I'm now a faculty at the CS department, the Hebrew university, Jerusalem, Israel.
My new webpage is http://www.cs.huji.ac.il/~shais.
This page is no longer maintained.







Shai Shalev-Shwartz




I am a research assistant professor at the Toyota Technological Institute at Chicago.

A picture of mine



Tutorial on Online Learning (ICML 2006)

Publications Source-code People Contact


Publications

Dissertations

"Online Learning: Theory, Algorithms, and Applications" Shai Shalev-Shwartz, The Hebrew University of Jerusalem. PH.d. thesis. July 2007. [Paper (corrected): pdf ]
(I'd like to thank Francesco Orabona for pointing out important corrections to Figures 5.2 and 5.4)
"Robust Temporal and Spectral Modeling for Query by Melody" Shai Shalev-Shwartz, The Hebrew University of Jerusalem. M.Sc. thesis. Jerusalem 2002 [Paper: pdf ]

Journal Papers

"Individual Sequence Prediction using Memory-efficient Context Trees" Ofer Dekel, Shai Shalev-Shwartz and Yoram Singer, To appear in IEEE'IT. [Paper: pdf ]
"Ranking Categorical Features Using Generalization Properties" Sivan Sabato and Shai Shalev-Shwartz, Journal of Machine Learning Research, 2008. [Paper: pdf ]
"Online Learning of Complex Prediction Problems Using Simultaneous Projections" Yonatan Amit, Shai Shalev-Shwartz and Yoram Siner, Journal of Machine Learning Research, 2008. [Paper: pdf ]
"The Forgetron: A Kernel-Based Perceptron on a Budget" Ofer Dekel, Shai Shalev-Shwartz and Yoram Singer, SIAM Journal of COMPUTING, Vol. 37, Issue 5, Pages 1342-1372, 2007. [Paper: pdf ]
"A Primal-Dual Perspective of Online Learning Algorithms" Shai Shalev-Shwartz and Yoram Singer, Machine Learning Journal, 69:2/3, pages 115 - 142, 2007. [Paper: pdf ]
"A Large Margin Algorithm for Speech-to-Phoneme and Music-to-Score Alignment" Joseph Keshet, Shai Shalev-Shwartz, Yoram Singer and Dan Chazan. IEEE Trans. on Audio, Speech and Language Processing. [Paper: pdf ]
"Efficient Learning of Label Ranking by Soft Projections onto Polyhedra" Shai Shalev-Shwartz and Yoram Singer, Journal of Machine Learning Research 7 (July), pages 1567-1599, 2006. [Paper: pdf ]
"Online Passive-Aggressive Algorithms" Koby Crammer, Ofer Dekel, Joseph Keshet, Shai Shalev-Shwartz and Yoram Singer, Journal of Machine Learning Research 7, pages 551-585, 2006. [Paper: pdf ]
"Smooth Epsilon-Insensitive Regression by Loss Symmetrization" Ofer Dekel, Shai Shalev-Shwartz and Yoram Singer, Journal of Machine Learning Research (JMLR), 6(May):711--741, 2005 [Paper: pdf ]

Conference Papers

"Stochastic Methods for $\ell_1$ Regularized Loss Minimization" Shai Shalev-Shwartz and Ambuj Tewari. ICML, 2009. [Paper pdf]
"Stochastic Convex Optimization" Shai Shalev-Shwartz, Ohad Shamir, Karthik Sridharan and Nati Srebro COLT, 2009. [Paper pdf]
"Learnability and Stability in the General Learning Setting" Shai Shalev-Shwartz, Ohad Shamir, Karthik Sridharan and Nati Srebro COLT, 2009. [Paper pdf]
"Agnostic Online Learning" Shai Ben-David, David Pal and Shai Shalev-Shwartz COLT, 2009. [Paper pdf]
"Mind the duality gap: Logarithmic regret algorithms for online optimization" Sham Kakade and Shai Shalev-Shwartz. NIPS, 2008. [Paper pdf]
"Fast Rates for Regularized Objectives" Karthik Sridharan, Shai Shalev-Shwartz, Nathan Srebro. NIPS, 2008. [Paper pdf]
"SVM Optimization: Inverse Dependence on Training Set Size" Shai Shalev-Shwartz and Nathan Srebro. ICML 2008. Received best paper award [Paper (corrected): pdf ],[Errata]
"On the Equivalence of Weak Learnability and Linear Separability: New Relaxations and Efficient Boosting Algorithms" Shai Shalev-Shwartz and Yoram Singer. COLT 2008. [Paper: pdf ][Talk slides]
"Efficient Bandit Algorithms for Online Multiclass Prediction" Ambuj Tewari, Shai Shalev-Shwartz and Sham Kakade. ICML 2008. [Paper: pdf ][Talk slides]
"Efficient Projections onto the $\ell_1$-Ball for Learning in High Dimensions" John Duchi, Shai Shalev-Shwartz, Yoram Singer, and Tushar Chandra. ICML 2008. [Paper: pdf ]
"Pegasos: Primal Estimated sub-GrAdient SOlver for SVM" Shai Shalev-Shwartz, Yoram Singer, and Nathan Srebro. ICML 2007. [Paper: pdf ] [Talk Slides: ppt ] A source code is available here.
A technical report with a generalized logarithmic regret and detailed proofs:
"Logarithmic Regret Algorithms for Strongly Convex Repeated Games" ,Technical Report [2007-42], The Hebrew University, May 2007. [Paper: pdf]
"Prediction by Categorical Features: Generalization Properties and Application to Feature Ranking" Sivan Sabato and Shai Shalev-Shwartz, COLT 2007. [Paper: pdf ]
A version with all the proofs: [Paper: pdf]
"A Unified Algorithmic Approach for Efficient Online Label Ranking" Shai Shalev-Shwartz and Yoram Singer, AISTAT 2007. [Paper: pdf ]
"Convex Repeated Games and Fenchel Duality" Shai Shalev-Shwartz and Yoram Singer, NIPS 2006. [Paper: pdf ]
A version with all the proofs: [Paper: pdf]
"Online Classification for Complex Problems Using Simultaneous Projections" Yonatan Amit, Shai Shalev-Shwartz, and Yoram Singer, NIPS 2006. [Paper: pdf ]
"Online Learning meets Optimization in the Dual" Shai Shalev-Shwartz and Yoram Singer, COLT 2006. [Paper: pdf ]
A version with all the proofs: Technical Report[2006-2], Leibniz Center, 2006. [Paper: pdf]
"Online Multiclass Learning by Interclass Hypothesis Sharing" Michael Fink, Shai Shalev-Shwartz, Yoram Singer and Shimon Ullman ICML 2006. [Paper: pdf ]
"The Forgetron: A Kernel-Based Perceptron on a Fixed Budget." Ofer Dekel, Shai Shalev-Shwartz and Yoram Singer, Advances in Neural Information Processing Systems 17, MIT Press, 2005. Received "Outstanding student paper award". [Paper: pdf ] [A journal version with proofs: pdf ] [Talk Slides: ppt ] [Poster: ppt ]
"Phoneme Alignment Based on Discriminative Learning" Joseph Keshet, Shai Shalev-Shwartz, Yoram Singer, and Dan Chazan. Interspeech 2005 [Paper: pdf ]
"A New Perspective on an Old Perceptron Algorithm" Shai Shalev-Shwartz and Yoram Singer, Proceedings of the Sixteenth Annual Conference on Computational Learning Theory, 2005 [Paper: pdf ] [Errata (thanks to Francesco Orabona for pointing out a mistake in the paper)]
"The Power of Selective Memory: Self-Bounded Learning of Prediction Suffix Trees" Ofer Dekel, Shai Shalev-Shwartz and Yoram Singer, Advances in Neural Information Processing Systems 17, MIT Press, 2004. [Paper: pdf ] [Talk Slides: ppt ]
"Learning to Align Polyphonic Music" Shai Shalev-Shwartz, Joseph Keshet and Yoram Singer ISMIR 2004 Webpage for the paper [Paper: pdf ] [Long version: pdf ] [Talk Slides: ppt ]
"Online and Batch Learning of Pseudo-Metrics" Shai Shalev-Shwartz, Yoram Singer and Andrew Y. Ng ICML 2004 [Paper: pdf ] [Talk Slides: ppt ]
"Online Passive-Aggressive Algorithms" Koby Crammer, Ofer Dekel, Shai Shalev-Shwartz and Yoram Singer, Advances in Neural Information Processing Systems 16, MIT Press, 2003. [Paper: pdf ] [Talk Slides: ppt ]
"Smooth Epsilon-Insensitive Regression by Loss Symmetrization" Ofer Dekel, Shai Shalev-Shwartz and Yoram Singer, Proceedings of the Sixteenth Annual Conference on Computational Learning Theory, pages 433-447,Springer LNAI 2777, 2003 [Paper: pdf ] [Talk slides: ppt ]
"Robust Temporal and Spectral Modeling for Query by Melody" Shai Shalev-Shwartz, Shlomo Dubnov, Nir Friedman and Yoram Singer, Proceedings of the 25rd Conference on Research and Development in Information Retrieval (SIGIR), 2002. [Paper: pdf ] [Talk slides: ppt mp3 files (tar) ]

Other publications (Tech Reports, Workshops, Demonstrations, ...)

"On the duality of strong convexity and strong smoothness: Learning applications and matrix regularization" Sham Kakade, Shai Shalev-Shwartz, Ambuj Tewari. [Report, Slides of a related talk]
"Trading Accuracy for Sparsity" Shai Shalev-Shwartz, Nathan Srebro, Tong Zhang. Technical Report TTIC-TR-2009-3, May 2009. [pdf]
"Agnostic Online Learnability" Shai Shalev-Shwartz. Technical Report TTIC-TR-2008-2, October 2008. [Report] A much improved version is going [to appear in COLT], together with Shai Ben-David and David Pal.
"Stochastic Convex Optimization" Shai Shalev-Shwartz, Ohad Shamir, Nathan Srebro, Karthik Sridharan. [To appear in COLT]
"Low \ell_1 Norm and Guarantees on Sparsifiability" Shai Shalev-Shwartz and Nathan Srebro. Sparse Optimization and Variable Selection, Workshop, ICML/COLT/UAI, July, 2008. [Extended abstract, Report] [Talk slides]
"Iterative Loss Minimization with $\ell_1$-Norm Constraint and Guarantees on Sparsity" Shai Shalev-Shwartz and Nathan Srebro. Technical Report, TTI, 2008. [Report]
"A Demonstration of a Query by Melody system" Shai Shalev-Shwartz and Yoram Singer, Presented in NIPS, 2003. [Movie file: mp4 ]


People

I have worked and written papers with: Yoram Singer, Shlomo Dubnov, Nir Friedman, Ofer Dekel, Koby Crammer, Adiel Ben-Shalom, Michael Werman, Andrew Ng, Joseph Keshet, Michael Fink, Dan Chazan, Shimon Ullman, Yonatan Amit, Nati Srebro, Sivan Sabato, Sham Kakade, Ambuj Tewari, Karthik Sridharan, Ohad Shamir.

Contact Info

Shai Shalev-Shwartz
TTI-C
E. 60th Street
Chicago IL, 60637

Phone: 773 834 6850
e-mail: My-first-name at tti-c dot org