.333333	-15.1591	sim-s | sim-k
0.333333	-15.1591	bo | a0bd
0.333333	-15.1591	“slept” | 1933
0.333333	-15.1591	suffixes | alia
0.333333	-15.1591	order,vr | ={a0,
0.333333	-15.1591	foreach | p(γi|h)
0.333333	-15.1591	s50s48 | s51
0.333333	-15.1591	parenleftbig | k∈out(vj)|pjk|
0.333333	-15.1591	spector, | ronnie
0.333333	-15.1591	null=nullnullnull | “john
0.333333	-15.1591	19.58 | 1.83
0.333333	-15.1591	(%) | 3899
0.333333	-15.1591	κ0, | κ0
0.333333	-15.1591	.30 | .42
0.333333	-15.1591	infobox | <sears_tower,
0.333333	-15.1591	cu | 95.9
0.333333	-15.1591	pobj | tmod
0.333333	-15.1591	a159 | a166
0.333333	-15.1591	connotation | war).
0.333333	-15.1591	
 | pu
0.333333	-15.1591	sm | _#
0.333333	-15.1591	¡ | 0::b0
0.333333	-15.1591	does | pas〉.
0.333333	-15.1591	0.1462 | pseudo
0.333333	-15.1591	require: | alli∈pdo
0.333333	-15.1591	γi(y) | uj∈n(ui)
0.333333	-15.1591	89.07 | 88.58
0.333333	-15.1591	losing”, | lose”,
0.333333	-15.1591	000 | 1,17
0.333333	-15.1591	x∈t | x∈t,r∈{x,∅}
0.333333	-15.1591	 | −→
0.333333	-15.1591	f) | s-feature
0.333333	-15.1591	agrees | )v
0.333333	-15.1591	hd-sv-lv-ex | head-complement
0.333333	-15.1591	catabolism | complexassembly
0.333333	-15.1591	�c | �
0.333333	-15.1591	a157 | a170
0.333333	-15.1591	0.322 | 52
0.333333	-15.1591	(tendency | (one
0.333333	-15.1591	apply | ecur
0.333333	-15.1591	3:59 | (4)
0.333333	-15.1591	bits | 15176585
0.333333	-15.1591	formally: | subsetsqequal
0.333333	-15.1591	martinis | carrot
0.333333	-15.1591	7628 | 7628
0.333333	-15.1591	moses | 24.4765
0.333333	-15.1591	szpakowicz, | girju
0.333333	-15.1591	shift | with(eyebrows))
0.333333	-15.1591	vv | 10−20
0.333333	-15.1591	haibanaahuaakottaige | kottaige
0.333333	-15.1591	mediates | interacts
0.333333	-15.1591	arg1, | (clause
0.333333	-15.1591	(conjunction) | conn+conn
0.333333	-15.1591	#docs | 4,830
0.333333	-15.1591	prop | morp
0.333333	-15.1591	7.9 | 4gr+ibm2
0.333333	-15.1591	rk | (s,n)
0.111111	-15.1591	 | bracketrightbig
0.125	-15.1591	future | work
0.0399002	-15.1596	summationdisplay | =
0.166667	-15.16	θ | ,y
0.0366599	-15.1601	a | model | on
0.222222	-15.1602	pos-4grams | pos-3grams
0.222222	-15.1602	c., | c.,
0.222222	-15.1602	absolutive | selecting
0.222222	-15.1602	(%), | precision
0.0420168	-15.1606	we | phrases
0.333333	-15.1607	corr(giza++) | corr(hmm)
0.333333	-15.1607	thesaurus | lexicon
0.333333	-15.1607	aki | alj.
0.0454545	-15.1611	it | translation
0.0945946	-15.1612	not | sentence.
0.222222	-15.1613	llbracketp;errbracketw | llbracket〈p;e;
0.153846	-15.1614	mooney, | 2006;
0.5	-15.1615	lasige | lisbon,
0.137931	-15.1619	missing | with
0.121951	-15.1621	a | largely
0.0864198	-15.1622	they | been
0.0137025	-15.1623	the | most
0.0853659	-15.1624	words | would
0.333333	-15.1624	timex3 | event–
0.222222	-15.1625	queries | (sender)
0.222222	-15.1625	: | 
0.0281174	-15.1631	that | performance
0.2	-15.1631	$word | =
0.0967742	-15.1633	wikipedia | al.,
0.047619	-15.1636	e | o
0.0909091	-15.1637	consider | not
0.0761905	-15.164	other | was
0.333333	-15.1641	nullnull4) | nullnull3)
0.0510638	-15.1649	approach | model
0.0110466	-15.1652	sentence | the
0.109091	-15.1653	scheme | by
0.0359281	-15.1653	our | language
0.039312	-15.1655	this | section
0.0348624	-15.1655	with | translation
0.0193732	-15.1656	at | a
0.125	-15.1659	span, | the
0.0875	-15.166	data | model.
0.0649351	-15.1663	of | until
0.176471	-15.1665	null | represents
0.0465116	-15.1666	distribution of | of
0.0492424	-15.1666	rules | be
0.5	-15.1666	noun-phrase | verb-phrase
0.5	-15.1666	side | poster’s
0.5	-15.1666	letters | apostrophe,
0.5	-15.1666	(estan | networking.
0.0453074	-15.1671	figure | shows the
0.333333	-15.1674	gemeinsames, | extremen
0.285714	-15.1677	in | runs,
0.0219459	-15.1677	a | source
0.0298103	-15.1678	in | three
0.0864198	-15.1679	was | best
0.0339893	-15.1683	models | al.,
0.0515021	-15.1685	0.4 | 0.8
0.047138	-15.1688	a | o
0.5	-15.1691	fujiki, | blackstone
0.0356436	-15.1693	in | distribution
0.137931	-15.1695	represented | features
0.137931	-15.1702	2 | question
0.0409207	-15.1703	with | these
0.0436047	-15.1704	2, | of
0.039801	-15.1705	applied | a
0.101695	-15.1707	if | still
0.0792079	-15.1707	pairs | data
0.333333	-15.1708	|e)
0.0433526	-15.171	we | instances
0.142857	-15.1713	scores | task.
0.222222	-15.1714	identify | intended
0.222222	-15.1714	0.001 | 0.01
0.5	-15.1716	binary-valued | min(
0.5	-15.1716	p(w|t,r
0.5	-15.1716	horse | raced
0.0463576	-15.1718	context | as
0.0259336	-15.1719	this is | to
0.285714	-15.1721	tg | (scenes
0.107143	-15.1724	top | table
0.222222	-15.1725	macrof1 | chinese
0.0761905	-15.1725	to | representation.
0.0875	-15.1726	beam | to
0.25	-15.1729	〉,int | modules.
0.0463576	-15.1732	column | of
0.285714	-15.1735	2 | i.e.,
0.285714	-15.1735	class, | may
0.333333	-15.1742	”olive | ”oil”,
0.5	-15.1742	arresting | dictator
0.0192853	-15.1746	we | of | which
0.0671642	-15.1746	of | window
0.103448	-15.1749	cases | than
0.0740741	-15.1752	by | sequence
0.0621118	-15.1754	2, | is
0.0294511	-15.1755	t | a
0.039604	-15.1756	2 | 5
0.222222	-15.1759	lex-syntax | clm
0.285714	-15.1764	mi/j; | mi/j
0.0816327	-15.1764	this | version
0.5	-15.1767	kentaro | sadao
0.5	-15.1767	f(eij|si)
0.5	-15.1767	authoritativeness | labelled
0.5	-15.1767	abductive | (hobbs
0.5	-15.1767	amount | acount
0.0909091	-15.177	more | so
0.0897436	-15.177	can | general
0.222222	-15.1771	ro | ro
0.2	-15.1772	german | sentences
0.0222393	-15.1773	the | 2010).
0.0107894	-15.1775	by | in
0.333333	-15.1776	th | ecen
0.0921053	-15.1779	prior | be
0.0967742	-15.178	using | entity
0.0131791	-15.1781	use | and
0.121951	-15.1781	size | different
0.0387409	-15.1782	we | problem | the
0.107143	-15.1789	we | cross-validation
0.214286	-15.1789	provide | target
0.00900225	-15.179	we | and | and
0.5	-15.1792	steinberger, | kabadjov,
0.5	-15.1792	|φ∩ψ|
0.5	-15.1792	transcription | p65
0.222222	-15.1793	according | manual
0.0591398	-15.1794	are | either
0.0636943	-15.1797	phrases | be
0.0109665	-15.1797	the | other
0.0671642	-15.1797	- | on
0.0921053	-15.1797	0 | w
0.0526316	-15.1803	the | of | aspects
0.0376106	-15.1805	to | gold
0.116279	-15.1809	connectives | that
0.0103696	-15.181	each | of
0.1875	-15.1816	r4) | squaresmallsolid
0.0352941	-15.1817	certain | the
0.5	-15.1818	5.87% | 3.27%,
0.5	-15.1818	relational | (getoor
0.5	-15.1818	mc | depth/density
0.0366379	-15.1821	technique | the
0.137931	-15.1821	means | has
0.0363248	-15.1826	the resulting | the
0.0193095	-15.1832	this | by | of
0.0526316	-15.1836	many | not
0.0547264	-15.1839	such | between
0.0692308	-15.1843	for | define
0.1875	-15.1844	laboratory | university
0.333333	-15.1844	types: | (accepted
0.333333	-15.1844	training+dev | (28
0.5	-15.1844	···wn) | whilecandidatearcs.isnotempty()do
0.0687023	-15.1846	the | pp.
0.0524017	-15.1847	was | data
0.0365591	-15.1851	table | performance
0.0462046	-15.1852	al., | 2010),
0.0747664	-15.1852	to | (a
0.333333	-15.1861	(tesauro | galperin,
0.333333	-15.1861	23.81 | 31.53
0.0582011	-15.1864	quality | is
0.0512821	-15.1866	probabilities | to
0.5	-15.1869	≈ | rows.
0.5	-15.1869	¨a | @æ
0.0219365	-15.187	model | that
0.0636943	-15.1872	ranked | to
0.108696	-15.1872	three | described
0.0232158	-15.1873	is | sentence
0.0172013	-15.1875	we | as | in
0.0886076	-15.1883	function | word
0.0322061	-15.1885	we | syntactic
0.0676692	-15.1888	we | textual
0.0886076	-15.1888	greatly | of
0.0519481	-15.1891	all | model
0.0234579	-15.1893	every | the
0.0552764	-15.1894	results of | are
0.0407609	-15.1894	to | node
0.333333	-15.1895	relations | signaled
0.5	-15.1895	+poslm | +loclm
0.5	-15.1895	2010b; | k¨otzing,
0.5	-15.1895	mpf | posbleu
0.0539216	-15.1898	we | tasks
0.0457516	-15.1898	probability | as
0.0595238	-15.1899	figure | accuracy
0.0111005	-15.19	the | of the | for
0.0336879	-15.1902	a | extracted
0.0124755	-15.1903	can | in
0.0115595	-15.1903	but | the
0.0628931	-15.1907	more | translation
0.013868	-15.1913	this | is | of
0.056701	-15.1915	the | 2002;
0.0255624	-15.1915	a | test
0.0591716	-15.1916	as | shown
0.5	-15.1921	length | j|.
0.5	-15.1921	msa’ | alsbt
0.00334588	-15.1923	of | of
0.0613497	-15.1925	form | be
0.0253807	-15.1927	a | learning
0.0486891	-15.1927	is | present
0.08	-15.1928	may | our
0.0383693	-15.1929	and | ,
0.0875	-15.1932	similarity | target
0.0833333	-15.1934	not | useful
0.0776699	-15.1935	= | 4
0.0328152	-15.1936	2 | that
0.0433437	-15.1936	were | not
0.0864198	-15.1939	of | sizes
0.125	-15.194	the | world,
0.0364026	-15.1943	even | is
0.176471	-15.1944	96 | in
0.0387409	-15.1945	extracted | in
0.0613497	-15.1947	m | i
0.1	-15.1947	allows | different
0.0462633	-15.1947	the | organized
0.333333	-15.1947	1|q| | ⃗fi
0.5	-15.1947	walde, | vlachos
0.5	-15.1947	|q|
0.5	-15.1947	enrolled | engineering
0.5	-15.1947	p(tildewidee,|˜f,f)p(˜f|f)
0.027512	-15.1948	figure | with
0.0287206	-15.1951	as | language
0.222222	-15.1953	np-c | vbd+vbn+p
0.0284605	-15.1953	clustering | the
0.0363248	-15.1954	their | that
0.214286	-15.1954	6879 | 6880
0.214286	-15.1954	� | �t
0.011267	-15.1957	our | in
0.133333	-15.1958	the | processor
0.0588235	-15.1959	these | same
0.0477941	-15.196	document | as
0.0636943	-15.1968	in | complexity
0.0632911	-15.1971	top | is
0.5	-15.1973	ccl | clumper
0.5	-15.1973	refering | (a.o.,
0.5	-15.1973	microblog | shou-de
0.5	-15.1973	10.8 | v/s
0.166667	-15.1974	to | doc-comb
0.0555556	-15.1976	by | order
0.0153214	-15.198	the | would
0.0625	-15.1983	makes | for
0.128205	-15.1983	<tree | </tree>
0.0185387	-15.1984	at | in
0.0394089	-15.1985	we | corresponding
0.129032	-15.1986	amount | large
0.0426136	-15.199	that | them
0.0204756	-15.1991	we | one | the
0.0164218	-15.1995	the | syntactic
0.0133751	-15.1996	have | to
0.333333	-15.1999	dr∈ldr | er∈ler
0.333333	-15.1999	lieber | mann
0.5	-15.1999	0.059 | 0.063
0.166667	-15.2	any | words.
0.0313972	-15.2007	we | better
0.0909091	-15.2009	| | (
0.0138638	-15.201	more | of
0.0105042	-15.2012	for | that
0.108696	-15.2013	models | set.
0.333333	-15.2016	on, | 3g4,
0.0291262	-15.2022	we show | the
0.222222	-15.2023	represents | results
0.5	-15.2025	‘m’ | ‘e’.
0.5	-15.2025	holes | kinds
0.5	-15.2025	filled | db:2
0.0230769	-15.2027	target | and
0.285714	-15.2029	dependencies | words.
0.142857	-15.2029	table | id
0.0909091	-15.2032	on | 2009)
0.108696	-15.2035	systems | by | et
0.0681818	-15.2039	no | it
0.111111	-15.2039	even | information
0.0875	-15.2041	sentence | verb
0.016223	-15.2041	however, | of
0.285714	-15.2044	graph-based | x
0.0340909	-15.2047	an | between
0.0625	-15.2048	be | high
0.0180319	-15.2048	work | to
0.116279	-15.2049	.1 | 0
0.0707965	-15.205	if | s
0.5	-15.2051	(watts | strogatz,
0.222222	-15.2057	tierra | y
0.222222	-15.2057	a, | b
0.0444444	-15.2057	is | pos
0.0360169	-15.2061	in | make
0.133333	-15.2062	a | wants
0.0681818	-15.2063	language | system
0.222222	-15.2069	sentiment | representations
0.0628931	-15.2071	it | similar
0.0194293	-15.2072	the | found
0.116279	-15.2074	hypothesis | each
0.0561224	-15.2076	scores are | the
0.0243446	-15.2076	in | models
0.5	-15.2077	p(wdn|βdni)
0.5	-15.2077	noch | gericht
0.5	-15.2077	˜p(wi|boprimeprime)
0.166667	-15.2078	methods, | al.,
0.176471	-15.2082	disease | disease
0.0967742	-15.2085	national | research
0.333333	-15.2086	boyer | raleigh,
0.0558376	-15.2094	includes | to
0.0504202	-15.2097	annotation | be
0.0355649	-15.2097	are | information
0.0289256	-15.2097	of | less
0.0681818	-15.2099	feature | have
0.0897436	-15.21	by | -
0.032368	-15.2101	translation | on
0.0754717	-15.2102	sentence | dependency
0.021875	-15.2102	more | for
0.5	-15.2103	sign | (lse)
0.333333	-15.2104	thesaurus | categories,
0.0621118	-15.211	we | (i.e.,
0.0823529	-15.2117	can | derived
0.0536585	-15.2118	are | described
0.0581395	-15.2119	of | property
0.0502092	-15.2123	differs | the
0.121951	-15.2123	that | influenced
0.0676692	-15.2124	was not | to
0.0769231	-15.2128	web | al.,
0.0237877	-15.2129	this | this
0.0555556	-15.213	in | phrase-based
0.5	-15.213	p(α|γ)
0.5	-15.213	(75%), | (12.5%)
0.0202218	-15.2135	the | computational
0.05	-15.2135	however, | or
0.1875	-15.2137	gives | where
0.0193822	-15.2137	in | features
0.0707965	-15.2144	we | slightly
0.157895	-15.2148	, | d’
0.0280612	-15.2153	information | is
0.133333	-15.2153	each | required
0.5	-15.2156	p(fj|ei)
0.0544554	-15.2162	used | at
0.222222	-15.2163	as | path.
0.222222	-15.2163	was | explained
0.285714	-15.218	instantiates | in
0.0447284	-15.2183	that | algorithm
0.5	-15.2183	maximalization | (18.a-c)
0.5	-15.2183	nvda | [nvda
0.0403226	-15.2187	annotations | of
0.0259459	-15.2192	is | target
0.0293296	-15.2196	we | work
0.04375	-15.2196	at | this
0.0923077	-15.2198	time | all
0.0491803	-15.2199	also | data
0.0284974	-15.2199	more | with
0.142857	-15.2202	criterion, | the
0.0452962	-15.2203	by | human
0.0479705	-15.2208	pages | june 2011.
0.0574713	-15.2208	pages | usa,
0.0747664	-15.2208	data | accuracy
0.285714	-15.221	substantial | compared
0.333333	-15.221	p(t|ˆd)
0.0671642	-15.221	i | 0
0.0864198	-15.2211	given | will
0.0558376	-15.2221	a | the | translation.
0.222222	-15.2222	for | created,
0.119048	-15.2223	such | corpus.
0.5	-15.2236	pseg(s|w,s,t,l,β,α)
0.0111811	-15.2237	to | which
0.0472727	-15.2239	1 | 6
0.0652174	-15.2239	metric | as
0.0923077	-15.2242	implementation, | the
0.0804598	-15.2246	for | method.
0.056701	-15.2248	all | different
0.119048	-15.2249	immediately | in
0.0364026	-15.2251	observed | of
0.0909091	-15.2253	finding | be
0.108696	-15.2254	this | section.
0.285714	-15.2256	is | labeled.
0.108696	-15.2259	procedings | on
0.0214815	-15.2262	the | precision
0.0625	-15.2263	linguistic | are
0.5	-15.2263	yields | p(t|d)
0.5	-15.2263	1,i | 1|t,k),
0.5	-15.2263	0.372 | 0.814
0.5	-15.2263	really | #craigyatcarnegie
0.5	-15.2263	pr(s|i)
0.5	-15.2263	conventional | speaking.
0.0853659	-15.227	between | event
0.0209235	-15.2271	the | ,
0.104167	-15.2276	tesla-m | the
0.333333	-15.2281	|pi|productdisplay
0.116279	-15.2283	it | greater
0.137931	-15.2283	to | frequencies.
0.0445205	-15.2285	acts | the
0.0502092	-15.2286	3 | that | of
0.285714	-15.2286	systems | somewhat
0.5	-15.2289	 have |  flu
0.00715355	-15.2296	the | an
0.333333	-15.2299	ep(γ|w)[f
0.333333	-15.2299	cause(coreentity|event)?, site(entity)?,csite(entity)?
0.333333	-15.2299	[y | [x
0.0595238	-15.23	keep | of
0.0886076	-15.2303	as | 2010;
0.032967	-15.2304	as | same
0.103448	-15.2306	entire | with
0.0747664	-15.2308	we | (i.e.
0.0617284	-15.2311	that | 1.
0.0462633	-15.2313	data | language
0.157895	-15.2316	did | significantly
0.0967742	-15.2317	k | r
0.176471	-15.2317	framework | general
0.176471	-15.2317	with | inference.
0.0307692	-15.2321	be | at
0.0188791	-15.2327	time | the
0.0823529	-15.2328	3 | sentence
0.0284168	-15.233	the | np
0.0115721	-15.2336	it | and
0.1875	-15.2337	shows | experiments.
0.0786517	-15.2337	first | semantic
0.00577847	-15.2339	the | we
0.108696	-15.2341	natural | tasks
0.0552764	-15.2343	as | x
0.5	-15.2343	kb | non-nil
0.5	-15.2343	(adverbs), | (determiners),
0.0661765	-15.2344	≤ | =
0.00371041	-15.2345	the | of the
0.0983607	-15.2348	findings, | are
0.0216049	-15.2348	semantic | a
0.0671642	-15.235	on | corresponding
0.333333	-15.2352	ﻲﻓ | ﻢﻟﺎﻌﻣ
0.113636	-15.2353	with | beam
0.0187244	-15.2353	model | is
0.0221493	-15.2353	applied | the
0.1	-15.2354	graph | can
0.111111	-15.2364	if | able
0.0235507	-15.237	we | a | set
0.333333	-15.237	interesting | kernels
0.333333	-15.237	philippe | ur1077
0.5	-15.237	κ0 | τj-s
0.0160067	-15.2379	to | set
0.0882353	-15.2383	one | should
0.153846	-15.2391	said | *
0.0293686	-15.2393	pairs | to
0.108696	-15.2394	structure | its
0.0428135	-15.2412	is | multiple
0.05	-15.2416	for | might
0.0393701	-15.2417	it | data
0.0544554	-15.2419	two | translation
0.0422961	-15.2423	if | more
0.2	-15.2424	generates | translation
0.5	-15.2424	p(h|c)
0.5	-15.2424	xk|∀xk∈xprime,βt1k
0.222222	-15.2425	would | colleagues
0.333333	-15.2425	when, | perpetrators,
0.0656934	-15.2428	our | should
0.0983607	-15.2428	complex | can
0.2	-15.2435	nguyen | al.
0.333333	-15.2443	ce | query-url
0.333333	-15.2443	(o | ra)
0.0412979	-15.2445	2002; | and
0.137931	-15.2447	even | due
0.0786517	-15.2448	therefore | are
0.176471	-15.245	functions | semantic
0.5	-15.2451	koc¸ | sariyer,
0.5	-15.2451	un | pared
0.5	-15.2451	|v|−
0.0536585	-15.2457	let | set
0.0188014	-15.2458	language | in
0.333333	-15.2461	p(ti|si)
0.0555556	-15.2464	been | 2010).
0.0606061	-15.2465	algorithms | al.,
0.25	-15.2466	+ | (i;t
0.111111	-15.2467	from | morphological
0.0210843	-15.247	first | is
0.0426829	-15.2471	a | topics
0.108696	-15.2471	however, | learning
0.0409357	-15.2474	same | by
0.0864198	-15.2477	segment | in
0.5	-15.2479	p(y|x;w)
0.0584795	-15.249	have | both
0.0459364	-15.2491	it | sentence
0.0408163	-15.2492	for | parallel
0.125	-15.2492	propositions | by
0.00557887	-15.2496	a | of | of
0.333333	-15.2497	ins | orth
0.0689655	-15.2499	online | in
0.0200669	-15.25	of | task
0.0528846	-15.25	system | model
0.0342052	-15.2502	in | relations
0.5	-15.2506	fwn | g1006
0.5	-15.2506	variance | greater.
0.5	-15.2506	stating | prior(askyn(α,β,at(β,λ),µ))
0.0595238	-15.2507	phrases | for
0.0391645	-15.2509	is | evaluation
0.222222	-15.251	lease | receivable,
0.125	-15.2513	summarization | sentence
0.0219334	-15.2513	table | with
0.055	-15.2516	are | topic
0.0733945	-15.2517	test | number
0.00240222	-15.2518	the | and
0.0337972	-15.2521	with | which | of
0.222222	-15.2522	from | tuple
0.125	-15.2527	next, | by
0.052381	-15.253	also | all
0.222222	-15.2534	60 | -
0.5	-15.2534	φ(x, | 4.1:
0.5	-15.2534	(bangalore, | fabrizio,
0.5	-15.2534	1|k)
0.0368664	-15.2536	and | task.
0.104167	-15.2536	the | rule:
0.028	-15.2544	the | complex
0.0190067	-15.2549	or | is
0.0661765	-15.2549	not | new
0.0671642	-15.2552	therefore | be
0.222222	-15.2559	2000 | 3000
0.5	-15.2561	joshi2 | 2dept.
0.0361174	-15.2562	would | that
0.0584795	-15.2564	entity | that
0.0666667	-15.2568	can | time
0.0666667	-15.2568	from | generated
0.137931	-15.2568	bleu | based
0.0983607	-15.2568	very | they
0.0923077	-15.2568	systems | table
0.333333	-15.2571	n1 | n′1
0.333333	-15.2571	tags, | resulting
0.333333	-15.2571	1−|a∩s|+|a∩p||a|+|s|
0.0777778	-15.2573	learning | data.
0.0786517	-15.2573	- | •
0.0544554	-15.2577	data | one
0.104167	-15.2581	has | focused
0.0518868	-15.2582	to | contexts
0.0217742	-15.2582	in | that | are
0.0472441	-15.2587	described | for
0.5	-15.2589	vast | form.
0.5	-15.2589	p(y|x),
0.5	-15.2589	cder | 0.0991
0.5	-15.2589	verbalized, | possessives
0.5	-15.2589	mountainbikes, | hollandr¨ader.
0.125	-15.2596	of | operating
0.05	-15.2597	collected | of
0.0727273	-15.26	assumption | be
0.0166282	-15.2602	of | use
0.0177324	-15.2603	relation | the
0.0777778	-15.2603	0.8 | 0.4
0.113636	-15.2607	different | annotation
0.0967742	-15.2607	findings, | in
0.333333	-15.2607	funny | boch
0.333333	-15.2607	c0 | w¡1
0.137931	-15.2614	approach | perform
0.0882353	-15.2615	development | training
0.5	-15.2616	student | (si)
0.5	-15.2616	favre, | brett
0.5	-15.2616	fr-en | 59.7
0.0206553	-15.2626	the | development
0.0777778	-15.2628	applied | from
0.0322004	-15.2629	this | be | for
0.137931	-15.2629	corresponding | given
0.052381	-15.2631	with | negative
0.333333	-15.2644	(yu | [9,
0.0495868	-15.2644	in this | we present
0.2	-15.2644	ktb | =⇒
0.2	-15.2644	δλ | δλ
0.2	-15.2644	40.64 | 65.47
0.2	-15.2644	[model | 3
0.2	-15.2644	theme(protein) | site(entity),
0.5	-15.2644	beneficiary | directness
0.5	-15.2644	had | ca-score
0.5	-15.2644	|φ−ψ|
0.0568182	-15.2648	more | training
0.0843373	-15.265	by | head
0.0373832	-15.265	2010) | the
0.0241449	-15.2651	and | given
0.15	-15.2653	bleu | 51.8
0.0428135	-15.2655	which | number
0.0173381	-15.2655	to | two
0.0695652	-15.2658	weight | with
0.0126073	-15.2659	the | and | not
0.0747664	-15.2659	precision | number
0.0404624	-15.2664	we | size
0.0528846	-15.2664	abstract | paper
0.137931	-15.2668	2. | approach
0.0256983	-15.2672	as | number
0.5	-15.2672	xml | caption,
0.0571429	-15.2676	most | at
0.119048	-15.2677	they | how
0.2	-15.2678	adm | chew
0.0612245	-15.2679	training | feature
0.0357942	-15.268	the | is defined
0.0291971	-15.268	by | using
0.0638298	-15.268	corpus | were
0.333333	-15.2682	‘was | punched’;
0.0786517	-15.2682	pairs | between
0.137931	-15.2683	algorithm | method
0.25	-15.2686	boolean | the | mentioned?
0.0260047	-15.2688	parsing | of
0.0591716	-15.2691	mooney, | and
0.108696	-15.2693	does | about
0.15	-15.2694	e |  score
0.0310966	-15.2698	semantic | as
0.037123	-15.2698	on | feature
0.25	-15.27	2000; | been
0.333333	-15.27	-> | ,eric
0.5	-15.27	hypothesizethat | didnotcarefullyexamine
0.5	-15.27	p(¯a|¯q))
0.0354767	-15.2704	, | is
0.1	-15.2706	results | –
0.285714	-15.2709	a | generations
0.285714	-15.2709	unlike | features.
0.0833333	-15.2709	using | topic
0.0452962	-15.2709	word | have
0.0608108	-15.2712	either | with
0.0497738	-15.2715	tree | by
0.0204369	-15.2715	single | the
0.157895	-15.2715	n | ≤
0.0461538	-15.2719	proceedings of | linguistics,
0.222222	-15.2719	2005) | english
0.222222	-15.2719	i· | parenleftbign
0.333333	-15.2719	|wi|
0.104167	-15.272	discussion | have
0.0289436	-15.2722	we | similar
0.2	-15.2722	linguistic | following
0.015466	-15.2723	can | a | of
0.0591716	-15.2727	the | b.
0.5	-15.2728	p(t|x)
0.5	-15.2728	column: | :column(x).
0.176471	-15.2733	a | bucket
0.0752688	-15.2739	data | 2010).
0.121212	-15.2739	vocabulary | each
0.0470588	-15.2742	o | e
0.0967742	-15.2742	kinds | are
0.0183486	-15.2743	of | source
0.148148	-15.2743	dog | dog
0.046875	-15.2744	all | into
0.222222	-15.2744	comp | 
0.0335306	-15.2746	corpus | by
0.05	-15.2748	nlp | to
0.0983607	-15.2749	best | use
0.032967	-15.2749	of | bleu
0.0291121	-15.275	note | is
0.15	-15.2753	eu25=bp20) | =+
0.5	-15.2756	(1) | marginalize
0.5	-15.2756	@ | ’a
0.5	-15.2756	|a
0.0282885	-15.276	a | machine
0.0436242	-15.2761	that | rather
0.0366972	-15.2766	and | does
0.0483871	-15.2766	3) | the | the
0.0869565	-15.2766	the | education
0.0497738	-15.2768	these | between
0.148148	-15.2768	ext | (1);
0.25	-15.277	dataset | times
0.0436242	-15.2771	model | based
0.104167	-15.2773	words | described
0.0470588	-15.2774	our | well
0.0218801	-15.2775	they | to | the
0.333333	-15.2775	gmcγ(cij|si)
0.0581395	-15.2778	empirical | edinburgh, | uk,
0.102041	-15.2782	have | opinion
0.00852483	-15.2783	the | and | with
0.5	-15.2784	3nlp | 1nlp
0.5	-15.2784	p(obj|si)
0.5	-15.2784	q∗(v|c(v))
0.5	-15.2784	(obj), | subjective-negative
0.0237154	-15.2789	table | that
0.0616438	-15.279	set | its
0.0689655	-15.2791	for | correlation
0.1875	-15.2792	ceafe | p
0.0216346	-15.2793	the | natural
0.0493274	-15.2794	source | language
0.333333	-15.2794	50, | 50”;
0.129032	-15.2796	translation | (och
0.0588235	-15.2796	thus, | that
0.0509259	-15.2797	two | where
0.125	-15.2798	nodes | v
0.133333	-15.2801	candidate | entity
0.06	-15.2802	results | same
0.176471	-15.2803	also | highly
0.0707965	-15.2805	abstract | machine
0.137931	-15.2807	on computational | usa,
0.0744681	-15.2808	were | set.
0.0183171	-15.2808	this | of | that
0.15	-15.2812	ok | sub
0.5	-15.2812	p(j|s,t)
0.00467082	-15.2816	the | with
0.025641	-15.2822	computational linguistics | and
0.0353982	-15.2829	accuracy | a
0.333333	-15.2831	have<vbhaver><pres> | to<pr>
0.037037	-15.2835	presents | of
0.0493274	-15.2836	shown | by
0.15	-15.2838	type | np
0.5	-15.284	confusion | protease
0.5	-15.284	(macqueen, | lloyd,
0.5	-15.284	p(sub|si)
0.027451	-15.2841	features | that
0.166667	-15.2841	inference | entities
0.0351648	-15.2841	of | w
0.222222	-15.2844	star | oj)
0.0744681	-15.2846	o | l
0.0543478	-15.2849	class | as
0.333333	-15.285	coarse-grained | day,
0.333333	-15.285	clitic | cliticdeterminertype
0.0980392	-15.2854	strength | a
0.0666667	-15.2855	table | shows | by
0.129032	-15.2855	(3) | not
0.0251366	-15.2856	for | system | the
0.0347826	-15.2857	document | is
0.166667	-15.286	statistical | proposed
0.0683761	-15.2865	is | is,
0.0672269	-15.2866	i | o
0.333333	-15.2869	yv | ye.
0.333333	-15.2869	noun_phrase{ | american<adj>
0.2	-15.2869	tagger | orig
0.5	-15.2869	death | presidency
0.5	-15.2869	appeal | cultures.
0.0429043	-15.287	all | table
0.0528846	-15.2875	in | experimental
0.133333	-15.2877	these | along
0.0424837	-15.2878	2 | an
0.0232126	-15.2886	of the | computational
0.333333	-15.2888	kendall | (flat)
0.333333	-15.2888	its | syntax,
0.0683761	-15.2888	this | question
0.104167	-15.2891	translation | words.
0.0714286	-15.2894	properties | as
0.5	-15.2897	pre-cluster | formthefinalclusterresults.
0.0318021	-15.29	and | the best
0.0568182	-15.2904	was | number
0.333333	-15.2907	b) | ;y
0.333333	-15.2907	d. | computaci´on
0.0200831	-15.2909	algorithm | of
0.0230627	-15.291	we | an | in
0.285714	-15.2918	letters | arabic
0.1875	-15.2919	figure | plate
0.0980392	-15.2921	,q | 1
0.0695652	-15.2925	opinion | are
0.039548	-15.2926	with | into
0.5	-15.2926	0s | 7s
0.5	-15.2926	ubiu | (zhekova
0.5	-15.2926	g(+g2)1 | (vprime,e+2
0.0736842	-15.2927	irrelevant | the
0.0392157	-15.2932	at | not
0.0319149	-15.2932	task | is
0.0316901	-15.2932	the | that | any
0.222222	-15.2933	(marton | 2008;
0.0384615	-15.2937	of | linguistics, pages
0.0661157	-15.2937	+ | )
0.121212	-15.2938	0.2 | accuracy
0.0427632	-15.2939	we then | to
0.25	-15.2941	applications | sentiment
0.0629371	-15.2942	his | of
0.0895522	-15.2942	language | any
0.133333	-15.2945	2011 | biomedical
0.333333	-15.2945	equivalent | questions.
0.222222	-15.2946	gulf | and
0.0364964	-15.2947	most | by
0.0625	-15.2948	methods | not
0.1	-15.2949	we | future.
0.0397727	-15.2952	of | dialogue | and
0.5	-15.2954	(dtks) | (ddtks).
0.0683761	-15.2955	been | 2006;
0.0967742	-15.2956	to | forms.
0.0355556	-15.2957	a | by | et
0.113636	-15.2959	finding | from
0.0531915	-15.296	of | g
0.0795455	-15.2962	g | s
0.333333	-15.2964	predictive | interface
0.333333	-15.2964	after | update,
0.0604027	-15.2968	(4) | to
0.0412979	-15.297	is | 3
0.0493274	-15.2974	of | process.
0.0318584	-15.298	between | by
0.5	-15.2983	(chou | juang
0.0666667	-15.2984	2 | +
0.285714	-15.2984	introduces | defined
0.0155528	-15.2984	if | to
0.0266497	-15.2986	analysis | to
0.0224417	-15.2988	same | to
0.0391061	-15.2989	we | task.
0.0409357	-15.299	given | set
0.0405797	-15.299	is | example
0.111111	-15.299	have | function
0.0540541	-15.2996	on | input
0.117647	-15.2997	22 | for
0.016525	-15.2998	to | a | with
0.0672269	-15.3	moses | et
0.333333	-15.3003	cen | ftc
0.333333	-15.3003	however, | streams,
0.0313043	-15.3008	sentences | for
0.222222	-15.3009	instructions, | user
0.0608108	-15.3009	use | target
0.0540541	-15.3012	as | whether
0.0744681	-15.3012	is | approach.
0.0580645	-15.3012	not | them
0.5	-15.3012	o(|g|2),
0.5	-15.3012	morse | dashes,
0.5	-15.3012	oil* | made*
0.5	-15.3012	1|t|
0.0276243	-15.3015	using | by
0.0190597	-15.3015	a | a | b
0.0252294	-15.3016	using | on
0.285714	-15.3017	tokens) | for
0.0804598	-15.3019	approach | well
0.00441216	-15.3022	is | of
0.333333	-15.3022	|u|
0.333333	-15.3022	indicate | variable.
0.00958926	-15.3031	more | the
0.0616438	-15.3039	source | phrase
0.0491071	-15.3041	each | source
0.0695652	-15.3041	model | pairs
0.113636	-15.3044	training | svm
0.0786517	-15.3044	exploiting | of
0.0149547	-15.3046	in | using
0.0257913	-15.3051	where | for
0.0356347	-15.3055	would | as
0.25	-15.3056	ng | gme
0.1875	-15.3059	the | “/
0.0777778	-15.306	consist | a
0.0493274	-15.3062	on | annotation
0.0255814	-15.3063	more | as
0.285714	-15.3066	literature | text
0.0385675	-15.3067	text | with
0.5	-15.307	county* | men*
0.5	-15.307	kanssa | with-post
0.0504587	-15.3078	it | that | or
0.064	-15.3079	show | which
0.0440678	-15.3079	as | negative
0.176471	-15.3085	finds | =
0.0299213	-15.3086	each | all
0.0217938	-15.3086	the | make
0.0588235	-15.3089	of | kinds
0.111111	-15.3092	lexical | form
0.00953137	-15.3093	this | and | the
0.0213816	-15.3099	we | approach
0.5	-15.3099	decca | 1,158,342
0.176471	-15.3106	storage | for
0.133333	-15.3107	at | respectively.
0.0604027	-15.3108	sentences | more
0.0310345	-15.3111	is | document
0.0786517	-15.3112	of | explored
0.0509259	-15.3113	since | not | to
0.25	-15.3114	compared | whereas
0.0683761	-15.3117	rule | =
0.032197	-15.3118	most | as
0.0474308	-15.312	as | tree
0.0961538	-15.312	the | factor,
0.0396601	-15.3122	systems | as
0.084507	-15.3125	grammar | or
0.0151577	-15.3125	other | of
0.102041	-15.3127	into | sets
0.5	-15.3128	unlike | webps
0.025	-15.3131	scores | of
0.0882353	-15.3131	→ | s
0.142857	-15.3132	bel(y | bel(y
0.142857	-15.3132	%mis | (10k)
0.142857	-15.3132	andcj | yif
0.285714	-15.3133	rankpp∞ | ghkm
0.285714	-15.3133	v | wxi
0.285714	-15.3133	r´eservation | une
0.285714	-15.3133	fact-driven | benefit
0.285714	-15.3133	(00) | 00)
0.285714	-15.3133	aobjetr´eserv´e | inform
0.285714	-15.3133	server-side | mashups
0.285714	-15.3133	alef | hamza
0.285714	-15.3133	75.47 | —–
0.285714	-15.3133	girl | dt-5
0.285714	-15.3133	g1004g856g856g910 | g1004g856g856g910
0.285714	-15.3133	„c1 | c2‟
0.285714	-15.3133	5.77 | asr-
0.285714	-15.3133	sex | reef
0.285714	-15.3133	basicemotion | oppositeemotion
0.285714	-15.3133	θa, | tg
0.285714	-15.3133	0.859 | 0.538
0.285714	-15.3133	sex | pirat
0.285714	-15.3133	cap-macro | sempos-bleu
0.285714	-15.3133	fac | ;
0.285714	-15.3133	mt04 | wb
0.285714	-15.3133	x1 | null4null:
0.285714	-15.3133	dell | hp’s
0.285714	-15.3133	s49s48 | s52s48
0.285714	-15.3133	gold | (gold/gold).
0.285714	-15.3133	d97 | d89
0.285714	-15.3133	a5a0a2 | a19a16a12a18a14a20a8a12a21a8
0.285714	-15.3133	(jj) | (great)),
0.285714	-15.3133	monopats | p(pi)
0.285714	-15.3133	r= | apart)
0.285714	-15.3133	l=2 | l=1
0.285714	-15.3133	a3a3a2 | a19a16a12a18a14a20a8a12a21a8
0.285714	-15.3133	(cl) | spelling
0.285714	-15.3133	braceleftbigg | conditions
0.285714	-15.3133	a18 | na80
0.285714	-15.3133	mwen | genyen
0.285714	-15.3133	mt-chat∗ | 2.2e-16
0.285714	-15.3133	mt05 | wb
0.285714	-15.3133	<lower | part
0.285714	-15.3133	^c | …,
0.285714	-15.3133	∂wj | x,y]
0.285714	-15.3133	(αprime,γprime) | tgt
0.285714	-15.3133	−1) | sim(i−1,j)
0.285714	-15.3133	structure> | structure”.
0.285714	-15.3133	bigram | (overlapping)
0.285714	-15.3133	(frame) | meant
0.285714	-15.3133	96.0 | 95.6
0.285714	-15.3133	ghussa | ali
0.285714	-15.3133	xnoun | mapsto→
0.285714	-15.3133	mw, | pmi(hw,
0.285714	-15.3133	a18 | a124wda124,
0.285714	-15.3133	1933 | +40
0.285714	-15.3133	turnstileleft | (σ|i|j|k,β,a)
0.285714	-15.3133	nullδt(ytv | ytp)
0.285714	-15.3133	list; | wie
0.285714	-15.3133	(fn) | x∈d
0.285714	-15.3133	-87 | -99
0.0961538	-15.3137	: | summationdisplay
0.0350109	-15.3142	for | similarity
0.25	-15.3143	genia | was
0.0393258	-15.3145	to | sequence
0.0446097	-15.3148	example, | word
0.0961538	-15.3148	2: | scores
0.0230326	-15.3151	the | important
0.0813953	-15.3152	2007) | et
0.076087	-15.3155	need | which
0.5	-15.3157	umri-001 | marianna.apidianaki@inria.fr
0.5	-15.3157	p(pi(αx)|nx
0.5	-15.3157	dosage | chemical
0.5	-15.3157	harvard | subcontract
0.5	-15.3157	interspeech | subchallenge.
0.125	-15.3164	firstly, | a
0.142857	-15.3166	nn; | jj
0.0449438	-15.317	since | from
0.0695652	-15.3174	= | r | =
0.333333	-15.3177	ter-plus | (snover
0.333333	-15.3177	dinu | p(z|o,c)
0.121212	-15.3177	two | clustering
0.0345572	-15.3178	nodes | of
0.0666667	-15.3181	: | 1
0.5	-15.3186	(zhao | al-onaizan,
0.5	-15.3186	82.4 | 21.4
0.5	-15.3186	∈ | k+m
0.5	-15.3186	0.47 | 2.49
0.0482456	-15.319	a | information.
0.0531915	-15.3192	when | training
0.0244941	-15.3192	null | and
0.1	-15.3193	our | term
0.0288316	-15.3205	word | is | and
0.0410095	-15.3208	training | each
0.0666667	-15.3209	that | simply
0.176471	-15.3209	just | into
0.0584416	-15.3209	that | correlation
0.0287443	-15.321	system | et
0.157895	-15.3213	same | time.
0.0387812	-15.3217	algorithm | on
0.0486726	-15.3218	if | set
0.0275862	-15.3231	on | it | the
0.0122962	-15.3233	that | on
0.1	-15.3234	show | while
0.121212	-15.3234	experiment | more
0.0493274	-15.3241	n | r
0.0701754	-15.3243	we | scores.
0.5	-15.3245	system-level | ordered
0.5	-15.3245	did | wellaswehadexpected.somepossiblereasonsare
0.5	-15.3245	no.5 | 100080,
0.5	-15.3245	0.081 | 0.130
0.5	-15.3245	communications | 18(11),
0.142857	-15.325	query | also
0.0795455	-15.3263	output | each
0.0923077	-15.3263	like | helpful
0.166667	-15.3266	approaches | often
0.0821918	-15.3266	an | future
0.114286	-15.3267	experiments | features.
0.0961538	-15.3267	parser | our
0.0155239	-15.3269	and | more
0.064	-15.327	in | parallel | a
0.0909091	-15.327	shown | 5.
0.0909091	-15.327	paraphrases | are
0.176471	-15.3272	several | make
0.129032	-15.3273	  | a 
0.5	-15.3275	aiµ | µhaµ
0.104167	-15.3278	addition, | or
0.0157374	-15.3279	which | for
0.0210526	-15.3284	and | sentences
0.0157303	-15.3287	only | and
0.0209003	-15.3289	n | the | of
0.1875	-15.329	( | 
0.1875	-15.329	single-word | single-word
0.09375	-15.329	0.96 | 0.96
0.0198864	-15.3292	while | and
0.117647	-15.3293	the | shifted
0.333333	-15.3295	p(tildewidee,˜f|f)
0.0458015	-15.3304	with | learning
0.5	-15.3304	negativity) | negativity
0.5	-15.3304	p(t,a|s,n)
0.222222	-15.3308	one | concept,
0.0833333	-15.331	not | candidate
0.027027	-15.331	the | score | a
0.125	-15.3312	2006; | et | 2007),
0.333333	-15.3315	pk,wd,n | wk,l,wd,n
0.333333	-15.3315	[slot1,slot2,slot3,slot4], | [slot1,slot2,slot3,slot4],
0.333333	-15.3315	factory | usine
0.0278552	-15.3316	+ | and
0.0304054	-15.3317	nullnull | null
0.114286	-15.3321	compare | three
0.114286	-15.3321	language | shared
0.0375335	-15.3321	at | which
0.0645161	-15.3323	and | span
0.0833333	-15.3324	8: | of
0.157895	-15.3325	and | marina
0.0736842	-15.3326	can | small
0.0178263	-15.3333	proposed | the
0.5	-15.3334	zd,n | p(zd,n
0.0512821	-15.334	preference | the
0.0437956	-15.3342	events | to
0.0318949	-15.3344	the | takes | the
0.021329	-15.3346	graph | the
0.0518135	-15.3348	corpus | using
0.0810811	-15.3349	their | lexical
0.0474138	-15.3349	(1) | (3)
0.0322581	-15.3352	main | is
0.0302521	-15.3353	and | tree
0.0286145	-15.3353	precision | and
0.333333	-15.3354	γprime | αprime
0.0634921	-15.3357	between | was
0.5	-15.3364	1|x)
0.5	-15.3364	luis | dabbish.
0.5	-15.3364	p(ei|e−i,x,dx,θx
0.5	-15.3364	45th | acl,
0.0284431	-15.3365	introduction | and
0.0744681	-15.3366	as | alternative
0.333333	-15.3374	µf→x | n(f)\x
0.111111	-15.3374	error | =
0.0786517	-15.3381	only | phrase
0.0253165	-15.3383	in | paper | the
0.0795455	-15.3386	topic | distribution
0.0736842	-15.3392	is | seed
0.102041	-15.3392	last | which
0.0446097	-15.3393	how | by
0.5	-15.3394	cr(cij|si)
0.5	-15.3394	approp(n, | agr).
0.0478261	-15.3397	to | represent
0.0810811	-15.3397	m | number
0.0427632	-15.3398	where | = | and
0.0616438	-15.3398	with | classifier
0.0167854	-15.3401	we | words
0.0402477	-15.3402	than | or
0.106383	-15.3406	7 | each
0.0377358	-15.3411	set | our
0.0721649	-15.3412	successful | and
0.333333	-15.3414	(ackerman, | professor
0.0436364	-15.3418	indicates | is
0.106383	-15.3418	× | +
0.5	-15.3424	0.039 | 0.128
0.5	-15.3424	hinrich | stuttgart,
0.05625	-15.3424	2 | system
0.0526316	-15.3425	by | data.
0.133333	-15.343	experiments, | each
0.137931	-15.3433	will | model.
0.0620155	-15.3435	l | i
0.0666667	-15.3439	translation | between
0.035545	-15.344	we | set.
0.0882353	-15.344	liu | al.
0.176471	-15.3451	usually | same
0.5	-15.3454	w=0.01 | w=0.1w=0.01
0.5	-15.3454	|cd|summationdisplay
0.5	-15.3454	playing | harp!
0.117647	-15.3462	according | following
0.0158066	-15.3462	from | that
0.0540541	-15.3464	p | s
0.166667	-15.3465	group | same
0.104167	-15.3465	sentence | vector
0.125	-15.3476	first | candidates
0.0258303	-15.3477	of | 3
0.0480349	-15.3477	based | their
0.0293638	-15.3478	corpora | of
0.0620155	-15.348	of | map
0.5	-15.3484	qe | qm,
0.157895	-15.3486	und | der
0.15	-15.3487	(2007) | our
0.0371353	-15.3491	corpus | a | a
0.0149938	-15.3491	in | set
0.333333	-15.3494	alexandra | bertoldi,
0.333333	-15.3494	(fung | yee,
0.333333	-15.3494	bigrams | hf(w3|w1w2)’s
0.333333	-15.3494	vertices | neighbors
0.0810811	-15.3505	two | terms
0.025641	-15.3507	source | in
0.111111	-15.3508	pair | probability
0.5	-15.3514	ˆl(u) | ˆl(v)
0.5	-15.3514	j | ∂ξi
0.5	-15.3514	instrument-agency | product-producer
0.5	-15.3514	(berger | jeon
0.5	-15.3514	10)|2
0.5	-15.3514	(rhodes | maes,
0.0281899	-15.3517	section | our
0.0257985	-15.3517	score | to
0.022977	-15.3521	best | and
0.0113906	-15.3521	used | and
0.125	-15.3522	learn | data.
0.0410095	-15.3524	an | training
0.157895	-15.3524	3: | ∈
0.0380435	-15.3531	our | semantic
0.0502513	-15.3534	are | positive
0.076087	-15.3536	chart | and
0.06	-15.3536	to | classifier.
0.0281899	-15.3544	english | a
0.5	-15.3545	p(ym|xm)
0.5	-15.3545	dts | defining
0.0297521	-15.3549	and | corpus.
0.0472103	-15.3549	however, | than
0.2	-15.3551	is | analogously
0.07	-15.3554	use | probability
0.0181488	-15.3554	to | only
0.0507614	-15.3556	words | between
0.0117317	-15.3562	it | in
0.0252708	-15.3572	results | as
0.5	-15.3575	σ∈sprime|σ|p
0.0458333	-15.3576	may | on
0.0810811	-15.3579	are | produced
0.0588235	-15.3579	a | for | machine
0.0149938	-15.3582	a | than
0.0380435	-15.3584	to | list
0.0149564	-15.3588	be | with
0.166667	-15.3596	corpora | 2003;
0.333333	-15.3596	w* | (x-y)
0.125	-15.3598	the | raters.
0.028065	-15.36	that | do
0.5	-15.3605	p(e|q)
0.5	-15.3605	auto | 0.150.2
0.5	-15.3605	t(e′|
0.5	-15.3605	gaifman | shapiro
0.0752688	-15.3607	performed | as
0.181818	-15.3608	boxshadowdwn | acoustic
0.157895	-15.3611	l | bow
0.333333	-15.3616	eng | ita
0.0512821	-15.3619	we | increase
0.076087	-15.3624	> | > | of
0.05	-15.3624	corpus | training
0.0278184	-15.3626	although | to
0.0204563	-15.3626	we | test
0.104167	-15.3629	1: | scores
0.157895	-15.363	can | constructions
0.0225989	-15.3631	of | portland, oregon,
0.0769231	-15.3631	clustering | al.,
0.176471	-15.3633	3.3.1 | of
0.333333	-15.3636	p(u|v)
0.111111	-15.3636	7) | the
0.2	-15.3636	were | japanese
0.5	-15.3636	p(witi|φ(wi−1i−n+1ti−1i−n+1))
0.5	-15.3636	), | ,x,b)
0.5	-15.3636	gimpel | loss
0.166667	-15.3637	would | extend
0.0707071	-15.3638	paths | and
0.0679612	-15.3642	for | f1
0.0444444	-15.3651	query | is
0.0252404	-15.3653	to | dependency
0.0416667	-15.3658	there | more
0.136364	-15.3664	thus, | most
0.0168421	-15.3665	more | a
0.5	-15.3667	guesswork | em.
0.0534759	-15.367	different | one
0.333333	-15.3677	ax | e
0.0164271	-15.3678	of | table
0.033543	-15.3678	feature | are
0.0068227	-15.3679	in | the | is
0.0779221	-15.3683	2 | y
0.222222	-15.3685	text | phrases.
0.133333	-15.369	paper | human
0.0210084	-15.3692	or | for
0.0304114	-15.3692	method | is
0.333333	-15.3698	alicante | montoyo}@dlsi.ua.es
0.333333	-15.3698	utterances | speaker.
0.333333	-15.3698	h | comps
0.333333	-15.3698	l | disjoint
0.5	-15.3698	pragmatic | development,
0.181818	-15.3698	penalty | normal
0.108108	-15.3701	which | binary
0.0895522	-15.3703	generation | an
0.0391566	-15.3704	test | by
0.0315399	-15.3707	five | the
0.166667	-15.3709	0.36 | 0.11
0.0769231	-15.3717	because | model
0.0214408	-15.3719	the | languages
0.333333	-15.3719	interval | {r∈r|r1
0.0142747	-15.3723	to | model
0.0283465	-15.3727	wikipedia | the
0.5	-15.3729	7.4 | 82.3
0.5	-15.3729	t(¯e|
0.5	-15.3729	nwi|zi
0.0882353	-15.373	measure | not
0.0153576	-15.3733	is | this
0.031746	-15.3736	english | for
0.166667	-15.374	workshop | cultural
0.222222	-15.374	pharmacy | minneapolis,
0.0650407	-15.3741	using | parallel
0.0882353	-15.3741	 | >
0.018159	-15.3746	of | and | an
0.166667	-15.375	jiang, | and
0.0980392	-15.3755	information | research
0.102041	-15.3759	) | o
0.0385757	-15.3759	based | which
0.5	-15.376	erabakirik | did
0.0744681	-15.3765	on | results.
0.016444	-15.3766	is | between
0.0252101	-15.3767	for | also
0.108108	-15.3768	use | al., 2010).
0.108108	-15.3775	intervals | and
0.00854348	-15.379	the | the | it
0.0371353	-15.3791	system | not
0.125	-15.3791	follow | al.
0.5	-15.3791	state* | made*
0.5	-15.3791	decides | local
0.114286	-15.3794	srl | that
0.017321	-15.3794	in | were
0.0869565	-15.3795	〈〈q | 2
0.333333	-15.3801	||t||
0.0882353	-15.3802	development, | the
0.0119956	-15.3804	the | of | data
0.2	-15.3809	cpred | wj
0.15	-15.381	frequently, | the
0.133333	-15.3814	± | na¨ıve
0.0321932	-15.3818	the | allows | to
0.0433213	-15.3818	other | it
0.0789474	-15.3821	then | section
0.333333	-15.3822	x2 | (deliver-s2’
0.5	-15.3822	et, | eprimet.
0.5	-15.3822	align | ei:j,
0.5	-15.3822	bkg, | res,
0.028754	-15.3823	a | between | in
0.157895	-15.3825	is | difficult.
0.0442804	-15.3831	an | different
0.0620155	-15.3834	aspect | to
0.2	-15.3834	emboldened | [terrorists]−
0.2	-15.3834	high | recall,
0.013986	-15.3843	and | our
0.0371353	-15.3847	sentences | be
0.0526316	-15.3848	on | extracted
0.142857	-15.3849	build | target
0.0925926	-15.3857	similarity | have
0.0789474	-15.3865	2 | approach
0.25	-15.3869	fork, | agent(the
0.0943396	-15.387	6.2 | we
0.0961538	-15.387	work | types
0.108108	-15.387	class | query
0.07	-15.3874	by | summationdisplay
0.023	-15.3877	of | machine
0.08	-15.3877	al., | supervised
0.0532544	-15.3881	these | based
0.0436364	-15.389	a | set | =
0.0759494	-15.3892	identified | by
0.176471	-15.3896	teams | quarterbacks
0.0882353	-15.3896	machine | scotland, | 30–31,
0.0869565	-15.3898	triggers | and
0.0478261	-15.3899	apply | in
0.0666667	-15.3902	only | over
0.0759494	-15.3902	this paper | follows.
0.333333	-15.3906	|c
0.0199387	-15.3908	we | other
0.129032	-15.3912	ner | is
0.0304114	-15.3912	and | negative
0.0415225	-15.3914	was | than
0.157895	-15.3915	matches | which
0.0392749	-15.3916	we | model,
0.5	-15.3916	argmin | vextenddoublevextenddoubled−ftwvextenddoublevextenddouble2
0.5	-15.3916	word-lattice | cn-based
0.0441767	-15.3919	we | wikipedia
0.0535714	-15.3922	to | components
0.142857	-15.393	then, | two
0.0610687	-15.3934	system | sentence
0.0314342	-15.3935	look | the
0.0225933	-15.3936	6 | of
0.166667	-15.3938	as | sampled
0.0184332	-15.3943	is | words | the
0.0391566	-15.3944	experiment | to
0.0149637	-15.3945	table | for
0.0909091	-15.3948	the | session,
0.142857	-15.3948	(2007) and | al.
0.142857	-15.3948	0 | ,β
0.5	-15.3948	2000) | (blaheta
0.5	-15.3948	`a | cayes.
0.181818	-15.3948	, | 好
0.027735	-15.396	by | set
0.121212	-15.396	developed | information
0.0597015	-15.3961	fact | be
0.028436	-15.3969	performance | is
0.333333	-15.3969	(‘to | decision’),
0.0260417	-15.3972	such | that
0.108108	-15.3972	general | corpus
0.024911	-15.3979	would | for
0.5	-15.3979	dg, | timestamp.
0.5	-15.3979	|τsprime
0.5	-15.3979	˜p(w|h)
0.5	-15.3979	vegematic | heuristic
0.5	-15.3979	|ar|
0.181818	-15.3982	challenge | set
0.08	-15.3983	t | ∈
0.0666667	-15.3987	on | negative
0.333333	-15.399	program | direct
0.0542169	-15.3991	that | be | into
0.0580645	-15.3993	following | word
0.0645161	-15.3993	from | pair
0.0877193	-15.3994	document | same
0.181818	-15.3994	word expansion | extraction
0.0116773	-15.3995	approach | the
0.25	-15.3996	scores | dataset.
0.0821918	-15.3997	part | grant
0.0693069	-15.4	conclusions | have
0.0769231	-15.4001	previous | use
0.0185676	-15.4004	it | with
0.0311891	-15.4005	dialogue | a
0.0116309	-15.4005	then | the | the
0.0458333	-15.4007	we | step
0.0601504	-15.4008	is | h
0.0679612	-15.4008	based | analysis
0.5	-15.4011	pp(wk,tk,gk|d)
0.5	-15.4011	prep(in) | actvp(derived)
0.0276498	-15.4015	• | and | of
0.0160643	-15.4024	of | target
0.0377907	-15.4025	a | -
0.114286	-15.4032	each | correspond
0.166667	-15.4033	positive | human
0.333333	-15.4033	unfortunately, | amount
0.0148494	-15.4034	the | those
0.0424028	-15.4034	t | n
0.0195489	-15.404	research | the
0.5	-15.4043	spelling | 0.825001
0.5	-15.4043	extensive | ebp
0.5	-15.4043	|t|log|s|
0.5	-15.4043	jmwe | 2011a).
0.0980392	-15.4046	which | perform
0.0744681	-15.4051	defined | null
0.166667	-15.4055	v | 3:
0.0523256	-15.4057	methods | scotland,
0.040678	-15.406	null | that
0.03386	-15.4063	it | a | with
0.181818	-15.4063	such | estimate
0.0686275	-15.4068	2005) | to
0.023918	-15.407	null | a
0.0160643	-15.4071	both | of
0.0135231	-15.4072	all | the | and
0.333333	-15.4075	telephone | franklin
0.5	-15.4075	p(pi|w(pi))
0.5	-15.4075	(open | system:
0.0909091	-15.4076	results are | by
0.064	-15.4082	and | space.
0.0365535	-15.4082	related | is
0.125	-15.4083	können:can | es:it
0.0892857	-15.4088	x | j
0.108108	-15.4089	this | detection.
0.0456432	-15.409	of | domain.
0.25	-15.4092	perfect | because
0.0191761	-15.4095	the | negative
0.0305344	-15.4095	the | of | assigned
0.0240275	-15.4098	1 | =
0.0660377	-15.4098	work | analysis
0.00976059	-15.41	is | as | the
0.0686275	-15.4107	methods | based
0.5	-15.4107	embraces | ability,
0.0127632	-15.4115	a | =
0.00690036	-15.4116	we | that | the
0.333333	-15.4118	p(x|r)
0.00766415	-15.4121	and | that
0.105263	-15.4123	work | either
0.111111	-15.4125	check | we
0.130435	-15.4126	restr | 
0.0714286	-15.4131	discourse | al.,
0.142857	-15.4131	feature | several
0.2	-15.4139	parameter | every
0.5	-15.4139	(50) | (58)
0.0857143	-15.4141	treebank | with
0.0167317	-15.4141	1 | to
0.0740741	-15.4152	model | small
0.2	-15.4152	∏ | =
0.25	-15.4156	pre-processing | +morph1+
0.08	-15.4158	whose | be
0.07	-15.416	with | frequent
0.0375723	-15.4168	do | that
0.0597015	-15.4168	f | p
0.0271493	-15.4168	but | be
0.5	-15.4172	mehr | transportieren.
0.0462185	-15.4175	c | n
0.0492611	-15.4176	5 | 20
0.0188679	-15.4181	several | the
0.136364	-15.4181	paper, | consider
0.0625	-15.4184	dialogue | or
0.0857143	-15.4186	prior | from
0.0857143	-15.4186	++ | ++
0.222222	-15.419	corpus | stories
0.0344828	-15.4191	the | time,
0.0505051	-15.4192	translation | training
0.181818	-15.4192	m | according
0.181818	-15.4192	(p1,p2) | and
0.0740741	-15.4195	0.1 | 0.4
0.157895	-15.4197	the | singletons.
0.5	-15.4204	za|y
0.0620155	-15.421	can | tree
0.0980392	-15.421	30 | 90
0.0759494	-15.4218	be | text.
0.105263	-15.4218	more | (e.g.
0.222222	-15.4219	cat=n | neq
0.0464135	-15.422	use | using
0.0679612	-15.4222	translation | trained
0.114286	-15.4224	one | (e.g.
0.0642202	-15.4225	gold | by
0.0230366	-15.4226	the | is | different
0.0789474	-15.4229	where | w
0.0265363	-15.4232	in | large
0.0379009	-15.4235	more | can
0.5	-15.4237	buildings | artists
0.102564	-15.4242	mw | %
0.102564	-15.4248	will | evaluation
0.0740741	-15.4248	conclude | is
0.0542169	-15.4249	over | all
0.0555556	-15.4252	on | parse
0.0721649	-15.4252	shown | 1
0.0456432	-15.4252	we | sentence.
0.047619	-15.4254	science | university
0.0231092	-15.4257	are | word
0.0642202	-15.4263	parsers | al.,
0.0323276	-15.4269	possible | is
0.5	-15.4269	ends; | leftdependency
0.333333	-15.427	inference: | nullwe,
0.033557	-15.4273	+ | -
0.05625	-15.4273	1 | :
0.057971	-15.4274	matches | a
0.0610687	-15.4276	was | translation
0.027439	-15.4288	to | systems
0.0487805	-15.4297	be | even
0.0779221	-15.4302	that | path
0.5	-15.4302	�” | (tsune)
0.5	-15.4302	(v,w) | (u,w)
0.0769231	-15.4304	translation | evaluation
0.0759494	-15.4307	np | dt
0.15	-15.4309	assume | use
0.181818	-15.4311	structure, | may
0.333333	-15.4313	yon | mouri
0.0648148	-15.4314	expressed | not
0.25	-15.4319	1|t,k)
0.2	-15.4322	affixes | are
0.0142682	-15.4325	but | of
0.0318471	-15.4331	and | value
0.111111	-15.4335	with | treated
0.0629921	-15.4335	experiments | model
0.166667	-15.4336	“happy | and
0.0145299	-15.4339	the | would | the
0.0326087	-15.4341	this | feature
0.0909091	-15.4341	4 | how
0.0209059	-15.4341	how | of
0.00966004	-15.4354	the | of | not
0.0779221	-15.4354	3. | 4.
0.00679897	-15.4362	the | word | the
0.0707071	-15.4362	required | is
0.5	-15.4368	almost | translations.
0.5	-15.4368	graph | f(x1,x2,x3,x4)
0.5	-15.4368	nowo-pcfg | wo-pcfg
0.5	-15.4368	fasho | coo
0.5	-15.4368	9000 | 2dept.
0.0117198	-15.4375	is | that | of
0.166667	-15.438	the | complicated.
0.0259918	-15.438	in | standard
0.114286	-15.4389	states, | is
0.0523256	-15.439	processing, | scotland, uk,
0.0198728	-15.439	the | example
0.0373563	-15.4392	syntactic | is
0.0392157	-15.4392	addition, | to
0.0094914	-15.4394	were | the
0.333333	-15.4401	|σ|+
0.333333	-15.4401	αi | γa(j)
0.102564	-15.4403	to | authors.
0.181818	-15.4407	tokens | will
0.2	-15.4414	shown | baseline.
0.333333	-15.4423	(etzioni | kozareva
0.166667	-15.4424	(u,v) | is
0.15	-15.4428	both | long
0.0256757	-15.4429	and | order
0.5	-15.4434	wants | millionaire
0.5	-15.4434	(out | support
0.5	-15.4434	%de | gezien
0.5	-15.4434	history | esp
0.222222	-15.4438	black; | >,
0.181818	-15.4443	upp | b
0.0216369	-15.4445	of | single
0.0648148	-15.4447	operations | and
0.0344828	-15.4463	summarization | the
0.125	-15.4467	%fa30 | (10k)
0.5	-15.4467	)|(m
0.5	-15.4467	βj | γa(j)
0.5	-15.4467	blac | le?
0.166667	-15.4468	(great) | (ew
0.166667	-15.4468	0.0000 | 0.0000
0.25	-15.4468	comps | semantics,
0.25	-15.4468	qsomeriverqeverycity | border
0.25	-15.4468	(frame)” | meant
0.25	-15.4468	jj) | vbn\
0.25	-15.4468	orcj | cj)
0.25	-15.4468	¯np | ¯sbar
0.25	-15.4468	compatible(rj,mi) | cand
0.25	-15.4468	d101 | d90
0.25	-15.4468	head, | abstractness,
0.25	-15.4468	freq | 0.003%
0.25	-15.4468	yif | mentionsagreeinbothnumberandgender;elsen
0.25	-15.4468	fold | τcv
0.25	-15.4468	i-cmd-task | i-loc-town
0.25	-15.4468	matsumoto | nilsson
0.25	-15.4468	r1) | r1null
0.25	-15.4468	100% | (misspelled
0.25	-15.4468	loom | lauren
0.25	-15.4468	yt | ytp)
0.25	-15.4468	x12 | cargmaxsize
0.25	-15.4468	| | partitive
0.25	-15.4468	exp | insert
0.25	-15.4468	(37) | f+
0.25	-15.4468	p(w,c,n) | 50686f7370686f72796c6174696f6e
0.25	-15.4468	parenleftbign | parenrightbigsummationtext
0.25	-15.4468	a24 | a44
0.25	-15.4468	europarl | en-ar
0.25	-15.4468	msg | agr
0.25	-15.4468	(<1%) | turn-initial
0.25	-15.4468	window | 0.7170
0.25	-15.4468	ex | a,3
0.25	-15.4468	topi | __#
0.25	-15.4468	balti | vaalii
0.25	-15.4468	90 | 8∗
0.25	-15.4468	mpred | wpred
0.25	-15.4468	stop | .25
0.25	-15.4468	-/ε, | ο/ί,
0.25	-15.4468	fluency | full∩a
0.25	-15.4468	user | straight)
0.25	-15.4468	abstractness, | instr,
0.25	-15.4468	porter,wordnet | rules
0.25	-15.4468	loom | spencer
0.25	-15.4468	vg | tmod
0.25	-15.4468	nist | josh-base
0.25	-15.4468	(1 | bracehtipupright
0.25	-15.4468	(〈s | 〉,〈s
0.25	-15.4468	conll-x | pro3gres
0.25	-15.4468	a29a36a43 | a52
0.25	-15.4468	25% | avgsl
0.25	-15.4468	geneva | martigny
0.25	-15.4468	low-level | l=2
0.25	-15.4468	2,3] | 2,3
0.25	-15.4468	cj) | andlin(2006)
0.25	-15.4468	a29a36a38 | a52
0.25	-15.4468	+pp | −pp
0.25	-15.4468	ip | *pro*
0.25	-15.4468	dinu | w5
0.25	-15.4468	claim | (able
0.25	-15.4468	∧i | ds
0.25	-15.4468	strip | il-10
0.25	-15.4468	40.51 | iwslt,
0.25	-15.4468	p(θ|x
0.25	-15.4468	summationdisplay | yuv
0.25	-15.4468	2111 | x12
0.25	-15.4468	i-loc-town | i-time-date
0.25	-15.4468	topi | 
0.25	-15.4468	sigmab | promoter
0.333333	-15.4468	|f|summationdisplay
0.0315126	-15.4472	however, | to | a
0.0110429	-15.4474	in | an
0.0372493	-15.4477	does | for
0.102564	-15.4478	rather | only
0.0666667	-15.449	order | two
0.166667	-15.449	a30a10a23a8a11 | a9a8a11a9a10a12a13a14a12a15
0.166667	-15.449	view | es
0.333333	-15.449	current | prediction.
0.0231278	-15.449	the | of | should
0.0471698	-15.4496	will | or
0.0707071	-15.4498	also | well
0.0731707	-15.4502	30 | 70
0.181818	-15.4504	r-1 | r-2
0.0133961	-15.4506	the | is | it
0.0409556	-15.4508	only | can
0.0393443	-15.4508	language | using
0.333333	-15.4512	det | (labelled
0.102564	-15.4513	our | corpora.
0.039604	-15.4514	2002; | et
0.142857	-15.4515	larger | would
0.0693069	-15.4515	one | test
0.0909091	-15.4516	addition, | can
0.0102608	-15.4519	not | to
0.0545455	-15.4522	in | y
0.0821918	-15.4523	made | be
0.157895	-15.4531	built | it
0.0317797	-15.4531	dialogue | in
0.00739351	-15.4533	for | is
0.0523256	-15.4533	np | nn
0.25	-15.4535	central, | switzerland),
0.333333	-15.4535	dar | algu´em
0.333333	-15.4535	(u,v) | (u,w)
0.117647	-15.4539	on | score.
0.0821918	-15.4539	have | terms
0.0163845	-15.4539	there | in
0.0740741	-15.4542	be | e.g.
0.166667	-15.4546	trp | 32
0.0473934	-15.4547	f | r
0.0418251	-15.4553	from | feature
0.0523256	-15.4554	same | it
0.0229258	-15.4555	as | set
0.166667	-15.4557	nsummationdisplay | xijk
0.333333	-15.4557	73.26 | 34.05
0.333333	-15.4557	pr(s|n,v)
0.0542169	-15.4558	be | form
0.0238949	-15.4558	paper | to
0.0648148	-15.456	can | rule
0.0862069	-15.4566	we | integer
0.0821918	-15.4567	would | two
0.0269865	-15.4568	this | translation
0.0162866	-15.4575	the | them
0.0517241	-15.4575	models | our
0.057554	-15.4576	learning | not
0.0332542	-15.4579	that | probability
0.166667	-15.458	case−link | unigramcase−
0.0188542	-15.458	we | section
0.0514286	-15.4584	to | improvements
0.0514286	-15.4588	emotion | and
0.222222	-15.4602	eng1 | ita1
0.25	-15.4602	nc | narrative
0.25	-15.4602	clasifier, | single-layer
0.25	-15.4602	haitian-english, | 0.358
0.142857	-15.4602	may | words,
0.333333	-15.4602	semiring | associative
0.0862069	-15.4607	) | d
0.057554	-15.4607	described | this
0.117647	-15.461	on | modified
0.125	-15.4618	how | integrated
0.25	-15.4619	graphics | included
0.142857	-15.4621	first | comparable
0.333333	-15.4624	yvyd | polymerase
0.333333	-15.4624	boxshadowdwn | prob.
0.0289855	-15.4626	may | that
0.0722892	-15.4633	language | institute
0.0654206	-15.4634	training | has
0.0779221	-15.4635	be | low
0.0714286	-15.4636	multiple | can
0.0722892	-15.4643	how | language
0.121212	-15.4643	reading | which
0.333333	-15.4647	(z | 0.606
0.0152672	-15.4654	null | the
0.0117918	-15.4654	features | and
0.07	-15.4656	we | 2005)
0.0494505	-15.4659	were | our
0.222222	-15.4662	selecting | according
0.0847458	-15.4662	work | relation
0.0610687	-15.4667	in | method,
0.0520231	-15.4669	for | articles
0.166667	-15.467	)) | nulllic
0.0282392	-15.467	of | reference
0.0281924	-15.4674	be | but
0.0198347	-15.4675	is | of | word
0.00866337	-15.4676	in | by
0.0532544	-15.4682	side | in
0.0625	-15.4691	surrounding | the
0.0184006	-15.4692	the | also | to
0.0673077	-15.4693	preliminary | in
0.00564379	-15.4694	the | for | and
0.181818	-15.47	araucaria | of
0.25	-15.4704	distractor, | relatum,
0.25	-15.4704	null2, | null5
0.0877193	-15.4704	will | context
0.00904746	-15.4706	we | to | in
0.02138	-15.4709	this | be | in
0.114286	-15.4711	first | trained
0.1	-15.4711	figure | curve
0.333333	-15.4715	headword | collocation’s
0.333333	-15.4715	interfaculty | institute
0.0224839	-15.4718	two | as
0.25	-15.4721	kilometers | miles)
0.0810811	-15.4725	it | task.
0.0252324	-15.4726	are | from | of
0.114286	-15.4727	thus | at
0.0465116	-15.4729	2 | or
0.0759494	-15.4729	query | which
0.0511364	-15.4732	the | usually | to
0.0961538	-15.4734	464 | 424
0.0731707	-15.4735	w | r
0.333333	-15.4738	china | 60736014,
0.0597015	-15.474	3 | 9
0.0398671	-15.4741	we | further | the
0.0810811	-15.4742	dc | dc
0.0642202	-15.4744	set, | with
0.0402685	-15.4746	that | words.
0.0666667	-15.4752	be | related
0.0330189	-15.4753	to | grammar
0.102564	-15.4756	the | ter.
0.0418251	-15.4761	we | current
0.0555556	-15.4761	rest | as
0.25	-15.4772	wuv | c(u,v)
0.25	-15.4772	p1s1: | p1s4:
0.121212	-15.4775	(2); | dep
0.157895	-15.4777	contrast | work,
0.097561	-15.4782	tools, | and
0.333333	-15.4784	seon | youngjoong
0.01151	-15.4787	used | of
0.157895	-15.4788	followed | verb
0.0141255	-15.4789	were | of
0.0293578	-15.4791	we | from | as
0.022293	-15.4794	the | 10
0.0697674	-15.4794	that | false
0.045045	-15.4801	the | easy
0.00453146	-15.4808	and | to
0.105263	-15.481	an | finding
0.0257511	-15.4812	y | the
0.0666667	-15.4816	translation | over
0.0833333	-15.4818	semantic | sense
0.0114848	-15.4819	a | it
0.0789474	-15.4822	data | –
0.0220126	-15.4822	in | english
0.0383387	-15.4823	may | by
0.25	-15.4824	ﺖﻄﺒﺗرا | زوﺮﺒﺑ
0.142857	-15.4826	may | significant
0.0373832	-15.4829	represented | and
0.0551724	-15.4834	automatically | as
0.0108028	-15.4838	use | of
0.25	-15.4841	j∈ji | cijgn
0.25	-15.4841	trad-r | ∀j,k1[yi,j∗
0.0608696	-15.4845	in | refers
0.166667	-15.4853	all | grouped
0.333333	-15.4853	overview | the trec
0.333333	-15.4853	τ | 1−wt
0.108108	-15.4853	insight | to
0.0224359	-15.4856	to | about
0.0273312	-15.4856	of | 4
0.0438596	-15.4859	as | verb
0.25	-15.4859	support | correr
0.0418251	-15.4864	on | 2010).
0.0740741	-15.4871	6th | on
0.0269841	-15.4871	in our | we
0.25	-15.4876	(han) | (tetreaultparse)
0.333333	-15.4876	raising | 𝝓
0.0317125	-15.4877	to | pair
0.0146319	-15.488	if | in
0.0330969	-15.4885	a | phrases
0.0740741	-15.4886	is | sensitive
0.023753	-15.4889	the | initial
0.056338	-15.4892	there | all
0.00807003	-15.4893	with | to
0.333333	-15.4899	leacock-chodorow | bda-euclidian
0.114286	-15.49	/ | c
0.12	-15.4901	(7) | be
0.0289331	-15.4906	of | corpus,
0.0369231	-15.4917	cs | d7 | cs
0.0847458	-15.4918	linguistics | science,
0.0318907	-15.4919	data | not
0.0456621	-15.4922	document | from
0.333333	-15.4922	(public | generalist
0.0170316	-15.4925	verb | the
0.0163749	-15.4925	in | has
0.25	-15.4928	(wrb | (nnp
0.0345745	-15.4928	have | word
0.0540541	-15.4929	since | used
0.0779221	-15.493	due | their
0.0497238	-15.4932	was | set
0.0740741	-15.4933	related | al.
0.0494505	-15.4938	given | table
0.0258621	-15.4941	the | added
0.108108	-15.4944	if | precision
0.0341207	-15.4946	2 | by
0.333333	-15.4946	denotes | hw.
0.00905923	-15.4948	the | different
0.0666667	-15.4955	example, | will
0.02331	-15.4957	this | than
0.0514286	-15.4959	an | while
0.0186428	-15.4962	will | a
0.0705882	-15.4965	sentiment | negative
0.0377358	-15.497	model | at
0.0847458	-15.4972	of | equally
0.0232288	-15.4976	show | a
0.0442478	-15.4976	lower | and
0.0511364	-15.4987	compare | for
0.25	-15.4998	rank | relatively
0.0636364	-15.5005	is | structure.
0.00850077	-15.5008	the | training
0.157895	-15.5008	dogbark | η
0.157895	-15.5008	95.1 | 95.1
0.0284192	-15.5009	whether | to
0.0434783	-15.501	used | same
0.0191235	-15.5012	association | computational | of
0.0128251	-15.5015	more | and
0.0523256	-15.5015	3 | using
0.2	-15.5015	almost | other
0.0436681	-15.5021	is | (e.g.
0.0280702	-15.5021	noun | a
0.0544218	-15.5028	they | this
0.0392157	-15.503	4 | are
0.0177112	-15.5031	parsing | the
0.25	-15.5034	(2000) | several
0.333333	-15.5039	synset | (marked
0.0239295	-15.5041	and | may
0.0349462	-15.5041	approaches | et
0.0383387	-15.5042	provide a | the
0.15	-15.5044	in: | international
0.0377358	-15.5045	meeting | association | 2011.
0.0559441	-15.5048	by | (e.g.,
0.102564	-15.5049	such | present
0.0284192	-15.5051	we | rules | the
0.0862069	-15.5053	has | through
0.0323326	-15.5054	comparable | and
0.181818	-15.5054	erabakia | ize
0.0847458	-15.5061	words | set.
0.333333	-15.5063	type: | same
0.0429185	-15.5064	based | features
0.0147689	-15.5067	the | and | at
0.0150559	-15.5068	they | of
0.25	-15.5069	three | domains:
0.12	-15.507	such | occur
0.0150778	-15.5075	we | training
0.0100634	-15.5075	an | in
0.142857	-15.5076	ganchev | al.
0.0583942	-15.508	experiment, | and
0.0544218	-15.5082	information | language
0.0511364	-15.5084	vector | with
0.0478723	-15.5084	f | to
0.333333	-15.5086	paske | because.
0.333333	-15.5086	(fp7/2011-2014) | no.
0.333333	-15.5086	(kani-zabihi | coles-kemp,
0.25	-15.5087	students, | globin,
0.097561	-15.5088	from | view
0.0211132	-15.5089	multiple | the
0.0714286	-15.5092	we | develop
0.25	-15.5104	g(v) | 1|θ)
0.0373832	-15.5109	meaning | to
0.0536913	-15.511	assign | a
0.333333	-15.511	expertaverage | problem%
0.0583942	-15.5114	no | yes
0.00617996	-15.5114	which | the
0.0559441	-15.512	model | probability | of
0.021215	-15.512	the first | to
0.0379747	-15.5124	this | so
0.222222	-15.5126	vbd+vbn | np-c
0.15	-15.5129	provides | these
0.0379747	-15.5129	translation | word
0.2	-15.5129	comment | type
0.0497238	-15.5132	proceedings | usa,
0.0779221	-15.5133	linguistic | can
0.0705882	-15.5134	are | string
0.0779221	-15.5139	linguistics volume | number
0.0383387	-15.5143	than | more
0.0603448	-15.5145	new | word
0.125	-15.5148	application | which
0.0355191	-15.5149	need | in
0.0163934	-15.5151	pair | the
0.0806452	-15.5152	for | section.
0.0237797	-15.5153	as | data
0.044843	-15.5155	conference | 2011.
0.222222	-15.5158	wmt08 | cs,
0.15	-15.5161	emotions | as
0.015873	-15.5163	task | of
0.0133911	-15.5168	these | a
0.0674157	-15.5171	pairs | =
0.0133386	-15.5171	of | where
0.0544218	-15.5179	higher | with
0.0436681	-15.5179	have | between
0.0308642	-15.5182	dependency | for
0.059322	-15.5188	expect | be
0.0106173	-15.519	the | based
0.0444444	-15.5192	english | et
0.25	-15.5194	(i;t | rem
0.0583333	-15.52	example | table
0.0514286	-15.5203	main | this
0.333333	-15.5205	100.0 | /people/person/place
0.333333	-15.5205	3.3 | 6:
0.333333	-15.5205	explanation | worth*
0.0224215	-15.521	the | selection
0.0174497	-15.5211	sentence | in
0.15	-15.5215	consider | verb
0.00459664	-15.5215	we | a
0.0181028	-15.5217	is | data
0.0224719	-15.5219	of | 2010).
0.0862069	-15.5219	maltparser | and
0.036036	-15.5219	shows | by
0.0333333	-15.522	we | time
0.0230415	-15.5224	show | to
0.0222222	-15.5225	of | very
0.25	-15.5229	pathway | entity/reaction
0.25	-15.5229	sentiwordnet | been
0.25	-15.5229	z | h(˜cn−1,˜cn,wn+2n−2)
0.142857	-15.5229	order | problem,
0.333333	-15.5229	gemeinsames | rechten
0.0325581	-15.523	a | scores
0.0654206	-15.5233	are | main
0.0491803	-15.5234	of | or | words
0.0285205	-15.5241	on | 2011.
0.0630631	-15.5248	proposed | this
0.0578512	-15.5253	event | which
0.333333	-15.5253	(you | get)
0.333333	-15.5253	p(ti | t|x)
0.0320366	-15.5254	correlation | of
0.0533333	-15.5257	not | text
0.0516129	-15.526	corpus | all
0.0508475	-15.5261	section | describe
0.25	-15.5265	scores | arcs
0.0681818	-15.5267	previous | used
0.0164223	-15.5276	this | that | a
0.0420168	-15.5279	our | accuracy
0.0123498	-15.5282	system | of
0.0277296	-15.5283	we would | to | the
0.0295276	-15.5288	form | to
0.181818	-15.5288	measures | highest
0.153846	-15.529	erabakia | erabaki
0.136364	-15.5291	that | 1999).
0.15	-15.5301	sample | shown
0.0483871	-15.5301	with | parser
0.166667	-15.5302	|d | |h)
0.075	-15.5302	computational | 37,
0.153846	-15.5302	linguistics | rue
0.00508015	-15.5305	the | is | and
0.0388693	-15.5314	work | or
0.0135468	-15.5316	of | corpus
0.0519481	-15.5326	compare | that
0.333333	-15.5326	ensembl | (77.3%
0.0357143	-15.533	that | methods
0.02	-15.5332	specific | the
0.25	-15.5338	detail | briefly
0.25	-15.5338	“items” | |g
0.0456621	-15.5339	were | data
0.0544218	-15.5348	and | towards
0.0394265	-15.5356	size | is
0.0536913	-15.5357	i=1 | of
0.0192791	-15.5359	we | the same
0.0294118	-15.536	+ | =
0.108108	-15.536	arc | for
0.0220022	-15.5361	and | work
0.0078336	-15.5365	if | the
0.125	-15.5365	to | ambiguity.
0.181818	-15.5367	.31 | .29
0.044843	-15.5368	whether | be
0.136364	-15.5371	exploratory | to
0.0228916	-15.5372	dataset | the
0.0877193	-15.5373	two | features.
0.00950495	-15.5376	we | the | on
0.0209581	-15.5388	score | and
0.0466321	-15.5391	, | =
0.0731707	-15.5393	which | sentence.
0.044843	-15.5394	number | at
0.0311804	-15.5397	et | 2005;
0.0714286	-15.5406	table | four
0.25	-15.5411	system, | achieves
0.0134365	-15.5418	the | standard
0.0209372	-15.5422	for | university
0.00458418	-15.5432	is | and
0.0769231	-15.5436	only | less
0.111111	-15.5439	against | set
0.102564	-15.544	words, | context
0.0182927	-15.5445	proceedings | the | 2011.
0.0125479	-15.5445	in | with | and
0.042735	-15.5445	can | applied
0.0603448	-15.5446	probability | null
0.019103	-15.5446	that | data
0.0404412	-15.5447	which | translation
0.166667	-15.5448	have | positions
0.0224972	-15.5458	consists | the
0.166667	-15.546	with | tweets.
0.115385	-15.547	phrase table | by
0.0266223	-15.5472	different | on
0.0305677	-15.5479	based | al.,
0.153846	-15.5482	vp@2 | pp@2null1
0.25	-15.5484	bertoldi, | moran,
0.0202952	-15.5491	a | any
0.028169	-15.5495	2 | on
0.333333	-15.5497	attack | role=attacker
0.136364	-15.5501	most | measure
0.0930233	-15.5502	size | language
0.25	-15.5503	52.4 | 64.5
0.25	-15.5503	p(z|c)
0.25	-15.5503	description | description.
0.01206	-15.5519	if | of
0.0909091	-15.5521	department | science,
0.333333	-15.5521	content-container | instrument-agency
0.15	-15.5521	(law | portland,
0.0666667	-15.5527	q | ∈
0.0275387	-15.5528	for | would
0.0356083	-15.5528	al., | 2009;
0.222222	-15.553	supertagger | on
0.102564	-15.5531	impact | model
0.00938124	-15.5531	such | the
0.0137812	-15.5542	we | =
0.0674157	-15.5554	paper, | our
0.0793651	-15.5557	almost | with
0.25	-15.5559	fo | l(u
0.0909091	-15.5569	achieve | over
0.0207305	-15.557	the following | of
0.333333	-15.5571	p(wk+1|wkk−n+2h−1−mgk+1)
0.0740741	-15.5574	consists | an
0.25	-15.5577	r-1 | r-l
0.0409836	-15.5582	for | compared
0.0315315	-15.5593	named | the
0.037037	-15.5595	and | ∈
0.25	-15.5596	years | work
0.333333	-15.5596	f(si→sj|t)
0.333333	-15.5596	types, | supervised
0.0625	-15.56	system | source
0.0292398	-15.5602	our | based
0.0930233	-15.5606	task, | our
0.166667	-15.5609	used | indicator
0.0135478	-15.561	that | than
0.0256798	-15.5612	2006; | al.,
0.0491803	-15.5617	we | summarization
0.0533333	-15.5618	∈ | do | ∈
0.0533333	-15.5618	work, | that
0.0390071	-15.5626	i | r
0.0304348	-15.5632	the | values | to
0.25	-15.5633	mt08 | 34.01
0.25	-15.5633	narration | 58.33
0.25	-15.5633	ja | perustuttava.
0.0757576	-15.5634	one | being
0.333333	-15.5645	argmaxynullnullnull | nullv
0.333333	-15.5645	uniquely | using
0.111111	-15.5645	naive/reweighted | :
0.222222	-15.5646	iwslt, | 70k
0.222222	-15.5646	_\)’, | \w*{w1}\w*_(\d+)
0.222222	-15.5646	abs, | rc
0.222222	-15.5646	s48s46s48s52 | s101
0.222222	-15.5646	iinofaeala | =⇒
0.222222	-15.5646	ﺔﯿﺑﺮﻌﻟا | with-rise
0.222222	-15.5646	53697465 | 50686f7370686f72796c6174696f6e
0.222222	-15.5646	nulllic | wants(x,
0.222222	-15.5646	〉,〈s | d,1
0.222222	-15.5646	nounsubj | <>
0.222222	-15.5646	sfp | ec
0.222222	-15.5646	s48s46s48s50 | s101
0.222222	-15.5646	0% | 8906
0.222222	-15.5646	maaranaa | (fist
0.222222	-15.5646	mods | multi-modifier
0.222222	-15.5646	bracketleftbig | angbracketrightbigg
0.222222	-15.5646	x1 | x2)
0.222222	-15.5646	𝒘 | 𝟐
0.222222	-15.5646	3670 | %)
0.222222	-15.5646	(da | ∼
0.222222	-15.5646	=+ | u105(x,y,turnm)
0.222222	-15.5646	u102(x,y,turnm) | =+
0.0155989	-15.5647	same | of
0.0348837	-15.5648	as | data.
0.0614035	-15.5652	dependency | sentence
0.25	-15.5652	target | “happy
0.040293	-15.5655	we | role
0.0295276	-15.5658	the | level of
0.0390071	-15.5661	final | to
0.25	-15.5671	match | patterns.
0.0625	-15.5675	avoid | and
0.0326633	-15.5678	0 | =
0.0625	-15.568	several | been
0.0103935	-15.5681	or | of
0.0473684	-15.5684	map | of
0.25	-15.5689	kanssa | commissioner
0.105263	-15.569	for | segments
0.333333	-15.5696	interaction | type.
0.0244898	-15.5705	the | computational | portland,
0.25	-15.5708	china), | bot
0.023199	-15.5712	can | which
0.115385	-15.572	work | adaptation
0.333333	-15.5721	18.18 | 12.83
0.333333	-15.5721	c1), | c2),
0.0536913	-15.5724	the | utility
0.0735294	-15.5725	6th | july
0.25	-15.5727	unfortunately, | significant
0.25	-15.5727	stock: | p1s4:
0.0161968	-15.5732	evaluation | of
0.181818	-15.5732	bleu | beam
0.166667	-15.5734	authors | comments
0.0174948	-15.5735	in | training | the
0.0438596	-15.5735	task, | in
0.0614035	-15.574	method | semantic
0.25	-15.5746	price[<500] | q:
0.05	-15.5746	if | is | this
0.0385965	-15.5749	using | than
0.0273504	-15.5753	our | used
0.0877193	-15.5765	most | al.
0.021097	-15.5765	describe | the
0.115385	-15.5781	pu | np
0.0203488	-15.5786	a | they
0.181818	-15.5787	2: | noun
0.0625	-15.5789	i | g
0.0888889	-15.5794	--- | -------
0.0533333	-15.5795	not | given
0.0452261	-15.5797	number | between
0.0631579	-15.5799	experiments | have
0.25	-15.5803	retains | use
0.130435	-15.5806	features | still
0.0150649	-15.5806	to | use
0.115385	-15.5808	some | difficult
0.0139201	-15.5813	the | word | is
0.0191972	-15.5816	of | 2
0.333333	-15.5822	street, | 100080,
0.0722892	-15.5823	a | subsequent
0.0336134	-15.5825	still | to
0.0387324	-15.5829	these | such
0.0331633	-15.5833	we | list
0.153846	-15.584	would | highest
0.0588235	-15.584	by | fact
0.0806452	-15.5843	performance | development
0.0347826	-15.5844	score | as
0.333333	-15.5847	pc | chc
0.037415	-15.5852	because | we
0.0473684	-15.5853	features | section
0.0288462	-15.5856	a | of | phrase
0.0309051	-15.5857	given | be
0.00820578	-15.5863	a | of | for
0.0263591	-15.5864	3 | is
0.00257704	-15.5868	the | a
0.333333	-15.5873	influenza |  have
0.0551181	-15.588	d7 | d7
0.0806452	-15.588	could | between
0.019214	-15.5881	example, | is
0.0454545	-15.5881	each | distribution
0.0608696	-15.5883	table | shows | our
0.181818	-15.5884	<subj> | <np>
0.0735294	-15.5888	our | relative
0.142857	-15.5888	clairlib | to
0.25	-15.5899	annotators | confused
0.25	-15.5899	action=build-city | game
0.0909091	-15.5899	for | sections
0.0414938	-15.5899	tasks | et
0.0429185	-15.5903	alignment | as
0.0578512	-15.5904	l | p
0.142857	-15.591	value | 1,
0.0292276	-15.5911	to | speech
0.25	-15.5918	ˆpnull | shortrightarrow
0.108108	-15.5919	0.086 | 0.086
0.0192644	-15.5921	we | a | an
0.0249267	-15.5924	evaluate | and
0.00836751	-15.5929	features | the | the
0.2	-15.5929	to | lse
0.0869565	-15.5932	relations | more
0.097561	-15.5933	sequence | using
0.25	-15.5937	matrix | z2,4i
0.0336134	-15.5941	higher | and
0.0329949	-15.5947	separate | the
0.0697674	-15.5949	this | cases
0.333333	-15.595	smb | smb/smt
0.0409836	-15.5956	latent | a
0.0588235	-15.5958	to | speakers
0.111111	-15.5959	speech | corpus
0.0268007	-15.596	data | be
0.0115373	-15.5961	the | and | each
0.0689655	-15.5962	they | translation
0.0120399	-15.5975	our | of | of
0.0306346	-15.5976	let | and
0.0280899	-15.5979	by | such
0.25	-15.5995	instead, | account
0.0681818	-15.5997	table | classifier
0.0481928	-15.5997	positive | are
0.0325	-15.6002	we | structure
0.059322	-15.6006	that | state-of-the-art
0.2	-15.6007	ein | gemeinsames
0.0484848	-15.6008	when | language
0.25	-15.6015	0.063 | 0.002
0.142857	-15.602	– | 242
0.0182685	-15.6025	target | of
0.0416667	-15.6028	can be | not
0.333333	-15.6028	new←[] | null):
0.00578218	-15.6031	we | for
0.142857	-15.6031	1.000 | 0.538
0.142857	-15.6031	hat | jungen
0.0113706	-15.6035	not | a
0.00796633	-15.6036	our | to
0.0120898	-15.6046	source | the
0.0806452	-15.6049	categories | et
0.00338763	-15.6057	of | in
0.0251969	-15.6059	we | a | based
0.12	-15.606	70 | 4:
0.0264463	-15.6064	be | if
0.142857	-15.6065	work | al
0.142857	-15.6065	well | (e.g.
0.115385	-15.6066	where | derived
0.046875	-15.6068	usually | in
0.0806452	-15.6068	pair | training
0.0481928	-15.6068	can | single
0.0519481	-15.607	1 | sentence
0.0659341	-15.6071	but | translation
0.25	-15.6073	m.d. | j.c.
0.0307329	-15.6073	assume | a
0.00608412	-15.6081	in | a | of
0.0746269	-15.6085	is | early
0.0159509	-15.6088	to | a | on
0.0888889	-15.6089	moreover, | et
0.25	-15.6093	andr´e | portugal
0.0806452	-15.6093	corpus | 2010).
0.0325815	-15.6101	to | example
0.0888889	-15.6103	have | be used
0.0714286	-15.611	and | linking
0.0534351	-15.611	note | all
0.0869565	-15.6115	problem | sentences
0.023166	-15.6129	1: | and
0.037415	-15.6131	a | that | but
0.142857	-15.6132	coarse-grained | on
0.333333	-15.6132	(which | classes).
0.166667	-15.6133	offolds | prating
0.0125954	-15.6133	language | of
0.00805827	-15.6137	the | have
0.0406504	-15.6142	with | task
0.0416667	-15.6143	is | 2.
0.0952381	-15.6144	be | temporal
0.181818	-15.6153	plots | that
0.0509554	-15.6154	suggests | in
0.142857	-15.6155	consists | candidate
0.0479042	-15.6156	from | model.
0.0305882	-15.6156	that | so
0.333333	-15.6158	involving | witness
0.333333	-15.6158	dots | dashes,
0.031477	-15.6164	information | al.,
0.130435	-15.617	learned | used
0.0395257	-15.6189	4 | results
0.0137804	-15.6194	and | language
0.0659341	-15.6196	have | document
0.0251969	-15.6196	for | it | a
0.166667	-15.6198	text | texts.
0.0769231	-15.6199	split | set
0.0177606	-15.6205	in | section | the
0.0598291	-15.6207	which | order
0.0606061	-15.6212	to | deal
0.115385	-15.6221	from | texts,
0.115385	-15.6221	3 | coreference
0.181818	-15.6225	level, | level.
0.0395257	-15.6226	of | linguistics:shortpapers,
0.0652174	-15.6227	also | any
0.0408163	-15.6241	i | s
0.130435	-15.6243	has | active
0.25	-15.6251	xb, | (but
0.25	-15.6251	depvsm | 0.47
0.0819672	-15.6251	(i) | (iii)
0.0191606	-15.6254	and | section
0.181818	-15.6254	consider, | example, | of
0.111111	-15.6255	integration | by
0.0255183	-15.6258	we would | to
0.0625	-15.6261	can | improved
0.0327869	-15.6261	or | word
0.0530303	-15.6262	that | side
0.0363036	-15.6265	it | number
0.0793651	-15.6285	score | more
0.12	-15.6288	wu | al.
0.0140982	-15.6289	while | of
0.25	-15.6291	sh | ra
0.333333	-15.6291	koc-combo | uedin
0.045	-15.6297	model | section
0.0473373	-15.6298	5 | work
0.136364	-15.6298	semantic | metric
0.0139826	-15.6298	figure | and
0.181818	-15.6298	mcccj | dict
0.0888889	-15.63	will | sense
0.027833	-15.63	word | not
0.0326087	-15.6303	with | also
0.0909091	-15.6306	for each | k
0.0869565	-15.6308	features, | or
0.25	-15.6311	per | cb
0.0445545	-15.6314	algorithm | an
0.333333	-15.6318	fairy | brothers
0.0714286	-15.632	this | before
0.0909091	-15.632	7 | paper
0.0408163	-15.6339	the | usage
0.0326975	-15.6344	has | used
0.333333	-15.6344	agr(ann1||ann2)
0.25	-15.6351	aspects | perform
0.181818	-15.6356	should | change
0.0888889	-15.6357	the | coercion
0.0952381	-15.6364	should | considered
0.0459184	-15.6366	set | table
0.0351438	-15.6369	of | edinburgh,
0.333333	-15.6371	decoder | zhang
0.0793651	-15.6374	all | error
0.0354839	-15.6379	improvement | to
0.00257773	-15.6382	we | the
0.166667	-15.6385	laboratory | department
0.136364	-15.6386	galley | al.
0.0746269	-15.6387	60 | 60
0.333333	-15.6398	cotton | political
0.0769231	-15.6408	underlying | for
0.0350318	-15.641	parsing | is
0.00948808	-15.641	the | use
0.166667	-15.6412	process | sentences.
0.25	-15.6412	window, | length-3
0.0793651	-15.6419	based | knowledge
0.0289256	-15.6424	accuracy | in
0.333333	-15.6425	one’s | pronoun)
0.0267857	-15.6428	like | for
0.029106	-15.6428	can | each
0.142857	-15.6429	learning | zettlemoyer
0.0383142	-15.6434	association | linguistics,
0.0209713	-15.6437	source | a
0.06	-15.6438	of | 500
0.02693	-15.6438	we | tree
0.00803276	-15.6442	the | from | and
0.0299539	-15.6448	class | a
0.0335196	-15.6448	= | which
0.0618557	-15.645	set | probability
0.0414747	-15.6451	each | sentences
0.0143516	-15.6452	then | of
0.0186283	-15.6455	using | is
0.0131134	-15.6459	in | words
0.0746269	-15.6472	tags | pos
0.0514706	-15.6476	observation | to
0.333333	-15.6479	merkmal, | rechten.
0.0618557	-15.6485	vectors | are
0.097561	-15.649	paper, | only
0.0674157	-15.6499	output | al.,
0.0470588	-15.6502	c | association | computational
0.0952381	-15.6503	while | recall
0.0102902	-15.6505	the | section
0.333333	-15.6506	people | government
0.333333	-15.6506	d | œ
0.0296804	-15.6516	and | 1.
0.0503145	-15.6517	with | them
0.0526316	-15.6521	our | even
0.0674157	-15.6527	d | o
0.0606061	-15.6529	(2) | be
0.181818	-15.6533	cn | 
0.0378788	-15.6534	including | et
0.2	-15.6534	unigramcase− | case
0.033241	-15.6543	among | the | and
0.0503145	-15.6546	like | their
0.0952381	-15.655	regularization | is
0.0952381	-15.6558	bootstrapping | on
0.0674157	-15.656	+ | 2
0.0383142	-15.6562	is | a | distribution
0.00950226	-15.6563	we | that | to
0.0300231	-15.6565	each | at
0.0684932	-15.6569	2002), | of
0.25	-15.6575	49 | 45
0.0310263	-15.658	is | class
0.0225764	-15.6586	a | for | which
0.0952381	-15.6589	0.00 | event
0.0298851	-15.6589	that | possible
0.333333	-15.6589	event, | often
0.333333	-15.6589	viterbi | al.
0.0380228	-15.6605	word | language
0.0909091	-15.6609	on | parser.
0.103448	-15.6614	been | due
0.1	-15.6616	error | reduction
0.0227273	-15.6624	by | not
0.0152672	-15.6625	6 | the
0.0252101	-15.6626	translation | et al.,
0.111111	-15.6634	2 | evaluator
0.0177994	-15.6638	this | or
0.0757576	-15.6642	reasons | to
0.333333	-15.6644	yt, | vl,
0.333333	-15.6644	sides | rules);
0.0564516	-15.6647	natural | with
0.0227577	-15.665	the | is | figure
0.0638298	-15.6651	model | function
0.0181369	-15.6657	high | the
0.0420561	-15.6659	as | how
0.0434783	-15.6659	however, | et
0.0150376	-15.666	is | as | a
0.0481928	-15.6663	models | than
0.0164948	-15.6666	of | results
0.078125	-15.6666	first | best
0.103448	-15.6674	in | normalisation
0.0319149	-15.6683	this | they
0.0215311	-15.6688	1 | with
0.0522388	-15.6691	corpus | translation
0.0564516	-15.6698	constraints | as
0.0118662	-15.6699	be | by
0.1	-15.6699	0.859 | +
0.0247678	-15.6699	supported | the
0.333333	-15.6699	overview | bilingual
0.2	-15.6699	〉,〈s | e,1
0.2	-15.6699	··· | si+1
0.2	-15.6699	yif | thementionsarebothpronouns;elsen
0.2	-15.6699	(83, | (91,
0.2	-15.6699	flu | 4.22
0.2	-15.6699	bool | 1b-zz
0.2	-15.6699	vn | -y
0.2	-15.6699	leche | milk
0.2	-15.6699	stem+ | +morph2
0.2	-15.6699	3-tuples | single-word
0.2	-15.6699	λ1=0.2 | λ1=0.8
0.2	-15.6699	czc8 | cv
0.2	-15.6699	<wf | <wf
0.2	-15.6699	10] | ①
0.2	-15.6699	92.4 | 94.6
0.2	-15.6699	a1h | dda
0.2	-15.6699	rel-lda1 | rel-lda1
0.2	-15.6699	percentoftokenalignments | percentoftokenalignments
0.2	-15.6699	(+5): | literature,
0.0869565	-15.6699	mostly | on
0.0168006	-15.6707	semantic | to
0.0218978	-15.6709	of | described
0.0465116	-15.6711	into | translation
0.00396715	-15.6716	we | and
0.0334347	-15.6717	other words, | the
0.0542636	-15.6719	semantic | have
0.0136342	-15.672	the | of our
0.25	-15.672	f) | ﻞﻌﻟو
0.0930233	-15.6722	follow | al.,
0.0194363	-15.6723	its | a
0.0833333	-15.6727	were | due
0.08	-15.6732	v | all
0.1	-15.6733	large | better
0.133333	-15.6733	– | 69.5
0.0159337	-15.6734	for | used
0.0684932	-15.6745	0.01 | 20
0.047619	-15.6751	we | respect
0.0793651	-15.6752	between | entities
0.0930233	-15.6754	proceedings of | fifth
0.0152047	-15.6755	and | but
0.047619	-15.6755	a | tweets
0.097561	-15.6756	method | error
0.0240602	-15.6759	1 | as
0.0184049	-15.6759	data | is
0.25	-15.6762	adding | increased
0.153846	-15.6763	thus | lower
0.016	-15.6768	is | an | of
0.25	-15.6783	x←extract | bracketleftbigtask
0.0490798	-15.6794	language | university
0.0735294	-15.6798	two | parsing
0.0625	-15.6799	report | with
0.0174603	-15.6799	the | compared
0.0259965	-15.6806	in | terms
0.0441989	-15.6814	3: | on
0.125	-15.6814	the | penalty,
0.039823	-15.6817	paper | from
0.0168993	-15.6818	the | speech
0.166667	-15.6825	automatically | content
0.12	-15.683	coverage, | to
0.0380228	-15.6833	s | r
0.0816327	-15.6836	two | labeled
0.0735294	-15.6841	can | larger
0.2	-15.685	2. | real
0.0183888	-15.6855	are | an
0.125	-15.6856	using | leads
0.025641	-15.6865	to | following
0.25	-15.6867	<structure | viscus>
0.25	-15.6867	w | mislabeled
0.333333	-15.6867	empty | (*pro*)
0.333333	-15.6867	~ | in(on)
0.2	-15.6867	beatrice | marcinkiewicz.
0.0144308	-15.6872	an | as
0.0113523	-15.6882	a | this
0.25	-15.6889	machine | kauchak
0.333333	-15.6895	1|a
0.0909091	-15.6898	rather | use
0.133333	-15.6901	0.760 | +
0.12	-15.6902	computed | average
0.107143	-15.6907	provides | into
0.0345912	-15.691	as | dependency
0.0163701	-15.6915	all | is
0.2	-15.6918	system: | i
0.130435	-15.6919	(6); | (1);
0.166667	-15.6924	dpmm−0.1−0.01 | a71a71
0.021767	-15.6925	set, | the
0.181818	-15.6929	based | construct
0.0514706	-15.6935	this | still
0.0216606	-15.6937	use | et
0.0286344	-15.6937	sense | to
0.0219638	-15.6938	or | be
0.0542636	-15.6944	workshop | language
0.0869565	-15.6948	similar | work
0.25	-15.6952	ganz | interrogative
0.125	-15.6952	experiments | including
0.0514706	-15.6953	prior | that
0.0432692	-15.6963	our | they
0.0416667	-15.6964	does | with
0.0203644	-15.6967	to | then
0.0383142	-15.6967	use | such
0.2	-15.6969	kalean | kale
0.25	-15.6974	bleufeats | ext
0.25	-15.6974	ˆt | argmaxt
0.0507246	-15.6976	presented | this
0.0888889	-15.6977	evaluation | correlation
0.045977	-15.6979	evaluation | translation
0.125	-15.6984	if | always
0.0125	-15.6987	= | = | of
0.0675676	-15.699	sentences | parallel
0.0724638	-15.7	np | s
0.0542636	-15.7004	two | over
0.133333	-15.7004	descriptors | and
0.0566038	-15.7006	for | character
0.0612245	-15.701	text | other
0.0418605	-15.7013	to | named
0.0952381	-15.7022	figure | showing
0.0135988	-15.7028	is | each
0.0888889	-15.7031	to | equivalents
0.0292135	-15.7037	sets | in
0.2	-15.7038	ball | meters
0.00491803	-15.7048	we | the | to
0.0869565	-15.7053	in | ).
0.2	-15.7055	our | arc-eager
0.142857	-15.7063	they | separate
0.333333	-15.7067	brbo | :={}
0.030303	-15.707	on | both
0.0271845	-15.7071	consider | to
0.027897	-15.7074	the | of | 10
0.0212766	-15.7078	are | such
0.0510949	-15.7084	workshop | machine
0.0393013	-15.7088	the | of | chosen
0.0684932	-15.7089	pair | source
0.0851064	-15.7098	vs. | we
0.0306122	-15.7098	is | p
0.05	-15.7098	sense | be
0.1	-15.7099	means | such
0.25	-15.7103	k-best | parser
0.0414747	-15.711	for | make
0.0952381	-15.7121	scenarios | of
0.0694444	-15.7124	tested | for
0.125	-15.7124	improve | use
0.0304569	-15.7124	is | candidate
0.333333	-15.7125	affine | w
0.0152068	-15.7129	the | higher
0.0530303	-15.7132	if | does
0.0185185	-15.7133	one | the | for
0.0784314	-15.7133	(10k) | paper
0.0851064	-15.7134	consider | its
0.0278373	-15.7137	it | at
0.00954882	-15.7143	all | of
0.107143	-15.7147	(2010) | words
0.25	-15.7147	−→v | (−→ni⊗−→nj)
0.0357143	-15.7153	edinburgh, | uk,
0.166667	-15.7154	rusty | sky
0.153846	-15.7158	also, | number
0.0414747	-15.716	on | then
0.0156352	-15.7164	different | to
0.142857	-15.7174	group; | of
0.0869565	-15.7174	translation | 2003).
0.0503597	-15.7176	are | dialogue
0.0374532	-15.7179	based | word
0.029703	-15.7183	of | linguistics, | june
0.0142388	-15.7186	and | one
0.25	-15.719	(1999) | similar
0.0116831	-15.7195	a | and | that
0.0496454	-15.7195	(2) | with
0.0158311	-15.7204	translation | to
0.25	-15.7212	1450-1550 | 1550-1650
0.25	-15.7212	?s | geo:state
0.133333	-15.7212	the | dutch.
0.0684932	-15.7214	manually | each
0.0684932	-15.7214	seen | table
0.0205011	-15.7215	for | both
0.0675676	-15.7225	wi | is
0.0394737	-15.7226	this | focus
0.0224404	-15.7236	is | most | the
0.0545455	-15.7236	5 | used
0.333333	-15.7241	cause(coreentity|event)?,
0.0367647	-15.7244	also | such
0.2	-15.7247	sciences | delaware,
0.2	-15.7247	sep, | gainful
0.0470588	-15.7248	(1) | are
0.0462428	-15.7262	has | set
0.0294118	-15.7264	model | training
0.0425532	-15.7265	classification | al.,
0.2	-15.7265	bahutapara | banaa
0.0909091	-15.7268	(y | 
0.0444444	-15.7275	may | it
0.00980392	-15.7277	= | and
0.2	-15.7282	“left: | difficult”,
0.125	-15.7289	proceedings | 23-24,
0.0816327	-15.7295	were | who
0.25	-15.73	dersc | 98.6
0.0286976	-15.7301	vector | to
0.0492958	-15.7306	negative | with
0.0576923	-15.7308	our | research
0.0724638	-15.731	from | instead
0.25	-15.7322	coscos | iine
0.0361011	-15.7323	same | from
0.166667	-15.733	measures | (i.e.,
0.0178891	-15.7332	that | such
0.0909091	-15.7332	better | model.
0.0391304	-15.7335	more | at
0.2	-15.7335	overlapping | pair
0.0204082	-15.7343	with | can
0.0754717	-15.7349	to | pronouns
0.2	-15.7353	utterance | user’s
0.00898729	-15.7356	a | of | as
0.333333	-15.7359	rose-rank | rose-regpos
0.333333	-15.7359	chooses | distribution.
0.166667	-15.736	(sogrl) | p:
0.12	-15.7361	health | trials
0.015544	-15.7366	have | that
0.125	-15.7366	[1,4] | ▼
0.00883838	-15.7369	the | such
0.0329341	-15.7382	s | 1
0.0632911	-15.7387	also | them
0.0325444	-15.7388	7 | a
0.2	-15.7389	grammar | smith
0.2	-15.7389	area2 | manchester
0.181818	-15.7389	computer | amherst
0.0301508	-15.7393	s | e
0.00615258	-15.7402	in | that
0.2	-15.7407	0.356 | 0.346
0.0295567	-15.7408	starting | the
0.0769231	-15.7409	x | 0
0.0053139	-15.7418	that | a
0.333333	-15.7419	bulls | bulls.
0.333333	-15.7419	phrases | alone.
0.0108108	-15.7444	on | is
0.0304569	-15.7444	suggest | the
0.0289157	-15.7447	over | as
0.0618557	-15.7449	(1); | dep
0.333333	-15.7449	67.4 | −strength
0.0273109	-15.745	results | that | in
0.0232558	-15.7452	information | as
0.0316092	-15.7452	5 | that
0.00889178	-15.7453	table | the | of
0.0724638	-15.7458	version | al.,
0.0218281	-15.746	pos | of
0.0291262	-15.746	on | models
0.0343643	-15.7466	on | most
0.0196292	-15.7468	for | based
0.133333	-15.7473	:[c,x | :[c,x
0.2	-15.7478	meaning | input.
0.25	-15.7479	sales | shoulders
0.0632911	-15.7483	those | than
0.0784314	-15.7487	φ | y
0.0346021	-15.7487	if | all
0.00836979	-15.75	our | a
0.0754717	-15.7511	interest | discourse
0.0479452	-15.7523	can | similar
0.042328	-15.7526	not | been
0.0403587	-15.7527	moreover, | and
0.0769231	-15.7527	of | rte
0.0612245	-15.7528	this | four
0.0145858	-15.7529	is | words
0.0318841	-15.7537	and | us
0.0327381	-15.7542	contains | in
0.0503597	-15.7544	as | combination
0.0434783	-15.7546	conclusions | this
0.0271967	-15.7549	systems | et
0.0174978	-15.7553	that | using
0.0967742	-15.7558	consider | most
0.0123457	-15.7558	and | set
0.0754717	-15.7566	conclusion | new
0.133333	-15.7569	numbers | respectively.
0.333333	-15.7569	correct | readings
0.12	-15.7577	target | monolingual
0.0582524	-15.7577	significant | between
0.028436	-15.7578	• | that | the
0.0327381	-15.7587	corpus | be
0.0221607	-15.7595	2, | the
0.0632911	-15.7598	a29 | a29
0.0432432	-15.76	structure | that
0.0155091	-15.7605	in | section
0.0869565	-15.7608	of | consequence
0.0311615	-15.7612	methods | on
0.0888889	-15.7614	part | science
0.0416667	-15.7615	that | did
0.25	-15.7615	slot3, | −60.6,complval
0.25	-15.7615	four | fields
0.0560748	-15.7624	model | set.
0.1	-15.7624	date | to
0.2	-15.7624	f1 | 64.3
0.0110147	-15.7629	and | this
0.166667	-15.763	2. | parallel
0.333333	-15.763	explore | settings
0.0769231	-15.7639	difference | each
0.05	-15.7642	information | more
0.2	-15.7642	wn-1f | correlation
0.181818	-15.7652	(11 | robbery
0.181818	-15.7652	ñ, | 
0.181818	-15.7652	0.587 | nr
0.181818	-15.7652	oo | %%
0.181818	-15.7652	freq: | pairs)
0.181818	-15.7652	hdl | yale
0.181818	-15.7652	distance | jaro
0.181818	-15.7652	pasil | ▼
0.00703837	-15.7653	the | more
0.0437158	-15.7654	a | consists
0.0909091	-15.7669	example | :
0.00648968	-15.7683	the | between
0.0173762	-15.7685	same | as
0.0324484	-15.7688	search | is
0.133333	-15.7691	𝑣 | 𝑝
0.0196937	-15.7693	event | of
0.0185185	-15.7695	the | pages | oregon,
0.2	-15.7697	−i,z | excluding
0.0364372	-15.7699	to | need to
0.028169	-15.7706	6 | we
0.1	-15.7707	,0 | 0,
0.0250447	-15.771	scores | in
0.0146252	-15.771	this | can
0.0526316	-15.7713	like | thank | their
0.0357143	-15.772	word | than
0.028777	-15.7721	features, | of
0.0105721	-15.7724	a | of | are
0.0632911	-15.7726	them | from
0.25	-15.773	della | 56/c
0.0816327	-15.7737	roth, | et
0.0276596	-15.7747	we | corpus.
0.0555556	-15.7748	achieved | with
0.166667	-15.7753	sequence | speech
0.0521739	-15.7753	approach | models
0.0754717	-15.7753	undirected | and
0.333333	-15.7753	)φ | m−12
0.333333	-15.7753	progressive-verb | wh-adverb
0.09375	-15.7753	for | terminal
0.0888889	-15.7761	has | seen
0.0149626	-15.7761	given | to
0.0449438	-15.7761	for | e
0.0704225	-15.7763	topic | based
0.0740741	-15.7767	conclusion | proposed
0.153846	-15.7767	type | named
0.0566038	-15.7774	(2010) | et
0.333333	-15.7784	p(y,s|w)
0.0120733	-15.7793	data | to
0.0193548	-15.7805	if | as
0.0967742	-15.7807	for | dictionary,
0.0472973	-15.7812	sets | from
0.115385	-15.7817	a | correctly.
0.0535714	-15.7817	algorithm | can
0.0218579	-15.7826	of | label
0.0632911	-15.7826	compare | which
0.115385	-15.7828	listed | 1.
0.1	-15.7837	is | essential
0.0632911	-15.7837	which | phrases
0.0451613	-15.7838	aspect | of
0.00798983	-15.7841	that | be
0.333333	-15.7846	syntactic constraints | paraphrases
0.333333	-15.7846	was | gardening,
0.0594059	-15.7853	can | viewed
0.125	-15.7857	work, | evaluate
0.0352113	-15.7858	of | mention
0.0434783	-15.7858	even | with
0.166667	-15.7862	trained | amounts
0.0227964	-15.7865	of | like
0.0165877	-15.787	since | and
0.115385	-15.7882	models | within
0.2	-15.7883	various | made
0.0666667	-15.7884	both | based
0.0784314	-15.7884	evaluation | given
0.00548321	-15.7892	of | are
0.166667	-15.7893	connectives | level
0.0174672	-15.7894	more | is
0.08	-15.7894	team | in
0.142857	-15.7899	scores | clusters
0.2	-15.7902	fi∈f | g(x)
0.0595238	-15.7905	of | collection.
0.333333	-15.7908	hai | bahuta
0.142857	-15.7913	1 | 2)
0.0406091	-15.7914	shows | each
0.0218281	-15.792	context | in
0.153846	-15.7925	are | forms.
0.153846	-15.7925	(2005) | parsing
0.0283019	-15.7935	in | find
0.2	-15.7939	common | ones.
0.00446302	-15.7944	to | the | to
0.0545455	-15.7946	be | representation
0.0144665	-15.7948	between | a
0.166667	-15.7956	r., | m.
0.0234375	-15.7956	we | to | data
0.00498398	-15.7959	in | of | and
0.25	-15.7963	outscopes | (j,i)
0.103448	-15.7966	number | tweets
0.0215633	-15.7976	shown | table
0.117647	-15.7985	the | filter.
0.016835	-15.799	annotation | of
0.2	-15.7996	hierarchy | nonstative-v
0.0833333	-15.8011	--- | ------
0.153846	-15.8013	an | indirect
0.00650387	-15.8015	on | the | of
0.0486111	-15.8025	and | method.
0.0141343	-15.804	sentence | and | the
0.153846	-15.8042	types | coverage
0.0740741	-15.8045	summationtext | =
0.115385	-15.8046	with | standard.
0.0769231	-15.8049	424 | 344
0.0317003	-15.805	which | set
0.04	-15.8069	was | our
0.0138889	-15.8076	example, | and
0.0197216	-15.8076	on | it
0.25	-15.8082	library | historical
0.038835	-15.8085	found | are
0.153846	-15.8086	selected | well
0.133333	-15.8092	theme(protein) | theme(protein/event),cause(protein/event)
0.0609756	-15.8093	p5r5 | +**
0.0909091	-15.8095	text | similar
0.117647	-15.8097	ter | baseline
0.333333	-15.8098	data, | streaming
0.166667	-15.8099	empirical | test
0.0657895	-15.8104	be | often
0.0234114	-15.8104	as | figure
0.0266393	-15.8105	at | by
0.0109001	-15.8105	and | =
0.25	-15.8106	preference | xyzxyz
0.0571429	-15.8119	parameters | be
0.25	-15.813	constructions | alert),
0.2	-15.813	1-best | “leaving
0.2	-15.813	conclude | discussion | work.
0.117647	-15.8131	moreover, | will
0.0550459	-15.8137	as | word.
0.0495868	-15.8141	r | l
0.115385	-15.8145	x1 | i’m
0.0228311	-15.8146	to | relation
0.05	-15.8159	when | use
0.01481	-15.816	we | have
0.0545455	-15.8165	to | native
0.0769231	-15.8167	parser | al.
0.0323625	-15.817	et | 2008)
0.0378151	-15.8171	word | such
0.0714286	-15.8172	use | time
0.0666667	-15.8176	of | newswire
0.0222883	-15.8182	to | input
0.02019	-15.8185	such | al.,
0.142857	-15.8185	i | parenrightbig
0.142857	-15.8185	before | will
0.0242215	-15.8185	the | computational | 19-24,
0.0287958	-15.8187	shared | of
0.0560748	-15.8194	training | learning
0.0267857	-15.8194	to | likely
0.333333	-15.8194	brib | fasho
0.00692873	-15.8197	of | or
0.0754717	-15.8203	using | positive
0.2	-15.8207	نادﻮﺴﻟا | ةﺄﺸﻧ
0.038835	-15.8209	al., | al., 2010)
0.0186916	-15.8213	propose | the
0.00753643	-15.8225	a | a | is
0.111111	-15.8227	all | concepts
0.0258449	-15.823	work | from
0.117647	-15.8233	does | quality
0.0190275	-15.8235	words | for
0.0275229	-15.8235	of | translated
0.0617284	-15.8236	capture | are
0.0463576	-15.824	score | than
0.103448	-15.8244	as | measures.
0.0588235	-15.8244	each | represented | of
0.0574713	-15.8245	with | aligned
0.00796296	-15.8246	the | where
0.0165426	-15.8248	set | that
0.0816327	-15.8248	can | string
0.0857143	-15.8249	=+ | bel(x,
0.0412371	-15.8266	as | label
0.25	-15.8276	generation] | [draw
0.117647	-15.8279	for | prediction,
0.0701754	-15.828	training | corresponding
0.0566038	-15.828	attached | the
0.0833333	-15.8284	and | collections
0.037037	-15.8284	asr | the
0.0141928	-15.8285	the | could
0.0346154	-15.829	or | at
0.333333	-15.8292	allows | simultaneously
0.0111732	-15.8292	of | is | as
0.117647	-15.8302	cannot | applied
0.0382775	-15.8312	obtain | for
0.0290237	-15.8314	a | considered
0.0153097	-15.8316	1 | is
0.0714286	-15.8318	use | related
0.0358566	-15.8319	that | up
0.0151203	-15.8323	when | in
0.333333	-15.8324	terppara | fastr.
0.333333	-15.8324	belonging | scheme1
0.0857143	-15.8324	definition | such
0.0560748	-15.8336	can | classification
0.133333	-15.8337	k72 | k64
0.0211921	-15.8337	the | of | node
0.153846	-15.8339	system | better.
0.033557	-15.8343	from | has
0.040201	-15.8357	the | literature
0.333333	-15.8357	j=1 | j=i
0.2	-15.8363	entry | resource
0.0540541	-15.8367	we | starting
0.0460526	-15.8378	any | an
0.00515914	-15.8384	the | not
0.0769231	-15.8384	incidence | of
0.0574713	-15.8391	are | morphological
0.0156951	-15.8392	parse | the
0.00505708	-15.8398	that | in
0.0151934	-15.8403	we | feature
0.0119352	-15.8408	training | of
0.0240741	-15.8414	work | with
0.125	-15.8422	 | 
0.025	-15.8423	argument | of
0.25	-15.8423	yellow | myxedema,...
0.0568182	-15.8426	be | existing
0.0179641	-15.8436	that | training
0.0217077	-15.8437	in | human
0.0382775	-15.8456	when | at
0.0561798	-15.8478	divided | number
0.021097	-15.848	is | null
0.2	-15.8482	album | highlighted)
0.0406091	-15.8488	there | number
0.333333	-15.8489	return | estimation
0.0581395	-15.8499	these | shown
0.0701754	-15.8502	that | design
0.2	-15.8502	same | families
0.0967742	-15.8507	the | campaign.
0.0440252	-15.8511	segmentation | is
0.0182556	-15.8513	if | be
0.0545455	-15.8513	phase | of
0.25	-15.8522	robert | france
0.133333	-15.8522	author | thank
0.333333	-15.8522	international | s.p
0.166667	-15.8523	{plant, | η
0.166667	-15.8523	<cell | ...
0.166667	-15.8523	=> | smartcorrect
0.166667	-15.8523	agrnull | posnull
0.166667	-15.8523	cow, | instrument(with
0.166667	-15.8523	|dc| | |df|
0.166667	-15.8523	¯np | (rb
0.166667	-15.8523	sg | nom/acc
0.166667	-15.8523	treetagger | 1994)
0.166667	-15.8523	l(η;y,θ) | 12
0.166667	-15.8523	h23 | h41}
0.166667	-15.8523	||| | bien
0.166667	-15.8523	0.1385 | pseudo
0.166667	-15.8523	structure> | abdominal
0.166667	-15.8523	kleindienst, | mooney,
0.166667	-15.8523	hypertensive | disease.”
0.166667	-15.8523	α(i,j) | 
0.166667	-15.8523	6879 | 00004568
0.0284974	-15.8528	e | i
0.111111	-15.8533	0 | 11,11ensure
0.0249042	-15.8536	also | be
0.0701754	-15.8537	features | 4
0.166667	-15.8539	stemmed | feature;
0.0332226	-15.8539	word | between
0.0265487	-15.8543	our | to | from
0.0588235	-15.8546	al., | 2011).
0.25	-15.8547	affect | (whissell,
0.0130617	-15.8549	a | if
0.0191226	-15.855	to | performance | the
0.0314465	-15.8551	have | such
0.0386473	-15.8559	assumption | a | the
0.0491803	-15.8564	for | get
0.0337079	-15.8572	in | k
0.125	-15.8572	for | fluency.
0.166667	-15.8573	ts | #t
0.25	-15.8573	wpred | +wpartial
0.0406091	-15.8574	between | other
0.0574713	-15.8575	since | =
0.0291005	-15.8576	approach | al.,
0.08	-15.8578	authors | like
0.0228385	-15.8579	to | accuracy
0.2	-15.8582	corpus | acoustic
0.0568182	-15.8585	event | one
0.00602746	-15.8587	  |  
0.0223285	-15.8591	use | by
0.04	-15.8595	section | 2
0.117647	-15.8605	0.34 | 0.05
0.166667	-15.8606	centroid | (roi
0.0561798	-15.8607	table | overall
0.0568182	-15.8608	which | which | are
0.0153061	-15.8609	of | dependency
0.0454545	-15.8618	approach | than
0.0131234	-15.8619	about | the
0.25	-15.8623	def= | (λ,µ)
0.2	-15.8623	se | moins
0.025641	-15.8634	named | and
0.166667	-15.864	kiefer | anmachen
0.1	-15.8643	using | known
0.25	-15.8648	controversy | chilean
0.0535714	-15.865	given | same
0.0357143	-15.8655	section | this
0.031348	-15.8658	is | 2010).
0.0967742	-15.8662	compounds | for
0.0701754	-15.8671	and | w.
0.037037	-15.8671	evaluation | al.,
0.0311526	-15.8673	counts | of
0.25	-15.8674	operators | derivation
0.107143	-15.8677	this property | of
0.0118214	-15.8681	sentence | of
0.0316456	-15.8682	can | when
0.117647	-15.8689	in | cascade
0.111111	-15.869	vaalii | vaalaa
0.0625	-15.8693	(see | et
0.00750174	-15.8703	the | only
0.133333	-15.8711	k54k61k72k67k65k74k20 | k64
0.0310559	-15.8723	most | be
0.166667	-15.8725	phrases | parent
0.0159617	-15.8728	uses | the
0.037037	-15.8732	different | not
0.0495868	-15.8736	baseline | data
0.0857143	-15.8736	for | heads
0.117647	-15.8737	chunk | which
0.0320513	-15.8739	we | dialogue
0.0432099	-15.8742	by | possible
0.00318999	-15.8744	the | that
0.0530973	-15.8749	text | such
0.25	-15.875	yields | fewer
0.0689655	-15.8752	no | both
0.133333	-15.8765	instead | being
0.2	-15.8765	sentence-aligned | text
0.0727273	-15.8765	one | set.
0.0833333	-15.8767	from | taken
0.153846	-15.8772	evaluation | account
0.25	-15.8776	0.110 | (roi
0.0277078	-15.878	the | interest
0.0201511	-15.8801	work | by
0.00931762	-15.8802	each | to
0.166667	-15.881	step | boundaries
0.107143	-15.882	have | correctly
0.0228758	-15.8823	an | each
0.0432099	-15.8835	0 | 0 .
0.0204082	-15.8846	of | analysis
0.1	-15.8848	set | represents
0.025641	-15.8849	3: | and
0.0390244	-15.885	feature | all
0.25	-15.8853	pyysalo∗ | tsujii†
0.0882353	-15.8857	that | established
0.0277778	-15.8864	paper | on
0.0172084	-15.8866	in | there
0.0294118	-15.887	methods | as
0.0419162	-15.8875	most | can
0.0042173	-15.888	a | to | the
0.0188053	-15.8881	speech | the
0.107143	-15.8886	collects | the
0.0216049	-15.8898	= | if
0.153846	-15.8899	compares | results
0.0617284	-15.8906	layer | of
0.2	-15.8909	c | 
0.0754717	-15.8915	data | linguistic
0.0677966	-15.8916	for example, | when
0.0406977	-15.8922	that | random
0.0279188	-15.8927	in | obtained
0.0530973	-15.893	supported | science
0.166667	-15.8931	ner | trained
0.021875	-15.8932	on | between
0.0355556	-15.8938	2 | 7
0.0857143	-15.8939	a | post-processing
0.0727273	-15.8945	464 | 344
0.0588235	-15.8949	as | aligned
0.0351562	-15.8956	2: | a
0.0321429	-15.8957	proceedings of | uk, july
0.25	-15.8957	obtain | 2006).
0.0247934	-15.8959	of | contain
0.00914962	-15.8959	a | can
0.133333	-15.8972	they | represented
0.133333	-15.8972	theme(protein) | site(entity),csite(entity)
0.0203804	-15.8976	with | all
0.0784314	-15.8979	section | introduce
0.0581395	-15.898	according | or
0.166667	-15.8983	eθ | [w]
0.0151079	-15.8986	an | and | to
0.0213415	-15.8987	of | single | the
0.00939636	-15.8989	the | of | between
0.0214724	-15.9009	work | as
0.25	-15.9009	eθ | parenleftbigw
0.25	-15.9009	clusters, | word.
0.25	-15.9009	change | wrong
0.2	-15.9014	1011079100200 | withdrew-9
0.166667	-15.9018	tool | provided
0.25	-15.9035	0.289 | 0.239
0.2	-15.9035	parle | moins
0.2	-15.9035	about | importance
0.046875	-15.9044	related | be
0.0701754	-15.9056	2005; | 2010).
0.0214395	-15.9058	in | better
0.25	-15.9062	mild | baseline’s,
0.037037	-15.9066	supported | by
0.105263	-15.9074	0, | 6,
0.2	-15.9077	probabilities | classifiers
0.060241	-15.908	build | on
0.0180659	-15.9082	used | is
0.0129032	-15.9085	be | or
0.0223642	-15.9089	parallel | of
0.0123946	-15.9092	we | a | on
0.00570895	-15.9094	a | as
0.00684932	-15.9095	of | is | in
0.2	-15.9098	systems | explored
0.0568182	-15.9101	conclusion | on
0.0101798	-15.9105	one | and
0.166667	-15.9106	task | aligning
0.0461538	-15.911	section | experimental
0.0130081	-15.9111	of | based
0.0404624	-15.9112	= | r
0.0349345	-15.9114	word | where
0.0789474	-15.913	in | mst
0.0206897	-15.9137	addition, | the
0.0311526	-15.9146	linguistics | computer
0.0461538	-15.9169	correspond | a
0.0531915	-15.9171	training | examples
0.028133	-15.9172	e | n
0.0233393	-15.9172	is | data.
0.0458015	-15.9172	them | that
0.00848656	-15.9174	the | on | in
0.0411765	-15.918	so | by
0.0225694	-15.9187	evaluate the | of
0.111111	-15.9188	similar | assigned
0.25	-15.9195	after:, | 0.1879
0.0609756	-15.9204	as | monolingual
0.0113493	-15.9205	this | we | to
0.0375587	-15.9213	c©2011 association | department
0.0343511	-15.9216	table | baseline
0.0568182	-15.9222	that | word,
0.00734137	-15.9223	figure | the
0.2	-15.9226	narasimhan | zabin,
0.1	-15.9238	meaning | they
0.117647	-15.9242	would | too
0.25	-15.9248	29.76 | 13.07
0.25	-15.9248	at | past.
0.0833333	-15.9248	x|john | x|sees
0.111111	-15.9248	=> | out-of-the-box
0.0483871	-15.9253	than | best
0.2	-15.9269	nil | .0005
0.0288184	-15.9273	which | been
0.0040721	-15.9276	we | to | the
0.0245902	-15.9277	first | are
0.00969356	-15.9285	or | and
0.0058197	-15.9286	of | from
0.0294118	-15.9287	it | same
0.0543478	-15.9291	than | because
0.0124417	-15.9299	the | of this
0.166667	-15.9302	much〉 | x→〈x
0.166667	-15.9302	ˆpnull | enullnullnullnullnullnullnullnullnullnull
0.25	-15.9302	r-2 | 0.112
0.25	-15.9302	haitian-english, | (urdu-english)
0.153846	-15.9306	d | d≥1
0.0833333	-15.9311	cluster | two
0.046875	-15.9314	been | 2008;
0.09375	-15.9319	hold | to
0.153846	-15.9323	s51s48 | s49s48
0.153846	-15.9323	bool | dots
0.153846	-15.9323	.34 | .24
0.00989783	-15.9329	in | this
0.0124031	-15.9337	as | are | the
0.0322581	-15.9339	example, | from
0.142857	-15.9341	denotes | order
0.021374	-15.9345	set | be
0.2	-15.9356	therefore | developed
0.166667	-15.9357	(‘the | she
0.0789474	-15.9359	q | precedesequal
0.0119275	-15.9371	it | to | and
0.2	-15.9378	entries | umls
0.0441176	-15.9382	noun | or
0.125	-15.9383	form | 4
0.1	-15.9389	containing | sentences
0.0810811	-15.9389	through | using
0.0150376	-15.9392	shows | of | the
0.0833333	-15.9392	between | could
0.117647	-15.9395	10 | some
0.0810811	-15.9397	source-language | the
0.0330882	-15.9398	that | function
0.25	-15.9411	bilm | 43.12
0.133333	-15.9429	𝑣 | 𝑍
0.00579581	-15.943	the | all
0.0810811	-15.9433	while | dependency
0.032967	-15.9434	a | distance
0.2	-15.9443	transcriptions, | 5.77
0.0727273	-15.9443	has | precision
0.0308219	-15.9446	the | the | become
0.0754717	-15.9455	baseline | set.
0.0343348	-15.9459	syntactic | al.,
0.105263	-15.946	constraints | rules
0.0701754	-15.9464	5.1 | by
0.0108347	-15.9476	as | we
0.0419162	-15.9479	the | motivated
0.2	-15.9487	range | vary
0.2	-15.9487	ˆk1 | ,gk1
0.0113339	-15.9491	the | of | only
0.0191816	-15.9508	and | context
0.00997672	-15.9515	we | with | of
0.0444444	-15.9517	for | l | l
0.0380435	-15.9528	null | :
0.0833333	-15.953	at | previous
0.2	-15.9531	recently | 2005).
0.09375	-15.9534	one | words,
0.0909091	-15.9536	text | 2008).
0.0344828	-15.9542	to | utterances
0.0361991	-15.9553	2 | this
0.2	-15.9553	monotone | were
0.0857143	-15.9557	summationdisplay | .
0.0505051	-15.9576	in | entry
0.1	-15.9576	cap-micro | 0.086
0.25	-15.9576	oj, | oi.
0.133333	-15.9576	(2) | should
0.0833333	-15.9576	✓ | ✓
0.111111	-15.9576	randomly | data.
0.0227273	-15.9579	for | null | null
0.0229682	-15.9582	results are | the
0.00306574	-15.9589	of | and | the
0.0588235	-15.9595	was | evaluation
0.2	-15.9598	configuration | configuration
0.0857143	-15.9604	, | function
0.0531915	-15.9606	order | words | the
0.00365079	-15.9611	the | the | as
0.010598	-15.9614	= | = | and
0.0184332	-15.9614	are | our
0.0234375	-15.9625	are | translation
0.00381176	-15.9628	for | the | of
0.0198675	-15.9634	if | with
0.0458015	-15.9647	) | y
0.0145138	-15.9659	the | case
0.125	-15.9659	method | building
0.125	-15.9659	standard | evaluate
0.25	-15.966	vprime≤v≤|f|,
0.0428571	-15.9662	can | obtained
0.2	-15.9665	while | physical
0.0833333	-15.9669	thematic | and
0.0158451	-15.9673	contains | the
0.0150198	-15.9684	based | to
0.00533256	-15.9684	by | of
0.0196592	-15.9695	the | linear
0.033195	-15.97	words | using
0.111111	-15.97	million | girl
0.125	-15.9701	also | having
0.0271003	-15.9706	the | university
0.166667	-15.9707	clust, | 84.21
0.0447761	-15.9732	to | choose
0.0160428	-15.9732	with | an
0.0199715	-15.9736	results | by
0.00229017	-15.9736	in | the | the
0.0581395	-15.9739	not | account
0.00784563	-15.9743	can | to
0.0625	-15.9751	opinions, | and | are
0.0296053	-15.9752	this | dependency
0.2	-15.9755	interested | having
0.0261905	-15.976	this | most
0.0318021	-15.9761	word | translation
0.0431655	-15.9774	following | al.,
0.0402299	-15.9775	was | words
0.2	-15.9778	sampo | tomoko
0.0615385	-15.9787	note | use
0.0789474	-15.9789	like | comments.
0.25	-15.9801	task | duc
0.0081948	-15.9807	is | in | in
0.0233918	-15.9809	on | has
0.037037	-15.9809	s | o
0.0252294	-15.9817	we | vector
0.0574713	-15.9822	language | parallel
0.0701754	-15.9825	model | individual
0.0343348	-15.984	2010) | a
0.0155172	-15.984	their | a
0.25	-15.9858	nnps | rbr
0.0174482	-15.9859	we | performance | the
0.0967742	-15.986	a | dependency-based
0.075	-15.9864	all | category
0.0112108	-15.9868	than | of
0.0366492	-15.9869	fully | of
0.0769231	-15.9879	supervised | features
0.166667	-15.9896	shooter’s | shooter’s
0.0217028	-15.9897	an | set
0.0625	-15.9901	to | wi
0.0338983	-15.9905	threshold | of
0.0120361	-15.9905	on | from
0.037234	-15.9909	performance | than
0.25	-15.9915	1,2 | bielefeld
0.0206186	-15.9921	be | such
0.0139665	-15.9922	and | based
0.0731707	-15.9924	can | left
0.0350877	-15.9925	for | generate
0.0515464	-15.9926	even | which
0.166667	-15.9935	a29a36a29 | a47
0.0458015	-15.9938	conclusions | future
0.0125068	-15.9939	but | the | of
0.142857	-15.9948	describe | about
0.0224719	-15.9958	we | systems
0.075	-15.9959	experiments | paper
0.2	-15.9961	therefore, | done
0.0549451	-15.9968	3 | models
0.25	-15.9973	mz | quantity
0.25	-15.9973	was | ].
0.0112511	-15.9974	the | and | our
0.049505	-15.998	with | under
0.0211382	-15.9985	7 | of
0.0187266	-15.9994	of | pair
0.0360825	-16.0019	the | incorporate
0.0561798	-16.0023	would | been
0.0151261	-16.0024	we are | the
0.00733083	-16.0026	is | to | in
0.2	-16.003	mst+malt | malt+mst
0.25	-16.0031	active | owns
0.0549451	-16.0032	based | work
0.0882353	-16.0047	those | its
0.0655738	-16.005	4 | best
0.142857	-16.0064	89.0 | 87.6
0.142857	-16.0064	;�w | �
0.00346928	-16.0074	the | in | and
0.09375	-16.0078	effectively | with
0.037234	-16.0081	, | 2
0.0180072	-16.0083	of | baseline
0.0731707	-16.0085	using | present
0.03125	-16.0092	meeting | for | 19-24,
0.0238095	-16.0095	as | both
0.0952381	-16.0098	{a,b} | b
0.0112057	-16.0104	in | used
0.0366492	-16.0106	can | source
0.0882353	-16.0109	o: | a29
0.2	-16.0124	authors, | respect
0.0164251	-16.0128	the | oregon, | 2011.
0.0561798	-16.0141	| | np
0.0731707	-16.0154	to | ideas
0.0520833	-16.0154	/ | -
0.0810811	-16.016	algorithm | following
0.0154242	-16.0162	not | be
0.0358974	-16.0164	features | but
0.0561798	-16.0167	7 | 3
0.166667	-16.0168	milk | cuajada
0.133333	-16.0171	syntactic | produce
0.0236052	-16.0172	(1) | in
0.0103528	-16.0176	the | word | a
0.166667	-16.0187	ahmed | qatar
0.00652504	-16.0187	the | one
0.0909091	-16.0191	particular | but
0.2	-16.0194	network | diameter,
0.166667	-16.0207	∂λ | ∂
0.2	-16.0218	∂xbleu | ∂sil
0.125	-16.0221	automatically | boundaries
0.0247191	-16.0223	be | any
0.0177305	-16.0232	but | that
0.05	-16.0236	built | for
0.00232558	-16.0236	we | the | the
0.0251142	-16.0256	on | natural
0.075	-16.0266	paper, | will
0.075	-16.0266	vector | (svm)
0.0655738	-16.0266	probability | )
0.0235546	-16.0269	e | s
0.0391061	-16.0275	in | 2005;
0.0441176	-16.0276	been | 2009).
0.0437956	-16.0276	needs | to
0.04	-16.028	edit | and
0.142857	-16.0283	are | eval
0.00255935	-16.0287	a71 | a71 | a71
0.00679934	-16.0289	1 | the
0.25	-16.0296	partition | ={a⊆v
0.0282132	-16.0304	with | source
0.049505	-16.0306	at | sentences
0.0615385	-16.031	three | results
0.0241758	-16.0311	the | optimal
0.105263	-16.0329	window | 2
0.0268097	-16.0331	there | or
0.00782123	-16.0331	have | of
0.142857	-16.0334	1.0 | .21
0.0113261	-16.0339	a | used
0.0253165	-16.0348	similar | is
0.026455	-16.0355	in | corpora
0.25	-16.0356	touch | sony
0.2	-16.0361	buy | kids,
0.0505051	-16.0364	compute | with
0.0526316	-16.0373	pairs | number
0.1	-16.0374	their | must
0.0606061	-16.0375	recommendations | necessarily reflect
0.0526316	-16.0379	as | resulting
0.0247191	-16.0385	of | annual meeting
0.00849934	-16.0386	be | as
0.00563218	-16.0389	word | the
0.030888	-16.0409	at | than
0.0597015	-16.0409	upper | on
0.0882353	-16.0424	can | path
0.0192044	-16.0425	this | number
0.0416667	-16.0431	in | approach.
0.0515464	-16.0431	of | methods,
0.0810811	-16.0431	based | 2007).
0.25	-16.0446	precision@10 | knowledge1
0.0485437	-16.0455	an | or | that
0.0351759	-16.0456	example, | have
0.1	-16.0458	represent | source
0.0374332	-16.0459	on | terms
0.0344828	-16.0465	this | novel
0.25	-16.0477	2005) | focused
0.111111	-16.048	n=10,m=20 | –
0.2	-16.0482	augment | knowledge
0.166667	-16.0487	1|t,k,r | 1|t,k)
0.2	-16.0507	fung | yee,
0.0458716	-16.0516	disambiguation | for
0.0247191	-16.0523	where | from
0.2	-16.0531	cash | caja
0.2	-16.0531	identification | approach.
0.2	-16.0531	amigos | decidido
0.0217786	-16.0539	the | annual | of
0.142857	-16.0542	-93 | -87
0.117647	-16.055	model | probability.
0.0301003	-16.0552	note | as
0.2	-16.0555	ngd-google | ngd-yahooboss
0.142857	-16.0559	comparable | yee,
0.166667	-16.0568	maaraa | go)
0.166667	-16.0568	understanding | expressions
0.0327869	-16.0571	models | which
0.00571707	-16.0575	the | words
0.142857	-16.0577	languages | expressed
0.0625	-16.0584	component | al.,
0.0198171	-16.0586	to | approach
0.111111	-16.0588	system | domain.
0.04	-16.0595	if a | it
0.25	-16.0599	training | limited.
0.0833333	-16.0599	air | contract
0.0304183	-16.0614	at | al.,
0.035533	-16.0615	when | the | model
0.0235043	-16.0616	while | be
0.111111	-16.0616	translated | where
0.011963	-16.0629	the | and | features
0.166667	-16.063	point | per
0.25	-16.063	hearst | pattern
0.25	-16.063	∂l2∂f | ∂θ1
0.0116044	-16.0631	in | =
0.0434783	-16.0635	measure | we
0.0262467	-16.064	precision | to
0.0597015	-16.065	our | showed
0.166667	-16.0651	+negative | 73.5
0.2	-16.0654	p(e|f),
0.2	-16.0654	i− | |h
0.0148392	-16.0656	the | 5
0.0230608	-16.0656	can | at
0.25	-16.0661	repeat(u1), | bnullcc
0.0276074	-16.0667	a | translations
0.0462963	-16.0679	trained | which
0.0454545	-16.0691	graph | be
0.0571429	-16.0697	each | must
0.045045	-16.0698	note that | from
0.0480769	-16.07	60 | 80
0.0547945	-16.07	of | merged
0.0909091	-16.0709	these | hence
0.166667	-16.0713	stories | not
0.023913	-16.0725	as | test
0.0300752	-16.0732	we | significantly
0.0344828	-16.0732	proceedings | empirical methods
0.142857	-16.0736	performs | articles
0.111111	-16.074	where | logical
0.0135714	-16.075	linguistics | of
0.0382166	-16.0751	language | was
0.166667	-16.0754	algorithm | knn
0.133333	-16.0754	s49s48 | s32s75s76s83s72
0.133333	-16.0754	=+ | bel(b,
0.133333	-16.0754	pp[from] | (1);
0.133333	-16.0754	hdl | hdl
0.133333	-16.0754	s49s48 | s32s83s83s72
0.133333	-16.0754	s49s48 | s32s40
0.133333	-16.0754	with”) | xj-word=“with”,
0.133333	-16.0754	5353 | 5345
0.00489396	-16.0755	the | word
0.0343137	-16.0768	and | 2011.
0.2	-16.0778	here, | labeled
0.0882353	-16.0783	source | bilingual
0.25	-16.0785	gain | (dzikovska
0.0186418	-16.0785	have | by
0.166667	-16.0796	identifying | input
0.166667	-16.0796	n-gram∈i | metric:
0.0200927	-16.0811	whose | the
0.0731707	-16.0812	vbd | np
0.117647	-16.0813	components | different
0.0588235	-16.082	j | j | ∼
0.0222672	-16.082	the | refer
0.0909091	-16.0822	other | (and
0.2	-16.0829	[draw | arg]
0.0216998	-16.0838	baseline | to
0.00761308	-16.0847	other | the
0.00726539	-16.0863	as | is
0.0344828	-16.0871	having | and
0.0731707	-16.0877	and | community.
0.0588235	-16.0879	subject | be
0.25	-16.088	lexicon | word-forms,
0.0282132	-16.0887	ratio | of
0.0462963	-16.0888	dynamic | in
0.0300752	-16.0891	event | that
0.142857	-16.0898	f-measure | wtm
0.00318967	-16.0898	the | is | of
0.00637339	-16.0905	for | we
0.2	-16.0905	took | it
0.0253165	-16.0907	bilingual | of
0.25	-16.0911	m2> | return
0.00886308	-16.0915	the | not | in
0.0336538	-16.0915	document | are
0.0338164	-16.0919	their | an
0.166667	-16.0922	#words | liberal
0.0789474	-16.0927	such | verbs
0.0148731	-16.0928	for | if
0.0148862	-16.0928	that | language
0.0606061	-16.093	2. | model
0.0126024	-16.0939	this | it
0.25	-16.0943	haifa | linguists,
0.166667	-16.0944	by | january
0.00889298	-16.0949	the | first
0.0833333	-16.0954	for | specifically
0.011079	-16.0956	the | null
0.166667	-16.0965	kinds | class
0.0203735	-16.0966	scores | to
0.075	-16.0969	again | for
0.00478215	-16.0976	of | on
0.025641	-16.0981	carried | the
0.0810811	-16.0986	list | other
0.0533333	-16.0997	if | higher
0.0100725	-16.101	in | more
0.0101916	-16.1013	of | its
0.016269	-16.1033	the | is | language
0.125	-16.1039	focused | methods
0.00766017	-16.1049	in | or
0.2	-16.1058	polarity | regarding
0.2	-16.1058	kernelized | proposed
0.047619	-16.1062	1 | ≤
0.00898012	-16.1064	the | any
0.166667	-16.1072	mode | adjs
0.0810811	-16.108	(4) | number
0.142857	-16.108	brothers | tales
0.0224033	-16.1089	into | as
0.0519481	-16.109	parallel | training
0.057971	-16.11	a | pruning
0.0731707	-16.1102	used | paper.
0.125	-16.1103	lin | al.
0.0519481	-16.1104	shows | scores
0.2	-16.111	potassium, | argon,
0.0540541	-16.1126	is | national
0.0731707	-16.113	system | second
0.075	-16.1133	œ | √
0.0697674	-16.1151	of | users’
0.0625	-16.1152	each | out
0.0295203	-16.1153	on | pages | edinburgh,
0.0833333	-16.1158	sentences, | an
0.0195489	-16.1158	about | to
0.125	-16.1168	null3 | nullnull
0.25	-16.1169	hybrid | hybrid+named
0.166667	-16.118	present | increasing
0.166667	-16.118	perform | equation
0.017744	-16.1183	the | senses
0.0131488	-16.1187	an | of | is
0.00736888	-16.12	language | the
0.25	-16.1201	scopes | oj,
0.142857	-16.121	shared | submitted
0.00856974	-16.121	where | of
0.0188679	-16.1215	approach | is
0.125	-16.1217	similar | was | by
0.0159915	-16.1224	in | paper, we
0.25	-16.1234	king, | b¨ogel
0.075	-16.1241	shown | al.
0.0122026	-16.1243	there | a | the
0.0571429	-16.1244	same | when
0.166667	-16.1245	 ast |  rev
0.0402685	-16.1247	both | more
0.017284	-16.1258	to | syntactic
0.0311284	-16.126	of | simply
0.166667	-16.1267	iprime∧k | 1∧k
0.0731707	-16.1274	another | different
0.0155794	-16.1288	to | test
0.2	-16.1293	1|t,k) | p(k|t,r=1)
0.0299625	-16.1296	note | we
0.00894855	-16.1298	of | which | of
0.037037	-16.1301	different | model
0.0258621	-16.1303	to | defined
0.142857	-16.1304	para | sundaradikhataa
0.0358974	-16.131	annotation | we
0.0331754	-16.134	and | tagging
0.035533	-16.1346	most | word
0.166667	-16.1355	kill/beat | jaao
0.0952381	-16.1355	eyebrows | eyebrows
0.00349719	-16.1356	and | a
0.018732	-16.1359	as | = | =
0.023753	-16.1359	this | models
0.00557069	-16.1368	this | to | the
0.0228311	-16.1372	is | noun
0.2	-16.1372	these | labelled
0.166667	-16.1377	|x | φ
0.0681818	-16.1381	you | or
0.0416667	-16.1383	of | dialogue,
0.125	-16.1399	rouge-n-r | summationtext
0.125	-16.1399	pp[of] | (1);
0.125	-16.1399	t(h),t(sib),d | —
0.0446429	-16.1406	c©2011 | technology
0.0833333	-16.141	recently | al.
0.0247525	-16.1412	c©2011 association | for
0.00953429	-16.1418	corpus | of
0.0506329	-16.1422	dialogue | act
0.0666667	-16.1426	number | considered
0.2	-16.1426	generally | found
0.0789474	-16.1431	attempting | of
0.0215686	-16.1455	shown | figure
0.0420168	-16.1461	) | __ | )
0.0810811	-16.1464	using | unigram
0.0434783	-16.1481	this | systems.
0.166667	-16.1489	patterns | seeds
0.142857	-16.1495	encoding | prior
0.0105119	-16.1497	it | be
0.111111	-16.1503	rate | lower
0.00574943	-16.1511	● | ● | ●
0.0869565	-16.1521	amount | but
0.00443871	-16.1531	of | with
0.0833333	-16.1533	eat) | ate)
0.0191693	-16.1544	of | higher
0.0115005	-16.1553	of | also
0.026087	-16.1553	set | data
0.0193548	-16.1555	that | source
0.0115385	-16.1561	a | based
0.0127005	-16.1565	is | of | or
0.00933573	-16.1573	over | the
0.0189274	-16.1573	will | that
0.05	-16.1574	x | +
0.0139918	-16.1584	paper, | the
0.0228311	-16.1589	it | such
0.0180556	-16.1591	be | using
0.0324074	-16.1593	conclusion | this
0.0160085	-16.1593	we | most
0.0506329	-16.1599	with | derived
0.125	-16.1601	metaphorical | is
0.0158228	-16.1612	and | task
0.0638298	-16.1614	third | from
0.00880914	-16.1618	this | a | a
0.0423729	-16.1627	given | al.,
0.028777	-16.1629	or | each
0.125	-16.1635	for | roleset
0.075	-16.1642	22 | 0
0.0652174	-16.1645	type | information
0.0833333	-16.1647	our | confirm
0.166667	-16.167	filip | tapio
0.1	-16.1683	dependency | structure.
0.1	-16.1683	not | product
0.056338	-16.1691	combination | translation
0.045045	-16.1692	i | y
0.0512821	-16.1694	extending | of
0.2	-16.1697	r-1 | r-4
0.0714286	-16.1699	translation | according
0.0150376	-16.1701	language | is
0.125	-16.1703	sentm1 | hatwidestalign
0.022779	-16.1706	results | et
0.0136327	-16.1724	is | different
0.00363236	-16.173	the | by | the
0.00279638	-16.173	to | and
0.00584295	-16.1739	the | for | for
0.0191083	-16.1741	of | obtained
0.117647	-16.1743	taxonomy | from
0.010989	-16.1745	two | to | the
0.0588235	-16.175	machine translation, | uk,
0.00781047	-16.1754	that | by
0.00518193	-16.1757	table | the
0.1	-16.1766	probabilities | same
0.142857	-16.1768	its | edges.
0.0157068	-16.1777	we | that | can
0.0321101	-16.1782	articles | a
0.0280702	-16.1787	s | n
0.142857	-16.1788	size | limited
0.00931793	-16.1789	we | that | is
0.117647	-16.1791	total | entities
0.0124021	-16.1791	figure | a
0.0350877	-16.1798	that | performs
0.007768	-16.1798	of | are | and
0.0413223	-16.1804	4 | features
0.0413223	-16.1804	clustering | that
0.0519481	-16.1807	most | sentence
0.166667	-16.1808	analysis | despite
0.166667	-16.1808	encn | )
0.0277778	-16.1815	table | feature
0.010245	-16.1816	to | at
0.0216535	-16.1817	of the | linguistics,
0.0833333	-16.1819	second, | other
0.0236967	-16.1832	can see | the
0.2	-16.1835	les | j’
0.2	-16.1835	after:, | before:,
0.105263	-16.1837	two | require
0.0278746	-16.1842	many | be
0.00405424	-16.1846	as | of
0.0481928	-16.1847	this | represented
0.117647	-16.1857	features | few
0.2	-16.1863	6.42 | hybrid:
0.142857	-16.1867	frequency | had
0.117647	-16.1873	& | axioms
0.166667	-16.1878	still | strong
0.06	-16.1883	as | class.
0.0105213	-16.1889	in | number
0.0462963	-16.1893	then | end
0.0291971	-16.1897	is | classification
0.0512821	-16.1915	with | mean
0.00700786	-16.192	was | the | the
0.0205607	-16.1922	task | et
0.0909091	-16.1922	obtain | target
0.0147738	-16.1924	be | an
0.0413223	-16.1937	represented | as
0.029661	-16.1938	for | usa
0.117647	-16.1939	with | sense.
0.00442219	-16.194	are | of
0.0869565	-16.1941	possible reasons | the
0.0100349	-16.1942	of | sentence | the
0.0454545	-16.1953	adjective | and
0.0242588	-16.1957	this | source
0.0186335	-16.1961	since | for
0.0321101	-16.1971	figure | two
0.117647	-16.1973	acknowledgements | funded
0.117647	-16.1973	its | called
0.015748	-16.1975	and | also | the
0.125	-16.1982	systemalignment: | +morph1+
0.00694591	-16.1983	we | we
0.166667	-16.1995	seed | patterns.
0.0311111	-16.1996	conference | empirical | natural
0.2	-16.2003	soldiers | combatant.
0.08	-16.2009	clustering | similar
0.0612245	-16.2022	could | its
0.0201835	-16.2029	we | human
0.0252101	-16.203	5 | with
0.2	-16.2032	(look | (feel
0.042735	-16.2034	between | both
0.06	-16.2043	better | results
0.142857	-16.2048	proposed | (2002)
0.0638298	-16.2049	training | small
0.0215054	-16.2053	1 | by
0.0243243	-16.2061	figure | example
0.166667	-16.2066	reasoning | knowledge
0.00831601	-16.2071	semantic | the
0.2	-16.2089	r-1 | 10.4
0.00563574	-16.2094	to | we
0.0382166	-16.2104	our | alignment
0.00855513	-16.2106	is | not
0.142857	-16.2109	in | 2007a).
0.0833333	-16.2113	tell | a
0.0252809	-16.2121	bleu | to
0.0196429	-16.2124	is | i
0.00294907	-16.2138	a | to
0.125	-16.2142	bit | mepv
0.0588235	-16.2156	research, | in
0.166667	-16.2161	his | journalist
0.0423729	-16.2168	approach | both
0.0247934	-16.2169	english | is
0.0540541	-16.2183	and | li
0.111111	-16.2185	however, | derived
0.105263	-16.2195	g | primeprime
0.125	-16.2196	that | opinion.
0.0909091	-16.2207	the | v),
0.0869565	-16.2214	present | parallel
0.0547945	-16.2219	all | according
0.0321101	-16.2251	acknowledgements | the
0.0393701	-16.2265	4 | corpus
0.0423729	-16.2266	works | et
0.0333333	-16.2267	a | always
0.0714286	-16.2275	some | type
0.166667	-16.2282	substantially | baseline
0.0261438	-16.2287	set | can
0.0264026	-16.2288	have | language
0.1	-16.2291	example | sentence.
0.166667	-16.2306	word+nw | +lem+lem
0.0576923	-16.2308	to | realistic
0.00561719	-16.2313	it | to
0.0420168	-16.2314	experiments | language
0.142857	-16.2316	that | cultures.
0.105263	-16.2317	because | examples
0.0103578	-16.2318	in | such
0.0185759	-16.2326	specifically, | the
0.06	-16.2329	− | 1
0.166667	-16.2331	02-21 | mods
0.0454545	-16.2359	excluded | the
0.166667	-16.2379	(rule) | (x1
0.111111	-16.2379	in | rumor
0.0228426	-16.2385	of | had
0.00345139	-16.2389	in | the | in
0.0869565	-16.2391	section | algorithms
0.142857	-16.24	passvp(expressed) | actvp(derived)
0.0681818	-16.2403	experiments | experimental
0.166667	-16.2404	kkuthnayta | sicakhata/
0.125	-16.2415	%age | 88.9
0.166667	-16.2429	feature | w0
0.166667	-16.2429	hack | obama
0.0271186	-16.2433	this | al.
0.0403226	-16.2434	may | one
0.0143678	-16.2447	the | of | i
0.0769231	-16.2447	while | cases
0.125	-16.2452	any | semantics
0.08	-16.2461	experiments | mt
0.0164349	-16.2475	more | are
0.0137575	-16.2478	include | the
0.0588235	-16.2483	not | appropriate
0.0207039	-16.2499	a | found
0.0487805	-16.2501	we | increasing
0.0307018	-16.251	when | can
0.0247934	-16.2514	annual meeting | portland,
0.0363636	-16.2519	in | 2011).
0.2	-16.2527	divided | events
0.0227848	-16.2536	so | be
0.0588235	-16.2536	this | improving
0.00493963	-16.2561	the | on the
0.047619	-16.257	able | which
0.111111	-16.2577	 | udef_q_rel
0.111111	-16.2577	angbracketleftbigg | 
0.111111	-16.2577	á, | 
0.111111	-16.2577	κ | 0.71210
0.0125647	-16.2579	each | that
0.2	-16.2587	encouraging | using
0.0158014	-16.2589	to | 1
0.0203666	-16.2591	in | we | use
0.0697674	-16.2593	least | least
0.0381679	-16.2598	with | system.
0.166667	-16.2602	course, | many
0.0274914	-16.2606	r | i
0.111111	-16.261	cap-micro | approx-restr
0.0547945	-16.2616	we experimented | of
0.2	-16.2617	choosing | gives
0.0512821	-16.2623	these | event
0.0291667	-16.2633	a | four
0.0909091	-16.2636	black; | >
0.0740741	-16.2644	results | nist
0.04	-16.2648	user | are
0.037037	-16.2649	amounts | the
0.0184564	-16.2652	based | as
0.0235602	-16.2661	some | et
0.0204082	-16.2668	there | between
0.0283401	-16.269	not | translation
0.00920368	-16.2691	our | we
0.0248447	-16.2691	be | system
0.0506329	-16.2697	have | developed
0.166667	-16.2703	bad; | dull;
0.00508936	-16.2729	the | =
0.00950806	-16.2737	this | be | the
0.0869565	-16.2743	algorithm | sets
0.0566038	-16.2747	2007) | with
0.00569544	-16.275	to | a | a
0.047619	-16.275	set | 2
0.166667	-16.2754	when | request
0.0299145	-16.2756	to | 2008;
0.0255591	-16.2757	and | 2001; | and
0.0186757	-16.2762	in | types
0.0305677	-16.2766	2: | on
0.0260586	-16.2768	c | e
0.2	-16.2768	spans | contain
0.2	-16.2768	0.856 | 0.811∗
0.125	-16.2772	,d(z=k)j | (z=k)
0.166667	-16.2779	characters | bar
0.0869565	-16.2796	with | 5).
0.0215054	-16.2799	less | of
0.0714286	-16.282	see | have
0.105263	-16.2822	cosine | measure
0.0233161	-16.2822	it | been
0.0373134	-16.283	t, | of
0.2	-16.283	suggests | practical
0.043956	-16.2836	generates | a
0.166667	-16.2856	lack | amount
0.125	-16.2868	sentence, | sentences.
0.00404549	-16.2871	for | to
0.0240964	-16.2873	to | instead
0.142857	-16.2874	2010 | task,
0.0111306	-16.288	the | of | they
0.0465116	-16.2885	human | at
0.0625	-16.2888	(2010), | et
0.0294118	-16.2903	this | score
0.0288066	-16.2904	pairs | be
0.166667	-16.2908	“play | role”
0.0526316	-16.2911	means | or
0.0430108	-16.2925	languages | not
0.111111	-16.2933	our | detects
0.166667	-16.2934	j’ | des
0.0952381	-16.2941	from | segments
0.04	-16.2942	figure | dependency
0.0165062	-16.2945	new | a
0.0244648	-16.2951	word | all
0.2	-16.2954	shown | affect
0.00640404	-16.296	this | we
0.125	-16.2965	able | reasonable
0.0154578	-16.2969	the | reported
0.1	-16.2985	see | even
0.142857	-16.2985	two | pairing
0.166667	-16.2986	) | )v
0.024024	-16.2993	meeting | oregon, june
0.0512821	-16.3001	vector | using
0.0125294	-16.3005	is | same
0.0152009	-16.3005	it is | is
0.0137615	-16.3006	presented | the
0.0141509	-16.3007	section | a
0.0254777	-16.301	i | e
0.037037	-16.3014	will | model
0.105263	-16.3018	contributions | model
0.00979478	-16.3018	and | than
0.08	-16.3023	• | yellow
0.125	-16.3024	1,i | 1|t,k)
0.0454545	-16.3028	language | natural
0.0192308	-16.3045	different | as
0.0140187	-16.3057	3 | to
0.0952381	-16.306	into | right
0.166667	-16.3064	<var | direction="following"
0.04	-16.3073	6 | 3
0.0833333	-16.3077	are | terms.
0.00450318	-16.3083	to | that
0.0393701	-16.31	training | pairs
0.0697674	-16.311	the | squared
0.0107466	-16.3111	a | training
0.111111	-16.3125	of | trigrams.
0.0107223	-16.3126	we | between
0.037037	-16.3148	section | work.
0.08	-16.3149	into | models.
0.0638298	-16.3153	most | machine
0.0714286	-16.3155	a | transcribed
0.0387597	-16.3158	d | t
0.0493827	-16.3181	generate | which
0.0155502	-16.3181	(2010) | the
0.031746	-16.3186	and | expressed
0.021327	-16.3187	we | a | but
0.142857	-16.3188	since | replace
0.0519481	-16.3198	task, | are
0.00604767	-16.3201	by | to
0.0142276	-16.3202	a | of | features
0.125	-16.3202	(αprime,γprime) | pb(γprime|¯t,γ,
0.025	-16.3203	is | set.
0.0833333	-16.3209	experiments | domain
0.111111	-16.3213	combined | performance.
0.0769231	-16.3216	models | trained | all
0.0075	-16.3258	a | not
0.0331492	-16.3272	research | under
0.0352113	-16.3273	we | derived
0.111111	-16.3285	numbers | types
0.00420962	-16.3293	can | the
0.0111663	-16.3299	test | of
0.0133929	-16.3307	we | information
0.0170213	-16.3312	is | corpus
0.00732026	-16.3325	for | by
0.00381624	-16.3332	of | in | of
0.0714286	-16.3338	:= | r
0.0409836	-16.3341	words, | be
0.0189394	-16.3341	by | two
0.125	-16.3343	15.2 | 15.1
0.0689655	-16.3347	(dictionary | 20
0.0273438	-16.3354	data | than
0.00945946	-16.3369	the | 2
0.0714286	-16.3373	quality | models
0.125	-16.3383	xfer-orig | full-short
0.0952381	-16.3396	using | partial
0.0517241	-16.34	with | just
0.0139165	-16.3403	could | of
0.0110227	-16.3413	that | it
0.1	-16.3416	twitter | a | twitter
0.0129534	-16.3417	in | paper, | the
0.0714286	-16.3419	as | contexts.
0.0147727	-16.3425	is | 1
0.0273438	-16.3429	t | r
0.0251572	-16.3435	we | representation
0.0198807	-16.3443	in | 4.
0.0188679	-16.3453	when | we
0.0545455	-16.3459	our | comparing
0.00326505	-16.346	by | the
0.0258303	-16.3468	rules | with
0.0181518	-16.347	the | few
0.0470588	-16.3475	as | lower
0.0909091	-16.348	finally, | lexical
0.0272374	-16.3482	to | assigned
0.0545455	-16.3495	reported | al.
0.05	-16.3498	these | are | than
0.0506329	-16.3505	mi, | mi,
0.0135922	-16.3515	for | has
0.00713012	-16.3515	and | or
0.045977	-16.3516	mainly | in
0.166667	-16.3521	levels | analysis
0.0228571	-16.3532	of | b
0.00451977	-16.3553	the | not | the
0.0740741	-16.3557	classifier | over
0.0261194	-16.3562	acknowledgments this | by
0.0666667	-16.3565	serves | for
0.0740741	-16.3581	each | column
0.0170015	-16.3586	based | that
0.125	-16.3589	highest | scores.
0.111111	-16.3593	is | bike
0.0625	-16.36	solutions | a
0.166667	-16.3603	testing | selected
0.0152941	-16.3606	to | so
0.0666667	-16.3631	experiments | chinese
0.166667	-16.3631	additional | describe
0.166667	-16.3631	bum | you
0.1	-16.3631	w=0.1 | s=0.8
0.1	-16.3631	� | �=�
0.0249221	-16.3645	model | between
0.0131234	-16.3649	with | al.,
0.00382128	-16.365	from | the
0.0307692	-16.3651	of | usa,
0.0625	-16.3652	cd | np
0.1	-16.3664	valid | any
0.00494254	-16.3665	to | that | the
0.166667	-16.3687	xxw1 | (slop=2)
0.166667	-16.3687	improvements | absolute
0.0301508	-16.3706	contrast, | to
0.0324324	-16.3718	algorithms | et
0.034965	-16.3718	proposed | an
0.0195695	-16.3729	annual | of
0.125	-16.3735	directly | standard
0.06	-16.3741	∑ | j
0.0387597	-16.3761	30 | 50
0.0199115	-16.3766	semantic | et
0.111111	-16.378	order | guarantee
0.037594	-16.3782	were | model
0.00403701	-16.3783	in | we
0.0158311	-16.3785	table | by
0.00827636	-16.3796	for | is | to
0.020316	-16.3799	tokens | of
0.0267176	-16.3818	corresponds | a
0.0535714	-16.382	scoring | with
0.0384615	-16.3825	as | parser
0.0517241	-16.3839	sets | two
0.0566038	-16.3841	our | demonstrate
0.166667	-16.3856	solution | ecur
0.0252708	-16.3862	section | results
0.1	-16.3884	entry | where
0.00798669	-16.3892	= | is
0.142857	-16.3896	seems | so
0.08	-16.3901	complexity | algorithm
0.00876494	-16.3908	all | to
0.0869565	-16.3909	are | uniformly
0.0199557	-16.3912	each | number
0.0166667	-16.3912	is | table
0.0231884	-16.3915	compare | in
0.04	-16.3935	t | )
0.166667	-16.3941	[appeasement] | [appeasement]−
0.0769231	-16.3943	which | well.
0.00738427	-16.3949	we | the | which
0.0179856	-16.3954	we | on | from
0.00617038	-16.3957	in | we | a
0.0150376	-16.3962	is | figure
0.0952381	-16.397	specific | might
0.0833333	-16.397	building | an
0.0508475	-16.3984	we | valid
0.0301508	-16.3988	to | added
0.00539541	-16.3995	where | the
0.0430108	-16.4003	2 | algorithm
0.0215054	-16.4018	that | because
0.1	-16.4022	be | negative,
0.0689655	-16.4041	fifteenth | oregon,
0.0740741	-16.4047	summationdisplay | ∈
0.0769231	-16.4049	our | papers
0.08	-16.4052	annotation | may
0.00932401	-16.4053	no | the
0.0545455	-16.4058	words) | of
0.0769231	-16.4062	the | guesses
0.0909091	-16.4064	official | is
0.0261194	-16.4068	(koehn | al.,
0.142857	-16.4069	[x | α•yβ,h]
0.0164671	-16.407	relations | of
0.08	-16.408	all | t.
0.00838095	-16.4101	the | model | a
0.0952381	-16.4102	would | good
0.0116533	-16.4104	this | can | the
0.0131086	-16.4115	in | performance
0.0191083	-16.4116	learning | is
0.142857	-16.4118	counts | threshold
0.038835	-16.4135	french | for
0.125	-16.4143	vbd | vp-c
0.00509853	-16.4156	for | the | is
0.04	-16.4165	we | cross
0.0135841	-16.417	that | number
0.1	-16.4179	to | minimum.
0.1	-16.4179	scfg-hmd | over
0.0714286	-16.4182	this | occurring
0.125	-16.4187	set | predicates,
0.047619	-16.4194	∈ | be
0.1	-16.4197	mt, | is
0.0104287	-16.4213	language | a
0.166667	-16.4232	recent | potential
0.0769231	-16.4238	between | instance
0.0769231	-16.4238	also | combining
0.0454545	-16.4239	addition, | be
0.0340136	-16.4244	table | precision
0.125	-16.4253	way | performance.
0.0362319	-16.4273	learning, | the
0.00551194	-16.431	we | that | of
0.00693888	-16.4317	of | and | with
0.0689655	-16.432	introduce | used
0.0117739	-16.4321	and | test
0.00811509	-16.4339	a | words
0.0318471	-16.4346	test | used
0.142857	-16.4346	count(oi | oj
0.0364964	-16.4349	most | more
0.0170358	-16.4351	in | should
0.0129032	-16.4358	can | to | to
0.0555556	-16.4361	c©2011 association | learning
0.0612245	-16.4372	〉,int | 2
0.024735	-16.4373	between | each
0.0265487	-16.4389	workshop | portland,
0.0373832	-16.4394	(a) | (c)
0.0714286	-16.4398	same | set,
0.045977	-16.4403	similarity | sentences
0.0298507	-16.4406	decisions | the
0.0645161	-16.4413	terms | precision
0.04	-16.4414	for | percentage
0.0377358	-16.4422	is | explicitly
0.0491803	-16.4433	by | string
0.111111	-16.444	produced | statistical
0.0377358	-16.4449	the | depicted
0.0140086	-16.4473	same | a
0.0869565	-16.4477	3 | instances
0.0114123	-16.4479	in | will
0.00967742	-16.4496	and | such
0.166667	-16.4501	xij | max(i,j)
0.0119332	-16.452	in | based
0.166667	-16.4532	issue | domain
0.0909091	-16.4534	features | parameters
0.08	-16.4537	ordering, | the
0.0176367	-16.4554	they | with
0.0869565	-16.4556	terms: | the
0.0392157	-16.4562	is | explicit
0.0588235	-16.4565	can | limited
0.125	-16.4569	* | semantic
0.0508475	-16.4581	however, | al.
0.0209424	-16.4586	table | system
0.1	-16.4592	approaches | account
0.0205128	-16.4595	and | sentence.
0.0234899	-16.4606	the | relationship
0.038835	-16.4609	since | not | we
0.0250896	-16.4616	because | with
0.0625	-16.4627	gold | annotations
0.0714286	-16.4632	wm | a
0.0625	-16.4638	since | focus
0.125	-16.4638	said | good
0.1	-16.4647	current | model.
0.00698714	-16.4663	of | our
0.1	-16.4666	overall, | both
0.0331126	-16.4683	how | our
0.111111	-16.4684	stage | method
0.111111	-16.4705	technology | text
0.0412371	-16.4711	make | from
0.0207254	-16.4713	combining | the
0.166667	-16.4716	fusion | multistream
0.05	-16.4721	for | reliable
0.0126354	-16.4726	the | 2009).
0.00717489	-16.4728	of | between
0.00952381	-16.4739	of | these
0.0120586	-16.4739	are | al.,
0.0625	-16.4754	that | decreases
0.0171233	-16.4762	also | et
0.047619	-16.4764	for | learning | university
0.0140515	-16.477	the | main
0.0326797	-16.4784	has | of | has
0.142857	-16.4791	street | 1993).
0.0241379	-16.4807	possible | as
0.0447761	-16.481	named | entity
0.0344828	-16.4813	described | al.,
0.0625	-16.4824	results | chinese
0.125	-16.4824	nell | knowledge
0.0114155	-16.483	to | feature
0.05	-16.4843	gestures | the
0.036036	-16.4849	by | examples
0.047619	-16.4862	where | associated
0.142857	-16.4871	report | among
0.166667	-16.4872	merkmal | gemeinsames
0.0220386	-16.4875	and | described in
0.0318471	-16.4878	we | capture
0.0566038	-16.4886	the | spans.
0.166667	-16.4903	will | below,
0.0740741	-16.4906	before | used
0.00879894	-16.4908	for | not
0.0357143	-16.4912	related | language
0.0689655	-16.4916	acquisition | et
0.00475185	-16.4922	the | of | we
0.0107167	-16.4925	from | we
0.047619	-16.4925	nodes | between
0.037037	-16.4934	20 | 80
0.0119352	-16.4941	of | is | than
0.00964114	-16.4954	section | of
0.0555556	-16.4966	model | alignments
0.0232558	-16.4973	event | for
0.1	-16.4985	m/f | nom
0.0276498	-16.4985	two | than
0.142857	-16.5007	para | bahuta
0.0333333	-16.5011	(e.g., | al.,
0.0290698	-16.5016	links | and
0.0344828	-16.5022	or | those
0.1	-16.5023	(forall | defeasibly,
0.0689655	-16.5035	comparison | task
0.0430108	-16.5037	to | issue
0.0909091	-16.5044	0.08 | 0.12
0.0606061	-16.505	to | degradation
0.0166667	-16.5052	found | to
0.166667	-16.5062	0.66 | 0.40
0.166667	-16.5062	describes | used.
0.0157143	-16.5062	to | do
0.0164204	-16.5073	table | our
0.0412371	-16.5077	is | connected
0.0483871	-16.5077	from | role
0.00516917	-16.5078	number | the
0.0689655	-16.5088	dirichlet | topic
0.0320513	-16.5095	with | (see
0.0588235	-16.5107	system | takes
0.00821693	-16.5109	to | between
0.0212766	-16.5119	we | that | other
0.0197044	-16.5122	to | do not
0.166667	-16.5126	end | words,
0.00941423	-16.5128	the | single
0.0606061	-16.5132	the | container.
0.0353982	-16.5144	a | (or | of
0.00232128	-16.5151	a | and
0.1	-16.5158	represented | represented
0.0174419	-16.5164	sentence | be
0.166667	-16.5191	type="term" | <rel
0.166667	-16.5191	according | measures.
0.0322581	-16.5193	for | 2007;
0.0545455	-16.5213	improve | system
0.0208333	-16.5231	2 | from
0.030303	-16.5232	combine | a
0.039604	-16.5261	model | n
0.0299401	-16.5265	this research | by
0.0909091	-16.5274	method | texts.
0.03125	-16.528	task | this
0.0380952	-16.5292	evaluated | by
0.125	-16.5305	uses | task.
0.1	-16.5315	simplification | simple
0.0434783	-16.5316	minimum | rate
0.00294163	-16.5326	as | the | the
0.0124164	-16.533	a | that | are
0.0205656	-16.534	semantic | on
0.0625	-16.5342	responses | by
0.0909091	-16.5345	by | bar
0.0258621	-16.5353	computer | and
0.0344828	-16.5358	part | from
0.0909091	-16.5363	item | pair
0.00453487	-16.5384	and | et
0.0134228	-16.5395	word | be
0.0045403	-16.5401	we | of | in
0.0284091	-16.5402	department | science
0.012037	-16.5408	features | as
0.1	-16.5414	present | step
0.00659805	-16.5417	example, | the | the
0.0118745	-16.5418	is | used | the
0.0689655	-16.5426	comparability | of
0.0292398	-16.543	language | pages | 27–31,
0.0909091	-16.5436	training | follows.
0.00782123	-16.5441	the | models | the
0.0240964	-16.5454	instance | is
0.0909091	-16.5454	from | affected
0.08	-16.5454	phones | in
0.021148	-16.5468	of | next
0.1	-16.5474	p1 | p3
0.00773481	-16.5474	can | as
0.00917899	-16.5474	the | types
0.0282486	-16.5499	in | corresponds
0.0272727	-16.5505	better | on
0.0555556	-16.551	can | selection
0.0526316	-16.5517	do | when
0.0316456	-16.5524	(and | and
0.0428571	-16.5529	we | briefly
0.0214067	-16.5543	approach is | to
0.0252101	-16.5552	on | processing,
0.0571429	-16.5557	to | relate
0.00803213	-16.5562	to | words
0.0301205	-16.5562	data | number
0.0109091	-16.5582	1 | for
0.0102881	-16.5624	for | have
0.0240964	-16.5629	x | x1 | x
0.125	-16.563	sie:it | vorbei:by
0.0571429	-16.5638	task | system.
0.0126984	-16.5643	word | are
0.016129	-16.5646	the | social
0.1	-16.5656	increase | time
0.0833333	-16.5656	learning | improving
0.142857	-16.5656	p(z|o,c)
0.0213415	-16.5669	show | as
0.0216718	-16.5669	results | shown
0.018	-16.5672	under | of
0.0689655	-16.5677	either | these
0.0689655	-16.5691	citation | that
0.0909091	-16.5693	, | phone
0.00758945	-16.57	we | the | our
0.0138714	-16.5705	the | are | an
0.00228597	-16.5709	the | and | and
0.0130435	-16.5717	to | first
0.0336134	-16.5723	addition, | are
0.0461538	-16.5729	description | be
0.142857	-16.5744	selectively | al
0.142857	-16.5744	contexts, | should
0.046875	-16.5775	linguistics | word
0.037037	-16.5777	then | null
0.00184275	-16.5785	  |   |  
0.0508475	-16.5788	tests | for
0.0241935	-16.5792	and | paraphrase
0.111111	-16.5793	belong | any
0.00506867	-16.5798	the | and | by
0.0428571	-16.58	to | gains
0.0909091	-16.5805	idea | like
0.0183066	-16.582	proceedings of | annual
0.0447761	-16.5825	new | features
0.08	-16.5829	8 | accuracy
0.142857	-16.5832	a | hypergraph.
0.00146328	-16.5839	the | to
0.142857	-16.5862	refers | times
0.00771605	-16.587	to | or | the
0.00936123	-16.5879	to | from | and
0.0294118	-16.5881	given | our
0.0909091	-16.5881	showing | over
0.0193237	-16.5882	consistent | the
0.0508475	-16.5883	in | ner
0.0689655	-16.589	system | good
0.0540541	-16.5896	on | structures.
0.0833333	-16.5897	breast, | cancer,
0.0357143	-16.5907	user | which
0.0645161	-16.5916	distractors | the
0.142857	-16.5922	|v
0.0290698	-16.5924	is | correlation
0.0366972	-16.5947	retrieval | al.,
0.0147059	-16.5957	achieved | the
0.0666667	-16.596	has | reported
0.0606061	-16.5977	university | york,
0.00664932	-16.5985	the | of | more
0.0117012	-16.5988	the | tokens
0.111111	-16.6002	not | spanish.
0.023622	-16.6026	our | does
0.0136307	-16.6029	the | identify
0.0666667	-16.6031	s, | set
0.0153139	-16.6047	task | for
0.0290698	-16.6072	for | l
0.0625	-16.6073	represent | information
0.0526316	-16.6083	we | hits
0.0689655	-16.6094	1.000 | tagger
0.03125	-16.6099	alignment | be
0.012685	-16.6113	a | tree | the
0.0173913	-16.6115	since | as
0.125	-16.6126	k=s | ρkj(1)
0.031746	-16.613	linguistics:shortpapers, | june
0.142857	-16.6133	form | smt
0.0327869	-16.6139	would | data
0.0280899	-16.6139	knowledge | that
0.0177384	-16.6158	be | number
0.0327869	-16.616	consists | from
0.0483871	-16.6169	in | 3.2.
0.0239044	-16.6181	classes | in
0.111111	-16.6191	via | topic
0.0447761	-16.6194	associated | which
0.0689655	-16.6198	course | are
0.0282486	-16.621	(papineni | al.,
0.0251046	-16.6213	figure | different
0.111111	-16.6215	a) | k,b,c
0.0422535	-16.6221	uses | data
0.0228137	-16.6235	parsing | with
0.00548482	-16.6237	this | that
0.0410959	-16.6246	all | output
0.00575731	-16.6248	the | table
0.0769231	-16.6254	nullcc | bel(x,
0.00382519	-16.6262	the | this
0.0606061	-16.6275	links | words
0.0314961	-16.6277	given | at
0.0441176	-16.6285	proceeds | the
0.0416667	-16.6288	can | class
0.0428571	-16.6291	semantic | noun
0.036036	-16.6307	for | (in
0.0245902	-16.6313	as | set.
0.0625	-16.6315	learning | statistical
0.142857	-16.6319	annotated | marked
0.00680272	-16.6323	of | used
0.04	-16.6331	and | and/or
0.0606061	-16.6341	output | target
0.031746	-16.6343	present | an
0.0245902	-16.635	source | be
0.142857	-16.635	prophet | nns
0.0176211	-16.6352	the | follows.
0.0588235	-16.6352	3.2 | semantic
0.0347826	-16.6352	grammar | be
0.0571429	-16.6375	phenomena | by
0.0491803	-16.6378	de | ,
0.046875	-16.6384	whether | have
0.0075643	-16.6384	of | but
0.0740741	-16.6385	might | these
0.0252101	-16.6385	for | r
0.111111	-16.6385	fa | by
0.0526316	-16.6391	a | viewed
0.04	-16.6393	which | task.
0.0104244	-16.6396	the | research
0.0441176	-16.64	statistical | edinburgh,
0.0769231	-16.6405	enough | such
0.0833333	-16.6415	features | key
0.0666667	-16.6419	stage, | are
0.0625	-16.6438	mentioned | more
0.0158172	-16.6448	into | is
0.0233463	-16.6458	system | than
0.0909091	-16.6467	current-speaker(x)) | nulllic
0.0290698	-16.6476	empirical methods | edinburgh,
0.017316	-16.6483	previous | et
0.0230769	-16.6492	set | where
0.111111	-16.6507	3.2 | sense
0.0217391	-16.6507	these | can | of
0.0555556	-16.6508	verbs | they
0.1	-16.6508	that | explored.
0.0182648	-16.6511	all | not
0.011178	-16.6517	is | that | be
0.0909091	-16.6528	component | specific
0.0909091	-16.6528	see | outperforms
0.00901917	-16.6528	methods | the
0.0172786	-16.6532	the | developed
0.0125714	-16.6539	these | we
0.0135468	-16.6564	of | set.
0.0290698	-16.6566	on | in | scotland,
0.010906	-16.6573	a | will
0.0689655	-16.6576	lexical | wordnet
0.0148588	-16.6579	on | pages
0.00917958	-16.658	to | features
0.0441176	-16.6596	to | practical
0.142857	-16.6603	pos | chunking,
0.0035303	-16.6608	and | is
0.00513196	-16.6615	in | we | to
0.0338983	-16.6617	of | 2004)
0.1	-16.6619	system | segmentor
0.00359712	-16.6628	to | a | of
0.0276243	-16.6631	node | with
0.0169851	-16.6634	to | positive
0.0384615	-16.6645	linguistics | university,
0.0769231	-16.6645	fact | tend
0.0023783	-16.6656	in | to
0.05	-16.6675	figures | with
0.0769231	-16.6679	then | normalized
0.0540541	-16.6696	we | disambiguate
0.0276243	-16.6713	thank | reviewers
0.0336134	-16.6717	pos | table
0.00267221	-16.6724	of | for
0.04	-16.6732	a | empirical
0.0588235	-16.6745	partial | not
0.0137931	-16.6757	this is | a
0.0909091	-16.6774	(kim | 2009).
0.1	-16.6778	per | category
0.125	-16.6789	information | others.
0.125	-16.6789	all | interpolated
0.031746	-16.679	each | described
0.142857	-16.6797	system | mt05
0.142857	-16.6797	polarity | sentence.
0.05	-16.68	are | communication
0.0176211	-16.6817	a | less
0.0194986	-16.6819	dependency | et
0.142857	-16.683	feasibility | approach
0.0645161	-16.685	blog | from
0.011846	-16.6855	can | used
0.0310078	-16.686	work | two
0.0181818	-16.6861	are | sentences
0.111111	-16.6885	100.0 | (b)
0.047619	-16.6896	bilingual | target
0.111111	-16.691	student | was
0.0555556	-16.6911	requires | number
0.00914634	-16.6912	sentences | of
0.0394737	-16.6919	and | area
0.0100647	-16.693	a | it | to
0.125	-16.6933	>b | 
0.0137741	-16.6936	we | score
0.111111	-16.6936	showing | other
0.0248756	-16.6937	in | language | pages
0.026738	-16.6939	techniques | et
0.0909091	-16.6941	tree | insert
0.00322699	-16.6952	in | we | the
0.0689655	-16.6954	0.3 | 0.9
0.0064384	-16.6984	model | to
0.0625	-16.6991	model | estimation
0.0714286	-16.6996	ex | 〉,〈s
0.00878156	-16.7	1 | in
0.023166	-16.7012	the | natural | pages
0.0147541	-16.7022	important | and
0.0164948	-16.7023	al., | 2009).
0.0375	-16.7029	technique | on
0.00293076	-16.7032	that | the | of
0.0689655	-16.7034	follows: | (1)
0.0175055	-16.7041	be | between
0.0454545	-16.7047	each | considered
0.0141844	-16.706	this | system
0.0252525	-16.7067	association | institute
0.0428571	-16.7076	concatenation | the
0.0606061	-16.7083	university | pittsburgh,
0.0769231	-16.7106	nlu | that
0.125	-16.7108	task | bb
0.00543024	-16.7112	also | the | the
0.0833333	-16.7118	translation, | will
0.1	-16.715	current | 2010).
0.0144	-16.7155	machine | to
0.0714286	-16.7198	inserts | the
0.0110803	-16.7199	learning | the | of
0.00849708	-16.7203	on | are
0.0625	-16.7227	defined | sequence
0.0214286	-16.7228	edinburgh, | 2011.
0.0243902	-16.7238	section | related
0.0571429	-16.7238	the | ironic
0.111111	-16.725	output | item
0.125	-16.7257	marginal | word
0.047619	-16.7265	steps | are
0.0190217	-16.7275	computer | university
0.0173697	-16.7276	way | and
0.125	-16.7287	cover | null2)
0.0185185	-16.7295	and | compared
0.00851517	-16.7308	syntactic | the
0.00755985	-16.7313	the | is | not
0.0625	-16.7317	k | between
0.125	-16.7317	whether | associated
0.0222222	-16.7326	of | sets.
0.0139535	-16.7338	of | association | oregon,
0.0625	-16.7347	used | algorithms
0.125	-16.7347	outputs | mt
0.0909091	-16.735	0.27 | 0.26
0.0833333	-16.7357	in | portable
0.028777	-16.7374	cs | cs
0.0204082	-16.7374	of | together
0.0189702	-16.738	examples | the | in
0.0769231	-16.7382	combine | information
0.00300401	-16.7397	in | of | of
0.0153551	-16.7416	table | using
0.0167015	-16.743	parse | to
0.0155642	-16.7436	as | into
0.125	-16.7438	(imp) | represent
0.125	-16.7438	p(wi|wi−11
0.0229358	-16.7442	instead | we
0.0121547	-16.7456	in | would
0.0147783	-16.7465	more | on
0.0361446	-16.7468	small | which
0.125	-16.7468	address | automated
0.0375	-16.7469	as | clusters
0.0714286	-16.7474	baseline | f-score
0.0078125	-16.7532	approach | of
0.00463111	-16.7543	to | by
0.0625	-16.7546	dependency | 2006;
0.03125	-16.7568	also | sentences
0.0259067	-16.7583	examples of | of
0.0115546	-16.7587	performance | to
0.0625	-16.7592	sentence | several
0.00975244	-16.7592	of | al.
0.125	-16.7592	words | median
0.00609756	-16.7602	table | and
0.1	-16.7611	prediction | annotated
0.0327869	-16.762	average | number
0.111111	-16.7657	contribution | type
0.031746	-16.7662	the | (we
0.0909091	-16.7663	sky | path
0.0833333	-16.7664	q∗t(f) | 0.046
0.00474576	-16.7666	can | the | of
0.0214286	-16.7683	pages | uk,
0.1	-16.7685	have | expressions.
0.111111	-16.7685	probability | ∝
0.0144462	-16.7692	for | machine
0.0298507	-16.7692	attribute | and
0.0105727	-16.77	for | linguistics
0.0606061	-16.7701	iterated | the
0.0223214	-16.7702	from | figure
0.0441176	-16.7759	6th | scotland,
0.0441176	-16.7759	translation, | edinburgh, | 30–31,
0.0441176	-16.7759	the | 30–31,
0.037037	-16.7759	text | their
0.1	-16.7761	algorithmic | be
0.0625	-16.7764	more | making
0.0454545	-16.7765	,x,b) | l
0.0208333	-16.779	of | represented
0.0625	-16.7795	used | argument
0.0241546	-16.7801	table | best
0.1	-16.7811	2. | ,
0.0454545	-16.7811	data | treebank
0.0410959	-16.7812	5 | presented
0.00518672	-16.7816	not | of
0.00527821	-16.7818	we | that | and
0.0540541	-16.7822	which | state-of-the-art
0.0769231	-16.784	latent | distributions
0.0184211	-16.785	in | 3, | the
0.0136571	-16.7854	of | domain
0.0365854	-16.7854	we | otherwise
0.0571429	-16.7858	maximum | this
0.0588235	-16.7863	recently, | on
0.0352941	-16.7875	section | proposed
0.0666667	-16.7888	3 | having
0.0322581	-16.7889	in | emotion
0.047619	-16.789	generate | training
0.0352941	-16.7893	x | x1
0.0117271	-16.7896	well | of
0.00446497	-16.7897	in | on
0.125	-16.7906	offolds | localization%
0.0588235	-16.7923	may | them
0.0526316	-16.7925	search | algorithm
0.0194175	-16.7936	of | linguistics
0.111111	-16.7938	1n | i
0.00409836	-16.7944	have | the
0.0326087	-16.7955	information | been
0.0769231	-16.7958	necessary | such
0.0769231	-16.7958	sentence | length.
0.0588235	-16.7984	somewhat | are
0.0197368	-16.8005	is | analysis
0.00858615	-16.8007	than | and
0.0769231	-16.8017	of | kind.
0.00408701	-16.8027	is | of | in
0.125	-16.8035	little | training
0.0109127	-16.8039	by | al.,
0.0666667	-16.8042	this | challenges.
0.016	-16.8052	are | both
0.125	-16.8067	model, | examples
0.0714286	-16.8068	assign | training
0.0261438	-16.8071	(brown | al.,
0.0571429	-16.8079	word | certain
0.00805585	-16.8081	in figure | the | the
0.0714286	-16.8087	these | variations
0.00377964	-16.8091	in | the | for
0.111111	-16.8111	way | corresponding
0.0164706	-16.8113	model | can
0.00599829	-16.8134	to | not
0.0144665	-16.8136	need | of
0.0454545	-16.8138	show | evaluation
0.0264901	-16.8139	work | based
0.00357508	-16.8149	the | can
0.0357143	-16.8151	web | by
0.0115546	-16.8164	of | work
0.028777	-16.8172	of | identical
0.1	-16.8172	value | smaller
0.020979	-16.8211	in | times
0.00396267	-16.8213	table | the | the
0.0107108	-16.8234	training | is
0.0134128	-16.8244	two | be
0.0222222	-16.8251	are | 2.
0.0268456	-16.8252	2011 conference | processing,
0.0135135	-16.827	model | al.,
0.0194805	-16.8299	table | scores
0.01875	-16.8341	49th | association for
0.0217391	-16.8348	to | head
0.0379747	-16.8356	information | given
0.04	-16.839	interesting | be
0.0487805	-16.8392	results | differences
0.0352941	-16.8392	be | need
0.0769231	-16.8402	spans | all
0.0454545	-16.8403	70 | 0
0.0133136	-16.8405	comparable | the
0.00970874	-16.8428	the | related | the
0.0416667	-16.8432	cases | was
0.0219298	-16.8436	project | of
0.00578194	-16.8437	is | is | a
0.0769231	-16.8464	ranked | average
0.0010281	-16.8467	the | of | the
0.037037	-16.8475	we | yet
0.0555556	-16.8495	a | probabilities.
0.0170732	-16.8553	in | small
0.0035498	-16.8559	in | as
0.0108161	-16.8579	has | and | al.,
0.0158371	-16.86	be | only
0.1	-16.8601	semantic | labelling
0.111111	-16.8616	cea | a
0.1	-16.8629	structures | classification
0.00479294	-16.8632	on the | of
0.00309058	-16.8638	and | of | and
0.0444444	-16.8653	p5r3 | +**
0.0465116	-16.8662	small, | of
0.0135135	-16.8668	only | be
0.012931	-16.8688	et | 2010)
0.0909091	-16.8689	data, | extracted
0.0571429	-16.8691	this | again
0.0119474	-16.8696	is | will
0.0243902	-16.8699	training | set.
0.0425532	-16.8715	one | extraction
0.0714286	-16.872	analysis | disambiguation
0.0129125	-16.8721	that | results
0.00716846	-16.8727	our | to | and
0.1	-16.8739	→r | french
0.0221239	-16.8746	annual | linguistics,
0.0588235	-16.8755	figure 1 | this
0.05	-16.8766	word-level | in
0.016092	-16.8779	computational | pages
0.0833333	-16.8785	star | count(oi
0.0769231	-16.8824	if | node.
0.038961	-16.8826	with | 3.
0.0487805	-16.8834	database | with
0.0221239	-16.8839	just | to
0.0540541	-16.8845	that | summaries.
0.0909091	-16.8866	depends | whether
0.0454545	-16.8878	does | perform
0.0068328	-16.8887	for | al.,
0.1	-16.8907	detection | spoken
0.0588235	-16.892	let | :
0.111111	-16.8926	favre, | irvin
0.111111	-16.8926	once | documents
0.0588235	-16.8937	242d2e302e3a21 | 21
0.0909091	-16.8943	or | judgments
0.0555556	-16.8958	with | optimization
0.0101946	-16.8963	proceedings | pages
0.04	-16.8964	these | task.
0.0909091	-16.8969	their | understanding
0.00463392	-16.8995	the | of | each
0.0315789	-16.9015	word | associated
0.0588235	-16.902	6 | setup
0.0238095	-16.9028	work | has | al.,
0.00164943	-16.9031	for | the
0.0512821	-16.9042	p1r5 | +**
0.1	-16.9049	approach | conversion
0.017094	-16.9052	select | of
0.0487805	-16.9097	models | much
0.037037	-16.9116	the | are | 6.
0.0233918	-16.913	al., | 2009)
0.021645	-16.9145	to | summary
0.00144699	-16.9148	is | the
0.0909091	-16.9178	function | individual
0.0714286	-16.9186	concept | associated
0.0107411	-16.9211	– | + | –
0.0833333	-16.9213	do | decoding
0.00367342	-16.923	to | with
0.0168539	-16.9248	from | translation
0.1	-16.9252	evaluated | compared
0.00445553	-16.9255	and | be
0.0291262	-16.9279	data | given
0.047619	-16.9282	popularity | of
0.00293233	-16.93	the | or
0.0220264	-16.9301	be | many
0.111111	-16.931	knowledge | abstract
0.0714286	-16.9311	malay | is
0.0229885	-16.9311	mainly | of
0.0227273	-16.9317	use | one
0.0121457	-16.9319	used | that
0.0408163	-16.9359	in | linguistics.
0.0588235	-16.9362	not | correction
0.0909091	-16.9364	whose | less
0.0178042	-16.9366	alignment | for
0.1	-16.937	graph | denote
0.0714286	-16.9374	functions | following
0.0163043	-16.9379	fact, | the
0.0340909	-16.9384	phrase-based | as
0.00530839	-16.9403	a | set
0.0352941	-16.9418	expressed | of | not
0.00398974	-16.9419	the | than
0.00437675	-16.942	this | we | the
0.0512821	-16.9447	is | expectation
0.0104932	-16.9469	to | phrase
0.00274782	-16.9469	our | the
0.000976002	-16.9474	the | the | of
0.0154185	-16.9494	would like | the
0.00426929	-16.9497	in | a | to
0.0909091	-16.9499	=+ | next-speaker(x))
0.0081864	-16.9501	the | by | et
0.111111	-16.9508	precision | results.
0.030303	-16.9518	have | annotation
0.0231214	-16.952	note | our
0.0131148	-16.9531	classifier | and
0.010989	-16.9533	4 | to
0.0215517	-16.9552	abstract we | a
0.0190114	-16.9556	that | features.
0.0555556	-16.9559	in contrast, | model
0.0126582	-16.9566	association for computational | university
0.0625	-16.959	3 | report
0.0361446	-16.9598	each | time
0.0465116	-16.9602	a | subgraph
0.1	-16.9609	sometimes | other
0.0280374	-16.9609	target | set
0.047619	-16.9623	see | best
0.00501792	-16.9631	a71 | a71
0.00216844	-16.9636	of | is
0.0227273	-16.9637	in | people
0.0909091	-16.9637	from | representations.
0.0588235	-16.9645	name | method
0.0425532	-16.9654	– | results
0.0155211	-16.9657	achieve | the
0.0261438	-16.9658	with | parallel
0.0231214	-16.9674	have | 2010).
0.1	-16.97	embed | that
0.0454545	-16.9747	recall | words
0.00754717	-16.9751	the | can | as
0.00254854	-16.9806	are | the | the
0.0216216	-16.9816	we | p
0.0232558	-16.9823	conference | language | edinburgh,
0.0309278	-16.9827	when | were
0.0322581	-16.9847	calculated | for
0.0326087	-16.987	by | precision
0.0487805	-16.9871	organized | follows:
0.0714286	-16.9873	become | standard
0.0294118	-16.988	desired | the
0.0135135	-16.99	of | model,
0.047619	-16.9902	make | when
0.0769231	-16.9916	bigram | progressive-verb
0.0434783	-16.9917	a | library
0.0769231	-16.994	a | pitman-yor
0.0377358	-16.9952	participant | in
0.00673854	-16.9954	the | learning
0.0182482	-16.9955	the | approach | et
0.00391645	-16.9958	use | the
0.0666667	-16.9959	task, | bel(d,
0.027027	-16.9959	----- | ---
0.0666667	-16.998	ˆc | it
0.04	-16.998	dataset, | in
0.00243681	-17.0007	we | in
0.00913621	-17.0013	the | due
0.0169492	-17.0023	proceedings | on | language
0.00673854	-17.0037	we | the | =
0.00293157	-17.0046	of | that | the
0.0102881	-17.0074	sentiment | the
0.0227273	-17.011	1 | 2:
0.0588235	-17.014	aspects | influence
0.0275229	-17.0148	each | represented
0.05	-17.0154	have | –
0.00653595	-17.0156	3 | of
0.0769231	-17.016	used | approaches.
0.0105018	-17.019	model | et
0.0384615	-17.0234	from | probabilities
0.0588235	-17.0234	problem, | one
0.0714286	-17.0235	know | word
0.00159458	-17.0242	the | the | is
0.0588235	-17.0291	news | corpora.
0.0357143	-17.0292	analysis | most
0.0288462	-17.0297	all | data.
0.0309278	-17.0298	verify | the
0.0138614	-17.0298	depending | the
0.00797342	-17.0324	table | to
0.0235294	-17.033	languages, | and
0.0109622	-17.0333	a | of | target
0.0625	-17.0336	error | annotated
0.00820232	-17.0355	in | also
0.0236686	-17.0363	in | some | are
0.0272727	-17.0364	in | paper, | model
0.1	-17.0364	over | having
0.0326087	-17.0369	those | al.,
0.0115108	-17.0385	an | can
0.00648789	-17.0393	as | et
0.1	-17.0397	summary | overall
0.0127273	-17.044	can | not
0.00699301	-17.0457	in | between
0.00828157	-17.0475	figure | the | in
0.0285714	-17.0505	in | improves
0.00848111	-17.0509	that | each
0.0107784	-17.0545	this | language
0.0555556	-17.0562	analysis | considered
0.0185874	-17.0569	al., | 2010;
0.05	-17.0578	position | 94.4
0.0526316	-17.0579	they | domain.
0.0392157	-17.0589	shown | 2,
0.00665716	-17.0595	such | to
0.0392157	-17.0641	of | simultaneously
0.0526316	-17.065	that | minimize
0.0465116	-17.0656	0.5 | =
0.0101237	-17.0679	that | is a
0.00165049	-17.0692	that | the
0.0666667	-17.0697	decide | number
0.0280374	-17.0713	to | location
0.0526316	-17.0721	previous research | that
0.0229885	-17.074	be | input
0.0833333	-17.0758	evidence | significant
0.0392157	-17.076	compute | set
0.0666667	-17.0787	no | conditions
0.0344828	-17.0795	quality | data
0.05	-17.0798	of | anchored
0.0185185	-17.0806	improve | in
0.0338983	-17.0837	in | enable
0.0194553	-17.0838	r: | r:
0.0392157	-17.0854	80 | 60
0.00254113	-17.0877	this | of
0.0327869	-17.0895	applicable | a
0.0909091	-17.0901	shows | quite
0.0207254	-17.0924	not | does
0.0327869	-17.0963	perplexity | a
0.0833333	-17.0988	09] | al.
0.0241935	-17.1008	our | shows
0.0392157	-17.1031	more | sense
0.0135659	-17.1041	which | than
0.0218579	-17.1059	value | that
0.0769231	-17.1061	possible | analyses
0.037037	-17.1081	vs | and
0.0526316	-17.1084	each | itself
0.0111732	-17.1105	been | by
0.00956429	-17.1115	the | (1)
0.0338983	-17.1144	tree | parse
0.0245902	-17.1144	e.g., | in
0.0588235	-17.1147	has | reading
0.016129	-17.1154	structure | is
0.00812408	-17.1158	1: | the
0.024	-17.1164	word, | is
0.00449539	-17.1184	of | on | of
0.0377358	-17.1184	generative | which
0.0769231	-17.1197	optimized | set
0.00955414	-17.1203	it | to | for
0.0833333	-17.1222	subset | documents
0.00558854	-17.1222	in | we | for
0.0173611	-17.1226	49th | oregon, | 19-24,
0.0625	-17.1252	bit | than
0.0454545	-17.1284	a | hypergraph
0.047619	-17.1286	used | state-of-the-art
0.00391223	-17.1308	and | and | in
0.0350877	-17.1353	suppose | in
0.0166667	-17.1383	mapping | and
0.00575126	-17.1387	by | a | of
0.00566572	-17.1409	have | a
0.0218579	-17.1425	is | times | in
0.0225989	-17.1432	therefore | is
0.0275229	-17.1434	named | al.,
0.0588235	-17.1443	target | translation.
0.00638686	-17.1457	word | is
0.0126354	-17.1462	space | of
0.00697051	-17.1464	in | use
0.00115325	-17.1474	in | the
0.00915565	-17.1474	not | by
0.0121739	-17.1494	features | et
0.03125	-17.15	of | 2010.
0.0125673	-17.153	2010), | al., | et
0.0118846	-17.1532	et | 2007;
0.0247934	-17.1552	can | form
0.0136054	-17.1553	in | improve
0.0833333	-17.1554	two | multilingual
0.0833333	-17.1554	similarity | semantically
0.0769231	-17.1558	in | document’s
0.00673924	-17.1572	as | of | is
0.0118846	-17.1615	on | such
0.0769231	-17.1615	combine | more
0.0714286	-17.1643	were | always
0.0769231	-17.1644	focus | paper.
0.0588235	-17.1725	counting | as
0.0222222	-17.1732	this | multiple
0.0526316	-17.1733	models | lda
0.0425532	-17.1751	approach | about
0.0526316	-17.1753	pittsburgh, | 15213
0.0206186	-17.1758	approximately | the
0.0526316	-17.1773	as | tuples
0.00731707	-17.1774	a | other
0.047619	-17.1776	of | dublin
0.00731261	-17.1797	the | machine
0.0714286	-17.1803	because | provides
0.0392157	-17.1829	to | zhang
0.0104849	-17.1836	be | than
0.0526316	-17.1892	the | (schwenk,
0.00389036	-17.1899	we | and | to
0.00371327	-17.1899	and | by
0.0241935	-17.1907	with | are | be
0.0241935	-17.1916	low | that
0.0225564	-17.1964	(2010) | as
0.0217391	-17.1976	of | overlap
0.0245902	-17.1977	each | input
0.05	-17.1993	between | sets.
0.00219545	-17.1994	the | the | are
0.0254237	-17.2	by | shown
0.006241	-17.2005	that | which
0.00761246	-17.2018	and | has
0.017316	-17.2023	the | reviewers for
0.0184332	-17.2025	in | object
0.0190476	-17.2044	parser | et
0.0434783	-17.206	present | general
0.037037	-17.2064	number | label
0.0714286	-17.2076	construct | new
0.0130435	-17.2081	the | contextual
0.0162866	-17.209	of | last
0.025	-17.21	the | (och,
0.0254237	-17.2128	are | per
0.0392157	-17.213	2-way | the
0.025641	-17.2133	data, | from
0.00189211	-17.2156	are | the
0.0177778	-17.2162	of | short
0.00491756	-17.2167	be | for
0.0769231	-17.2172	along | corresponding
0.0214286	-17.2187	finally, | be
0.0357143	-17.2229	step | all
0.0357143	-17.2229	count | by
0.00780031	-17.224	table | on
0.0094451	-17.2261	the | of | scores
0.028169	-17.2267	6th workshop | statistical
0.0588235	-17.2268	walk, | the
0.0377358	-17.2272	“i | the
0.00412116	-17.2279	of | that | to
0.00644202	-17.228	we | from the
0.0833333	-17.2286	their | almost
0.047619	-17.229	set | determine
0.037037	-17.2296	training | order
0.0666667	-17.2306	score | within
0.0192308	-17.2316	the | narrative
0.0434783	-17.2333	have | focused
0.0307692	-17.2342	which | entities
0.0238095	-17.2367	between | they
0.0252101	-17.2412	which | previous
0.00249188	-17.2414	is | the | in
0.0131291	-17.2428	system | be
0.01059	-17.2435	a | algorithm | the
0.047619	-17.2442	paper, | techniques
0.0243902	-17.2454	sentences | at
0.00821018	-17.2465	to | over
0.00914634	-17.2471	of | natural
0.0174672	-17.2471	probability | be
0.0454545	-17.2476	more | unlabeled
0.0344828	-17.2482	attribute | as
0.0014856	-17.2492	the | and | of
0.0203046	-17.2502	the | formal
0.014245	-17.2505	two | can
0.0555556	-17.252	weight | term
0.0344828	-17.2524	language | probabilities
0.00329852	-17.2525	a | to | a
0.0377358	-17.2529	how | been
0.0225564	-17.2533	be | being
0.0555556	-17.2542	complexity | not
0.0408163	-17.2542	we | calculating
0.0107034	-17.2544	and | standard
0.0555556	-17.2565	penn | parsing
0.0298507	-17.2572	due to | our
0.0223881	-17.2599	this | funded
0.0114379	-17.2625	a | best
0.0142857	-17.264	pages | oregon, june
0.0192308	-17.2642	0.5 | a71
0.031746	-17.2662	for | virtual
0.0146628	-17.267	for | trained
0.011583	-17.2678	we | of | sentences
0.00740741	-17.2686	to | that | be
0.0118577	-17.2691	described | to
0.0377358	-17.2699	number | entities
0.0273973	-17.2703	search | can
0.026087	-17.2718	data, | be
0.0588235	-17.272	demand | and
0.0227273	-17.2784	to | information,
0.0137363	-17.2817	set | words
0.00898876	-17.2828	proceedings of | on
0.016129	-17.284	and | clustering
0.0230769	-17.2842	in | not.
0.0263158	-17.286	which | those
0.0247934	-17.2866	each | shown
0.0434783	-17.2866	from | becomes
0.05	-17.2877	assume | information
0.00857143	-17.2887	context | the | of
0.0175439	-17.2912	would | to thank
0.00643432	-17.2918	with | we
0.0338983	-17.2922	trees | from
0.0011305	-17.2923	the | of | of
0.0125261	-17.2924	is | likely
0.028169	-17.2947	example | such
0.00311216	-17.2961	we | to | and
0.04	-17.2962	the | significance.
0.0187793	-17.2967	and | lexicon
0.00242737	-17.2999	an | the
0.0185185	-17.3012	those | be
0.0555556	-17.3028	assignment | an
0.0454545	-17.3042	used | europarl
0.0134048	-17.3067	meeting of | for
0.0202703	-17.3085	we | obtain | of
0.0229008	-17.3094	2 | w
0.0115385	-17.3111	other | be
0.0178571	-17.3112	we | achieve
0.037037	-17.3123	features | binary
0.0416667	-17.3129	three | name
0.0344828	-17.313	german | on
0.0666667	-17.3133	can | recognized
0.00554608	-17.3143	order to | the | the
0.0105263	-17.3146	are | data
0.0377358	-17.3161	dynamic programming | of
0.0108527	-17.3168	of | association | 2011.
0.0215827	-17.3198	feature | words
0.0625	-17.3219	00) | (00)
0.0666667	-17.322	what | those
0.0714286	-17.3251	3we | for
0.025974	-17.3271	standard | such
0.05	-17.3306	applying | set
0.0266667	-17.3306	subsequent | to
0.0142857	-17.3313	we | system.
0.0588235	-17.3322	but | here
0.00621118	-17.333	number | to
0.0526316	-17.3334	performance | statistically
0.0151515	-17.3346	model, | for
0.0215827	-17.3349	for | 100
0.0225564	-17.3354	created | a | of
0.0363636	-17.3355	extract | or
0.00516034	-17.3376	word | to
0.0174672	-17.3401	this | by | grant
0.0338983	-17.3405	our | treebank
0.0416667	-17.3439	through | two
0.0625	-17.3439	systems | data,
0.0454545	-17.3439	semantic | logical
0.0214286	-17.3467	to | false
0.0434783	-17.3468	association | mutual
0.0035823	-17.347	of | an
0.047619	-17.3471	since | nodes
0.0526316	-17.3474	section | provided
0.015748	-17.3507	- | +
0.0217391	-17.3516	in | column
0.0454545	-17.354	model | reading
0.047619	-17.3556	with | transitions
0.0117647	-17.3606	word | al.,
0.0101892	-17.3618	be | into
0.0588235	-17.3637	otherwise, | are
0.0454545	-17.3643	understanding | such
0.0454545	-17.3663	distributional | contexts
0.0363636	-17.368	60 | 40
0.037037	-17.368	began | of
0.00686813	-17.3687	a | use
0.0625	-17.3692	are | items.
0.0322581	-17.3693	many | training
0.00456343	-17.3739	figure | of
0.0555556	-17.374	of | 2007);
0.0126263	-17.3741	the | times | and
0.00749376	-17.3749	different | a
0.0232558	-17.3749	phrase | it
0.00632184	-17.375	for | each
0.0588235	-17.3798	error | training (och,
0.0055814	-17.3839	is | can
0.0140449	-17.386	size | a
0.00542118	-17.3868	by | of | in
0.0666667	-17.3879	effort | language
0.014245	-17.3882	to | fact
0.0526316	-17.3905	inference | 2010).
0.0069541	-17.3931	the | 2011.
0.0142045	-17.3937	because | as
0.028169	-17.3949	the best | was
0.030303	-17.3957	are | distinct
0.0289855	-17.3972	similar | between
0.0588235	-17.399	figure | flow
0.015748	-17.3991	be | new
0.00491746	-17.3998	by | for
0.0344828	-17.4031	cn | s
0.0666667	-17.4035	is | convex,
0.0217391	-17.4036	the | essentially
0.0179641	-17.4044	we | to make
0.0588235	-17.4045	logic, | a
0.04	-17.4053	at | 50
0.0384615	-17.4092	for | ending
0.00829876	-17.4107	in | paper
0.0128866	-17.4121	annual | association
0.0138889	-17.4123	two | =
0.0384615	-17.4129	lwlm | the
0.00981767	-17.4134	are | set
0.0119904	-17.4171	able | a
0.0416667	-17.4189	since | interested
0.0141844	-17.4193	we | due
0.025	-17.4201	language | lexical
0.0150943	-17.4202	for | proposed
0.00141861	-17.421	of | to
0.0163265	-17.4241	opinions, | those | the
0.00861141	-17.4268	performance | in
0.00981767	-17.4275	the | provides
0.0186335	-17.429	is | collection
0.0588235	-17.4298	some | observed
0.0357143	-17.43	each | global
0.0384615	-17.4389	determines | for
0.00275166	-17.4409	the | that | a
0.0588235	-17.4413	on | stages
0.0182927	-17.4421	are | label
0.0232558	-17.4421	2011 | usa,
0.028169	-17.4449	verbs | which
0.00176073	-17.448	in | a
0.00212489	-17.4491	that | to
0.00771456	-17.4515	the | are | table
0.037037	-17.4524	hedges | in
0.0357143	-17.4528	with | conll
0.00513631	-17.4585	a | number
0.0103627	-17.4611	the | of | left
0.05	-17.4616	system | emotion
0.0434783	-17.4639	gains | bleu
0.00200901	-17.4641	to | the | in
0.025974	-17.4643	relation | words
0.00651042	-17.4647	a | information
0.0140845	-17.4666	is | follows:
0.0384615	-17.4694	probability | weight
0.0189873	-17.4696	verb | be
0.05	-17.4717	smallest | is
0.0357143	-17.4724	massive | of
0.0294118	-17.478	it | top
0.047619	-17.4785	means | where
0.0266667	-17.4832	analysis | such
0.00326068	-17.4877	with | a
0.00947867	-17.4915	models | et
0.0150943	-17.4915	a71a71 | a71
0.0434783	-17.4926	model | structure.
0.0232558	-17.4936	examples | table
0.04	-17.4963	candidate | pair
0.0625	-17.4967	given | document.
0.0246914	-17.4981	for | not.
0.00350946	-17.5021	of | of | that
0.0224719	-17.5041	evaluated | from
0.0158103	-17.5045	the | computational linguistics:shortpapers,
0.0235294	-17.5074	made | et
0.00370128	-17.508	the | of | can
0.00413223	-17.5097	a | each
0.0555556	-17.5101	many | while
0.0625	-17.513	percentage | pairs
0.0357143	-17.513	in | by | respectively.
0.0289855	-17.5138	model | rule
0.00340909	-17.5159	as | of | of
0.0174419	-17.5161	languages | as
0.0322581	-17.5164	the | ∏
0.0298507	-17.5184	which | .
0.025	-17.5209	combined | by
0.0416667	-17.5218	czech, | and
0.04	-17.5236	adding | training
0.0143885	-17.5248	that | al.
0.0125392	-17.5271	it | has
0.0196078	-17.5285	for | count
0.0217391	-17.5291	our | learn
0.00978793	-17.5292	baseline | in
0.0289855	-17.5323	also | because
0.0163043	-17.5323	2011 association | computational
0.0176471	-17.5324	evaluate | with
0.00918836	-17.5349	use | al.,
0.0526316	-17.5353	& | language
0.0130293	-17.5361	show | our
0.00791855	-17.5371	be | it
0.0194805	-17.5382	accuracy | data
0.0105932	-17.5427	example, | are
0.0124611	-17.5427	we | patterns
0.04	-17.5429	basic | has
0.0178571	-17.5445	acknowledgments | to
0.00952381	-17.5451	number | by
0.0357143	-17.5454	the | mind.
0.04	-17.5473	its | nodes
0.0154639	-17.5474	that | argument
0.05	-17.5477	soon | al.
0.0555556	-17.5487	0.0005 | 0.0010
0.0114155	-17.5512	be | same
0.0454545	-17.552	not | topics.
0.00878477	-17.5525	search | and
0.012945	-17.5538	2 | be
0.0102881	-17.5583	and | documents
0.00679117	-17.5586	and | al.
0.0289855	-17.5607	information | dialogue
0.0588235	-17.5638	tested | language
0.00928793	-17.5665	of | association | june
0.0273973	-17.567	is | accuracy.
0.013245	-17.57	language | pages | 2011.
0.028169	-17.5701	one | tree
0.0588235	-17.5703	number | 12
0.00578369	-17.5716	but | in
0.00221597	-17.5756	the | from | the
0.0133779	-17.5771	association | pages | june
0.0180723	-17.5772	semantic | such
0.0125	-17.5782	49th annual | computational
0.0115741	-17.5788	features | al.,
0.0132013	-17.5792	al., | 2010)
0.00445969	-17.5813	and | can
0.0107991	-17.5814	is | al.
0.0206186	-17.5821	a | keep
0.0454545	-17.5823	from | listed
0.00388299	-17.5834	the | they
0.00923077	-17.5851	c©2011 association | computational linguistics | and
0.047619	-17.5857	2 | ...
0.0294118	-17.5869	6th | translation, | uk,
0.030303	-17.5874	example, | correct
0.0104603	-17.5875	frequency | in
0.00206322	-17.5878	is | and | the
0.0182927	-17.5887	we | account
0.0253165	-17.5898	current | from
0.0240964	-17.5928	= | return
0.0196078	-17.5935	can | parsing
0.0215054	-17.5937	1), | the
0.0051474	-17.5937	the | top | the
0.0526316	-17.5937	due | because
0.00123033	-17.5939	to | of
0.00565611	-17.5945	example, | of
0.0273973	-17.5948	our | larger
0.0289855	-17.5982	similar | they
0.0357143	-17.5993	these | determine
0.0111607	-17.5997	data, | and
0.0212766	-17.6019	the | harmonic
0.05	-17.6038	to | seconds
0.0273973	-17.6042	have | form
0.0555556	-17.608	0.0015 | 0.0005
0.0344828	-17.6081	prune | a
0.0454545	-17.6109	tend | similar
0.016129	-17.6128	by | three
0.0126582	-17.6138	on | learning
0.025641	-17.6152	from | set,
0.0079096	-17.6165	on | training
0.0179641	-17.6172	the | summarized
0.0106838	-17.6187	a | general
0.00931677	-17.6191	pages | june
0.0196078	-17.6207	for | (section
0.0454545	-17.6215	we | typing
0.0031752	-17.6223	we | a | in
0.00236991	-17.6233	the | it
0.0555556	-17.6242	since | occur
0.0131579	-17.6327	pairs | from
0.05	-17.633	χ2 | to
0.0135135	-17.6338	to | perform
0.0204082	-17.6411	of | web.
0.025	-17.6465	make | our
0.0434783	-17.6501	although | al.
0.0157068	-17.6517	turn | and
0.0238095	-17.6526	were | out
0.03125	-17.6559	the | (shown
0.0208333	-17.6566	can | because
0.00581395	-17.6618	number | is
0.0111111	-17.6622	a | reordering
0.00680272	-17.6638	for | number
0.0344828	-17.6656	in | prep
0.00516529	-17.6665	to | used
0.012987	-17.6695	that | work
0.0163934	-17.6696	al., | 2009),
0.00837989	-17.6704	or | et
0.0384615	-17.6705	input | single
0.027027	-17.6714	workshop | of | june
0.0294118	-17.6739	-auto | -
0.0416667	-17.6745	methods | many
0.0294118	-17.6757	sparse | for
0.00294214	-17.6792	be | a
0.0023661	-17.6803	to | as
0.0333333	-17.6807	their | pairs
0.0196078	-17.686	workshop | july
0.0206186	-17.6926	segmentation | by
0.0176991	-17.6959	50 | 70
0.047619	-17.6968	as | naturally
0.025	-17.6974	data | better
0.00737619	-17.6974	type | of
0.0174419	-17.7039	empirical | natural | 27–31,
0.0229885	-17.7096	former | a
0.0136364	-17.7163	table 2 | and
0.04	-17.7179	experiments | single
0.0240964	-17.7185	example | al.,
0.03125	-17.7186	100 | training
0.0168539	-17.7195	been | has
0.0322581	-17.7197	a | replacement
0.0133929	-17.7246	edge | a
0.0088968	-17.7254	data | al.,
0.030303	-17.7254	this is | case,
0.0162162	-17.7258	which | when
0.0185185	-17.7262	as | individual
0.00188904	-17.727	of | as
0.0116618	-17.728	for | vector
0.0322581	-17.7281	furthermore, | used
0.0169492	-17.7292	the | annotate
0.0224719	-17.7293	of | case.
0.0434783	-17.7293	be | question.
0.025641	-17.7308	use | regression
0.0222222	-17.7318	prominent | the
0.0240964	-17.7326	language | most
0.0125786	-17.734	49th | of | june
0.00526624	-17.7355	the | work
0.0240964	-17.7358	= | edge
0.0294118	-17.7377	can | improvements
0.037037	-17.7395	strongest | the
0.00689655	-17.7425	this | al.,
0.000983815	-17.7429	to | the | the
0.00860585	-17.7431	relation | in
0.0166667	-17.7449	natural | scotland, | 2011.
0.0416667	-17.7466	unfortunately, | this
0.0294118	-17.7474	for | anaphora
0.0178571	-17.7482	can | standard
0.00834725	-17.7493	this | is | be
0.0114286	-17.7505	a | bilingual
0.00732601	-17.751	input | of
0.00475511	-17.7533	have | to | the
0.00361312	-17.7558	with | is
0.0181818	-17.7627	do | each
0.0144231	-17.7629	by | task
0.0322581	-17.7645	higher | also
0.0357143	-17.771	this | indicated
0.0322581	-17.7711	• | extracted
0.00344828	-17.7743	the | that | be
0.012605	-17.7749	to | joint
0.047619	-17.7751	three | significantly
0.0294118	-17.7772	we | czech
0.014218	-17.7821	in addition | a
0.0235294	-17.7837	no | all
0.0243902	-17.7875	can | implemented
0.0136986	-17.7885	are | even
0.027027	-17.7898	metrics | more
0.0183486	-17.7899	results | terms
0.0322581	-17.7911	aligned | two
0.0185185	-17.7935	there | difference
0.0103627	-17.7935	words, | to
0.0130435	-17.7938	find | are
0.02	-17.7957	n | g
0.0454545	-17.799	phrase-pairs | the
0.030303	-17.8054	that | class,
0.0190476	-17.8062	be | further
0.0181818	-17.8131	of | fine-grained
0.0178571	-17.8184	in | fragments
0.030303	-17.8227	then | length
0.0142857	-17.8245	in | known
0.0333333	-17.8255	with | linking
0.025	-17.8273	node | source
0.00238424	-17.8312	as | a
0.0434783	-17.8372	event | could
0.0122449	-17.838	the | is | significant
0.00422297	-17.8394	in | and | with
0.0115385	-17.8403	in | c
0.0294118	-17.8504	one | automatically
0.0011939	-17.8526	the | of | a
0.0238095	-17.853	language | 2010),
0.0163934	-17.8554	also | corpus
0.020202	-17.8596	model | so
0.0285714	-17.8621	3.3 | et
0.00150477	-17.8655	with | the
0.0238095	-17.8672	the | λadj
0.0052459	-17.8674	we | corpus
0.00319803	-17.8674	in | al.,
0.0285714	-17.8685	a | hashing
0.0285714	-17.8728	better | these
0.0238095	-17.8743	often, | the
0.00122071	-17.8761	the | the | for
0.0210526	-17.8852	can | whether
0.0277778	-17.8882	avg/max | avg/max
0.0131004	-17.8898	sentences | we
0.0322581	-17.89	identified | each
0.030303	-17.8901	a | ),
0.0238095	-17.8907	figure 1, | is
0.00455466	-17.8978	of | over
0.00444884	-17.9042	as | as | and
0.03125	-17.9045	the | unknown.
0.00705053	-17.905	c©2011 association | linguistics
0.00211104	-17.9062	the | and | as
0.0115385	-17.9065	work | from | and
0.0117647	-17.9099	in | performance.
0.0285714	-17.9102	learning | dependency
0.00681818	-17.9108	the | summationdisplay
0.00558659	-17.9129	the | 4
0.0222222	-17.915	about | while
0.03125	-17.9168	phrase-based | word
0.012987	-17.9177	experimental | for
0.00647948	-17.9184	and | learning
0.0188679	-17.9189	general | et
0.0127119	-17.9209	on | single
0.00350542	-17.9235	of | have
0.04	-17.9237	2: | various
0.00594732	-17.9254	evaluate | the
0.00810373	-17.9268	association | the
0.00854701	-17.9273	e | r
0.00472813	-17.9278	a | different
0.0222222	-17.9326	as | instances.
0.00678733	-17.9331	because | a
0.03125	-17.9341	realized | in
0.0103627	-17.9352	news | of
0.0222222	-17.9398	tweets | not
0.00133393	-17.9413	to | in
0.04	-17.943	experiments | 1.
0.0102302	-17.9439	of | able
0.0263158	-17.9444	translation | out
0.000788751	-17.947	a | the
0.030303	-17.9525	our | simpler
0.00323974	-17.9533	of | =
0.00807754	-17.9549	features | be
0.0222222	-17.956	two | under
0.0232558	-17.9609	on | subjectivity
0.0104167	-17.9638	49th | june | 2011.
0.0285714	-17.9701	denotation | the
0.0227273	-17.9706	for | candidates.
0.0344828	-17.9725	nodes | but
0.0165289	-17.9725	each | extracted
0.00191755	-17.9735	in | and | and
0.0384615	-17.9757	fully | al.,
0.018018	-17.9801	only | will
0.00530303	-17.9849	its | the | of
0.0136986	-17.991	used | also
0.012987	-17.9911	types | be
0.017094	-17.9955	understand | and
0.0294118	-17.9974	time | but
0.00274941	-18.0062	a | we
0.0333333	-18.0064	that | broad
0.00167018	-18.0069	and | to | the
0.025641	-18.0073	will | after
0.0106007	-18.0075	in | web
0.0322581	-18.0109	model | performing
0.0294118	-18.0125	patterns | were
0.037037	-18.0134	hovy | of
0.0176991	-18.0162	proceedings | the | international
0.0212766	-18.0178	described | use
0.0333333	-18.018	in | work, | semantic
0.0136054	-18.0213	and | million
0.030303	-18.0243	due | sentence
0.037037	-18.0264	with | above.
0.0106762	-18.0264	phrases |   |  
0.00932401	-18.0293	proceedings | annual | the
0.000685777	-18.0304	the | the | and
0.0322581	-18.0306	the | koppel
0.03125	-18.0315	project | under
0.03125	-18.0315	annotation | protein
0.0106007	-18.0316	standard | as
0.0065703	-18.0331	vector | of
0.0294118	-18.0384	specific | text
0.0133333	-18.0391	pmi | the
0.00268706	-18.0461	each | and
0.025641	-18.048	or | instances
0.00318471	-18.0494	for | be
0.0212766	-18.0518	it | increase
0.0227273	-18.0533	or | linguistic
0.0333333	-18.0535	linear | between
0.00359195	-18.0538	that | is | and
0.030303	-18.054	search | between
0.0277778	-18.0545	- | only
0.0024139	-18.0549	in | as | the
0.00439802	-18.0552	by | et
0.0204082	-18.0558	we used | described
0.0243902	-18.0573	data | within
0.0238095	-18.0595	data | thus
0.025	-18.0595	use | random
0.0204082	-18.0632	numerical | in
0.0208333	-18.069	i | b
0.0100671	-18.0712	the | valid
0.0344828	-18.0721	gmentation | a71
0.00202767	-18.0736	the | can | the
0.0115385	-18.0747	to | k
0.027027	-18.0768	spectrum | and
0.02	-18.0834	data | consists
0.00217391	-18.0835	a | a | to
0.0188679	-18.0838	knowledge, | on
0.0136986	-18.0848	experiments, | for
0.0222222	-18.0862	(8) | in
0.011236	-18.0895	how | that
0.0204082	-18.0915	+ | score
0.0232558	-18.0961	a | guide
0.0178571	-18.1112	be | ones
0.0153846	-18.1112	effort | of
0.0243902	-18.1119	updates | a
0.0153846	-18.1127	1 | feature
0.00386847	-18.1143	the | do
0.0071048	-18.1155	of | linguistics,
0.025641	-18.1171	of | (2006).
0.016129	-18.1215	1 | there
0.0263158	-18.1263	to | annotating
0.0263158	-18.1263	with | starting
0.015748	-18.1284	in | annotation.
0.000885471	-18.1319	is | the | the
0.00332668	-18.1341	its | the
0.03125	-18.1362	of | dots
0.0285714	-18.1378	which | relations.
0.0101351	-18.1384	future | for
0.0212766	-18.1421	all | small
0.00529568	-18.1424	2: | the
0.0217391	-18.144	pairs, | are
0.00146462	-18.1472	of | that
0.0106762	-18.1512	proceedings | uk,
0.0117647	-18.1525	2 | 4 | figure
0.0178571	-18.1531	order | some
0.00350331	-18.1535	or | to
0.00113721	-18.1582	that | the | the
0.0238095	-18.1596	then, | from
0.010274	-18.1634	the | regular
0.030303	-18.1639	they | best
0.0166667	-18.1665	m | where
0.00500417	-18.1696	for | and | from
0.00874636	-18.171	of | hypothesis
0.0104167	-18.1726	be | features
0.0263158	-18.1734	also | candidate
0.015748	-18.1771	this | to | models
0.0222222	-18.1773	syntactic | feature
0.00547046	-18.1798	the | bilingual
0.00631313	-18.1804	and | those
0.0227273	-18.1872	2011), | portland,
0.0294118	-18.1878	including | models
0.03125	-18.193	one | appropriate
0.0243902	-18.1947	supporting | for
0.0148148	-18.1956	of | increased
0.0212766	-18.196	network | from
0.000798101	-18.1968	a | the | the
0.0192308	-18.198	appear | than
0.0294118	-18.2	also | 1.
0.0196078	-18.2009	parameters | can
0.0114943	-18.2116	in | comparison
0.0107527	-18.2124	on | error
0.0208333	-18.2171	8 | number
0.0232558	-18.2197	the | nulli
0.000732499	-18.2198	the | to | the
0.00189708	-18.2221	a | is | of
0.00646204	-18.224	the | to | single
0.025	-18.2247	followed | al.,
0.0121951	-18.2308	tasks | for
0.025641	-18.2321	two | annotations
0.0163934	-18.235	important | note
0.0277778	-18.2388	approach | improvement
0.0169492	-18.2396	a | backoff
0.0208333	-18.2418	count | table
0.00958466	-18.2423	translation | model
0.00211416	-18.2487	be | in
0.0217391	-18.2529	obtained | system
0.00165047	-18.254	not | the
0.00192012	-18.2547	be | to
0.0144928	-18.2602	is | domain.
0.00983607	-18.2624	as | order
0.00294031	-18.2631	al., | al., | et
0.0222222	-18.2677	mining | on
0.0188679	-18.2686	evaluation | how
0.0166667	-18.2713	values | which
0.0263158	-18.272	me | and
0.010929	-18.2742	on | labeled
0.00490677	-18.2817	document | of
0.0119048	-18.2829	has | all
0.00864553	-18.2855	information | with
0.003432	-18.2925	as | be
0.027027	-18.2938	into | model,
0.00395034	-18.2943	to | same
0.0208333	-18.2955	lscript | the
0.0263158	-18.2959	has | web
0.00677966	-18.2966	the | latent
0.02	-18.2987	corpus | method
0.0075188	-18.3002	asked | the
0.0126582	-18.3003	method | this
0.00837989	-18.301	the | and | automatic
0.00195667	-18.3027	or | the
0.025641	-18.3069	applied | system
0.0133333	-18.3134	be | part
0.0178571	-18.3174	extracted | sentences
0.00573066	-18.3222	we | the | described
0.00583942	-18.324	mean | the
0.0058309	-18.331	we | to | set
0.025	-18.3352	2007; | 2010),
0.00887574	-18.3359	of | the | (i.e.,
0.0169492	-18.3364	one | standard
0.0163934	-18.3378	a | content.
0.00263992	-18.3399	and | not
0.0106383	-18.3427	this | shown | the
0.0100503	-18.3434	as | distribution
0.00233744	-18.3437	of | can
0.0185185	-18.3439	also | topic
0.00239913	-18.3472	et | et
0.0222222	-18.3479	obtained | were
0.00353179	-18.3489	to | =
0.00947867	-18.3499	all | over
0.00854701	-18.3516	matrix | of
0.00676819	-18.3529	where | that
0.00595238	-18.3545	on | words
0.0166667	-18.3553	by | texts
0.00921659	-18.3575	to find | to
0.0144928	-18.3582	model | representation
0.0111732	-18.3586	such | at
0.025	-18.3656	number | proposed
0.0142857	-18.3691	features are | on
0.00247036	-18.3707	of | it
0.00383632	-18.3738	be | from
0.0136986	-18.3763	an | available
0.0227273	-18.3865	where | name
0.00473485	-18.3872	used | et
0.011236	-18.3892	also | have
0.0232558	-18.3897	for | consistently
0.00700935	-18.3906	task | on
0.00689655	-18.3909	for | null
0.00777202	-18.391	table | data
0.0208333	-18.4017	can | thought
0.00819672	-18.4049	act | the
0.00261097	-18.4061	in | as | of
0.0142857	-18.407	associations | and
0.0104167	-18.4097	49th | for | linguistics,
0.0153846	-18.4112	a | purpose
0.0204082	-18.4143	path | =
0.0169492	-18.4146	not | true
0.0058651	-18.4199	of | whether
0.00413223	-18.4266	models | to
0.0243902	-18.428	as | clause
0.0107527	-18.429	s | figure
0.00062896	-18.4351	and | the
0.015873	-18.4372	this | treebank
0.00164042	-18.4384	for | a | the
0.00264085	-18.4397	table | the | and
0.016129	-18.4403	use | bilingual
0.00934579	-18.4419	system | set
0.0112994	-18.4451	framework, | the
0.0126582	-18.4535	types | more
0.00490196	-18.4543	section | we
0.0147059	-18.4563	types | all
0.012987	-18.4591	8: | for
0.00884956	-18.4598	results | it
0.0144928	-18.467	based | linguistic
0.0126582	-18.469	a | know
0.0204082	-18.4743	into | .
0.00626305	-18.4752	proceedings | the | linguistics,
0.015625	-18.4764	new | other
0.0208333	-18.4764	sigdial | dialogue,
0.00623701	-18.4811	the | add
0.0188679	-18.4823	advance | the
0.002491	-18.4866	we | to | for
0.012987	-18.4913	are | properties
0.00803213	-18.4916	of | percentage
0.00728155	-18.4923	of | 2011
0.0192308	-18.4924	is | antecedent
0.012987	-18.4931	section | results.
0.000667302	-18.4982	the | the | to
0.00921659	-18.5004	expressed | and
0.015625	-18.5004	that | approximately
0.00108522	-18.5034	of | is | the
0.0106383	-18.5039	to | object
0.00518135	-18.5049	1 | that
0.0131579	-18.5077	of | aims
0.015873	-18.5115	consisting | from
0.0116279	-18.5131	empirical | natural | edinburgh,
0.00995025	-18.5137	practical | the
0.0163934	-18.5209	is | lexicon.
0.0166667	-18.521	resulting | each
0.0172414	-18.5214	evaluate | corpus
0.00699301	-18.5286	and | sense
0.00357782	-18.5288	the | according
0.0208333	-18.5295	our | well.
0.00934579	-18.5317	we | impact
0.0175439	-18.5318	such | content
0.0222222	-18.5373	8 | or
0.0070922	-18.5382	proceedings | association for
0.0181818	-18.5405	that | policy
0.0142857	-18.5447	once | that
0.00484262	-18.5456	on | model
0.0138889	-18.55	language | bleu
0.00877193	-18.5564	conference | methods | language
0.0140845	-18.5568	adding | from
0.00334635	-18.5584	can | as | the
0.00638298	-18.5655	to | make
0.0138889	-18.5666	word | class
0.0149254	-18.5685	set | translations
0.00627615	-18.569	this | system | the
0.00904977	-18.5749	improvement | a
0.0212766	-18.5774	compare | both
0.0151515	-18.581	labeling | as
0.0181818	-18.5814	leverage | and
0.0102564	-18.5821	an | any
0.0107527	-18.5947	can | parse
0.00884956	-18.5973	in addition, we | the
0.00447427	-18.6007	we | when
0.0188679	-18.6047	use | system.
0.00199124	-18.6064	and | and | et
0.016129	-18.6082	that | strings
0.0125	-18.6089	r | l | t
0.0104167	-18.6115	table | examples
0.0169492	-18.616	source | order
0.0185185	-18.6204	bd | d7
0.016129	-18.6209	also | whether
0.0121951	-18.6258	the | initialized
0.00904977	-18.633	and | derived
0.0108696	-18.6336	single | word
0.00171956	-18.6355	the | of | et
0.0188679	-18.6408	queries | on
0.0166667	-18.6463	nn | vp
0.0103093	-18.6469	sentence | will
0.0169492	-18.6488	model, | set
0.0149254	-18.6508	with | vocabulary
0.00943396	-18.6531	in | variety
0.00766284	-18.658	first | the | second
0.00444939	-18.6594	under | the
0.0131579	-18.6612	are | weight
0.0196078	-18.6656	by | seed
0.0181818	-18.6656	by | results.
0.0113636	-18.6656	instance | by
0.00714286	-18.6656	could | by
0.00219727	-18.6743	set | of
0.0151515	-18.6758	network, | the
0.00427807	-18.6833	use | that
0.00980392	-18.6855	to | edge
0.00231348	-18.6865	the | into | the
0.0131579	-18.6878	found that | with
0.00414079	-18.688	using | et
0.008	-18.6899	we | algorithms
0.00118475	-18.6933	is | is
0.016129	-18.6956	different | well
0.00724638	-18.7026	show | by
0.00892857	-18.7036	is | follows.
0.0116279	-18.7051	2 | same
0.000782294	-18.707	the | as
0.00952381	-18.7145	1 | ●
0.0119048	-18.7186	our | tasks
0.00816327	-18.723	features, | in
0.0120482	-18.7235	by | framework
0.0163934	-18.7335	figure | (a)
0.015873	-18.734	in | tweets.
0.00787402	-18.7349	from | al.
0.0166667	-18.7406	to | 2001).
0.00160983	-18.7409	this | the | to
0.00961538	-18.7492	features | phrase
0.0153846	-18.7513	information | dependency
0.0103093	-18.7517	have | 2009).
0.00205286	-18.7531	example, | the
0.00446927	-18.7531	find | of
0.010989	-18.7537	speakers | a
0.00589391	-18.755	the | usually
0.00099283	-18.7575	on | the
0.00158753	-18.7636	to | to | and
0.0121951	-18.766	according | each
0.0166667	-18.7834	in | derive
0.00273848	-18.7894	is | number
0.0104167	-18.7952	an | every
0.0136986	-18.7999	nodes | and | as
0.0105263	-18.8027	scientific | and
0.0056391	-18.8056	corpus | et
0.0142857	-18.8088	used | parallel
0.0126582	-18.8107	the | incorporates
0.00398406	-18.8143	similar | of
0.00970874	-18.823	fragments | to
0.00909091	-18.8371	is | interesting
0.00649351	-18.8415	objective | of
0.00649351	-18.8415	in | work, | to
0.0111111	-18.8435	bayes | a
0.011236	-18.8435	the | far.
0.0028393	-18.8468	we | with | to
0.0147059	-18.8568	if | form
0.00560224	-18.8587	we first | and
0.0142857	-18.8595	shown | 3
0.00826446	-18.8613	contrast | a
0.00722022	-18.8618	fact | is
0.00511945	-18.8652	from | words
0.00687285	-18.8695	the | resolution
0.011236	-18.8707	first | but
0.00909091	-18.8719	that | paraphrase
0.00524476	-18.8727	in | described
0.0123457	-18.8729	opinions, | those
0.00909091	-18.8756	vp | and
0.00352734	-18.8784	in | paper,
0.0135135	-18.881	each | d
0.00826446	-18.885	this | each | with
0.00980392	-18.89	this | differences
0.00943396	-18.8923	structure | which
0.0151515	-18.8945	of | focusing
0.0121951	-18.9057	there | three
0.00326531	-18.908	order to | of
0.011236	-18.925	and | input.
0.00408163	-18.927	the | computational | oregon,
0.00738007	-18.9287	function | as
0.00727273	-18.9353	texts | a
0.00265534	-18.9356	we | al.,
0.011236	-18.9421	can | expressed
0.00877193	-18.9436	way, | a
0.0060241	-18.948	this | algorithm
0.00724638	-18.9486	1 | this
0.0133333	-18.9489	showed | on
0.00588235	-18.9507	by | these
0.0117647	-18.9573	hypotheses | by
0.00970874	-18.9609	have | applied
0.0126582	-18.9746	generative | with
0.0121951	-18.9752	which | parallel
0.00380228	-18.976	of | in table
0.00045106	-18.9849	to | the
0.0111111	-18.9859	our | vector
0.00806452	-18.9861	in | —
0.00813008	-18.9892	be | sequence
0.00439239	-18.9927	the | the final
0.00537634	-18.9948	science | of
0.0120482	-18.9985	none | a
0.0128205	-19	– | word
0.00397351	-19.0036	to | that are
0.00763359	-19.0083	for | refer
0.000910913	-19.0094	is | the | and
0.0138889	-19.0105	to | state.
0.00302801	-19.0108	the | in | where
0.00666667	-19.0112	the | messages
0.00558659	-19.0154	this paper, | in
0.010101	-19.017	using | about
0.00316706	-19.0225	the | either
0.0060241	-19.0226	this | that | more
0.00302115	-19.0253	be | on
0.00322321	-19.0371	a | was | the
0.010101	-19.0437	that | variety
0.00591716	-19.0448	we | same | as
0.00763359	-19.0522	40 | figure
0.00680272	-19.0559	used | from | in
0.00892857	-19.0622	newly | the
0.00840336	-19.069	chosen | a
0.00724638	-19.0748	an | human
0.00641026	-19.0808	level | to
0.00560224	-19.0852	in | respectively.
0.00493827	-19.0873	to | representation
0.0070922	-19.0909	of | split
0.00451467	-19.0928	method | as
0.00970874	-19.0981	using | model,
0.00763359	-19.0982	each | original
0.00980392	-19.101	number | length
0.00934579	-19.1043	the | positives
0.00182482	-19.1102	a | et
0.00345224	-19.1167	of | function
0.00405954	-19.13	of | where | to
0.005	-19.1316	apply | and
0.00694444	-19.1319	written | to
0.00607903	-19.1371	words in | to
0.00657895	-19.1427	the | challenges
0.0120482	-19.1429	or | single
0.0113636	-19.1447	and | leads
0.00793651	-19.1452	feature | based
0.00597015	-19.1487	and | does not
0.00689655	-19.1593	messages | of
0.00729927	-19.1708	by | gold
0.00434783	-19.1741	in | being
0.00877193	-19.1848	adaptation | in
0.0108696	-19.1862	training | both
0.0046729	-19.1901	more | et
0.00775194	-19.2035	feature | if
0.00689655	-19.2175	of | leads
0.00584795	-19.219	e | p
0.00314136	-19.2193	system | is
0.0078125	-19.2204	ranking | with
0.00207727	-19.2256	the | model | in
0.00645161	-19.226	via | to
0.00350058	-19.2298	translation | is
0.00729927	-19.2354	in | share
0.00090854	-19.2448	for | a
0.00662252	-19.2641	in | explore | of
0.00408163	-19.2666	of | news
0.00724638	-19.2688	| | and
0.00653595	-19.2692	to | predicate
0.00877193	-19.272	is | includes
0.00617284	-19.2778	opinions, | and
0.00970874	-19.281	discourse | on
0.00909091	-19.283	30 | 60
0.00372439	-19.2887	the | annual
0.00425532	-19.2956	with | sentence
0.00505051	-19.3064	that | without
0.00833333	-19.3127	be | languages
0.00375235	-19.3238	association | for
0.00595238	-19.3314	is | weights
0.00358423	-19.3328	and | set.
0.0060241	-19.3528	from | parse
0.00472813	-19.3627	specific | a
0.00684932	-19.3634	which | first
0.00348432	-19.3729	is | possible | the
0.00584795	-19.3785	the | modern
0.00869565	-19.3822	for | includes
0.00819672	-19.3837	with | improve
0.00694444	-19.3893	sentence | target
0.0075188	-19.3893	as | process
0.00242522	-19.3975	an | be
0.00207684	-19.398	those | the
0.000464296	-19.405	and | the | the
0.00847458	-19.411	was | al.
0.00374532	-19.4152	in | section,
0.00346021	-19.4206	of | pages | 19-24,
0.00456621	-19.4233	any | of | and
0.00469484	-19.4268	as | additional
0.0078125	-19.4317	right | in
0.00787402	-19.4437	learning | features
0.00564972	-19.4494	a | of | arguments
0.00735294	-19.4539	mainly | to
0.0037037	-19.4628	of | translations
0.0043956	-19.4662	suggests | the
0.00402414	-19.467	the | of | tag
0.00292398	-19.4689	or | are
0.00184502	-19.4859	lexical | the
0.00232019	-19.4865	this | we | for
0.0021751	-19.4972	as | al.,
0.00746269	-19.4973	tagging | in
0.00680272	-19.5103	for | larger
0.00310559	-19.5227	corresponds | the
0.00704225	-19.5271	central | the
0.00429185	-19.5402	the | suggest
0.00591716	-19.5423	chosen | and
0.00406504	-19.5775	was | each
0.00564972	-19.5791	pair | be
0.00537634	-19.5873	al., 2003; | and
0.00529101	-19.5899	other hand, | and
0.0022338	-19.6039	is | have
0.0046729	-19.6068	is | right
0.0049505	-19.6194	only | model
0.000902425	-19.6306	the | a | for
0.00628931	-19.6402	) | ,
0.00218023	-19.6494	the | associated | the
0.00291971	-19.6538	in | paper, | a
0.00188442	-19.6668	same | and
0.00537634	-19.6817	nlp | et
0.00537634	-19.6867	this | domain
0.000363164	-19.6893	of | the | of
0.0070922	-19.7024	of | predicting
0.00284091	-19.7117	in | 4. | the
0.00657895	-19.7161	experiments | section | the
0.00657895	-19.7193	through | as
0.00540541	-19.722	frame | of
0.00286944	-19.7248	the | amount
0.00561798	-19.7257	to | confidence
0.000643542	-19.7356	the | which
0.0019698	-19.748	and | where
0.00595238	-19.7613	even | it
0.00636943	-19.7738	figure | performance
0.00371747	-19.7925	million | the
0.00341297	-19.8008	and | annotations
0.00469484	-19.8026	phrase-based | a
0.0014556	-19.8036	in this | we
0.00416667	-19.809	present | as
0.000813575	-19.8209	the | to | for
0.00458716	-19.8209	large | be
0.00403226	-19.8245	of | aspects
0.00132319	-19.8359	in this | we | the
0.00148754	-19.8387	in | of | on
0.00396825	-19.8405	and | articles
0.000904841	-19.8411	a | be
0.00383142	-19.8713	that | error
0.00164204	-19.8716	language | to
0.005	-19.8752	learning | be
0.00390625	-19.8779	number | features
0.00192926	-19.8804	the | average
0.00413223	-19.8919	is | model,
0.00507614	-19.9029	this work | supported
0.00331126	-19.9121	conference | natural
0.00420168	-19.9407	as | similar
0.00274725	-19.9412	three | to
0.00490196	-19.9436	in | temporal
0.003003	-19.9502	a | good
0.00193424	-19.952	if | for
0.00177515	-19.9724	in figure | the
0.00478469	-19.9796	might | for
0.00326797	-19.9805	on | portland,
0.00342466	-19.9815	2 | on | and
0.003663	-20.0055	in | appear
0.00280112	-20.009	table | which
0.00245399	-20.0106	c©2011 association | computational linguistics
0.00374532	-20.0236	discourse | a
0.00220507	-20.0377	random | the
0.00341297	-20.0486	intended | the
0.00330033	-20.053	of | count
0.000858369	-20.069	is | a | and
0.00259067	-20.0711	it | if
0.00431034	-20.0781	a | future
0.00458716	-20.0989	than | but
0.00401606	-20.1198	language | target
0.00347222	-20.1215	proceedings | for | 19-24,
0.002331	-20.1287	a | examples
0.000967352	-20.1308	in | we | in
0.00429185	-20.1339	news | in
0.00353357	-20.1392	us | of | and
0.00337838	-20.1489	i | that
0.000816727	-20.1567	on | to
0.003367	-20.1571	the | language | 2011.
0.003367	-20.1647	of | metric
0.00242718	-20.1685	the | annual | pages
0.00174368	-20.1999	dependency | of
0.000740521	-20.2022	of | et
0.00362319	-20.2142	experimental | a
0.003003	-20.2229	three | as
0.00185874	-20.2516	has | al.,
0.0026738	-20.2527	the | is | sentence.
0.00141443	-20.258	learning | of
0.00214133	-20.2902	= | set
0.00367647	-20.298	user | for
0.0021645	-20.3033	and | being
0.000793808	-20.3412	which | and
0.0020284	-20.3426	the | annual meeting
0.00115518	-20.3457	as | and | in
0.00193424	-20.351	the | vs.
0.00101078	-20.366	we | of | with
0.00252525	-20.375	is | corpus.
0.00165017	-20.4138	corresponding | the
0.00215517	-20.4319	results | be
0.0017094	-20.4388	from | as | in
0.00167084	-20.4549	a71 | a71 | xt
0.0018315	-20.4556	in | show
0.00194553	-20.4711	morphological | the
0.000578302	-20.4827	be | of
0.000960615	-20.4976	in | can
0.00308642	-20.5048	a | thus
0.00157853	-20.5118	● | ●
0.00206186	-20.5256	the | corresponds
0.00099272	-20.5352	the | is | or
0.00243309	-20.5358	be | they
0.00117578	-20.5439	on | al.,
0.00103734	-20.5646	a | have
0.00077821	-20.5754	work | the
0.000749251	-20.5942	has | the | the
0.00171233	-20.6725	on | have
0.00244499	-20.6935	model | our
0.000365726	-20.7067	the | the | with
0.00126582	-20.7408	for | and | =
0.00139082	-20.7437	of | languages
0.00128535	-20.7898	is | there
0.000633045	-20.7995	the | of | number
0.00195695	-20.8077	a | amount
0.00083647	-20.8163	and | between
0.00124688	-20.8166	as | than
0.00177936	-20.8357	based | et
0.000913659	-20.8435	we | or
0.00108401	-20.8516	a | of | al.,
0.00158983	-20.8518	number | with
0.00138504	-20.9038	our | than
0.00102722	-20.9347	we | word | the
0.00102564	-20.9546	of | large
0.00127389	-20.9832	we | the | probability
0.0017301	-20.9999	high | and
0.00186916	-21.0129	model | on | and
0.00105042	-21.0244	among | the
0.000825423	-21.0305	to | data
0.000550358	-21.0397	to | the | with
0.000977517	-21.0963	the | of | entity
0.000523834	-21.1313	a | are
0.000554785	-21.1586	the | 1
0.00106157	-21.1655	proceedings | the | computational
0.000413052	-21.23	the | number
0.000397298	-21.2355	of | which
0.000124496	-21.2569	of | the
0.000724638	-21.2688	of | sentence
0.000634719	-21.2947	this | be
0.000904159	-21.3047	in | as a
0.000670916	-21.3411	an | is | the
0.000884956	-21.3411	compute | the
0.0010917	-21.4281	context | and
0.000688705	-21.5376	classification | the
0.000484966	-21.6511	of | this
0.000325521	-21.6807	the | of | be
0.000158791	-21.7913	a | of | the
0.000467508	-21.7994	the | does
0.000591017	-21.8043	n | the
0.000305297	-21.8076	that | for
0.000769231	-21.8134	to | when
0.000508388	-21.8172	a | data
0.000353982	-21.8272	for | of | in
0.000398883	-21.8889	new | the
0.000281294	-22.0002	of | of | for
0.000551268	-22.1347	standard | the
0.000478011	-22.1898	the | and | used
0.000463392	-22.3054	in | of | from
0.000450248	-22.4233	when | of
0.000170576	-22.5352	of | the | for
0.000392927	-22.5595	are | to | to
0.000318066	-22.6358	between | of
0.00030349	-22.7228	we | be
0.000257003	-22.9808	on | and | and
7.99488e-05	-23.2916	as | the
9.31706e-05	-23.4656	of | by
3.01814e-05	-25.1487	a | the | of