Our group at the University of Waterloo studies computational linguistics and natural language processing. We are generally interested in the following topics:

• Language grounding: learning to connect language with the world.
• Generalization: building models that can generalize across languages and domains.
• General language science: using computational methods to find and explain linguistic phenomena.

Advising Philosophy. PhD programs are designed for in-depth training and education of independent researchers, and thesis-based Master's programs intend to prepare students for further academic studies. By the date of graduation, I expect PhD students to be able to conduct independent research and Master's students to be able to conduct research with some reasonable guidance.

Group Structure. I look forward to taking students with sufficient backgrounds in one or multiple of the following disciplines (which can be usually demonstrated by completing a Bachelor's degree with decent performance in relevant disciplines):

• Computer science;
• Cognitive science;
• Linguistics;
• Social sciences.

In the near future, I expect the group to reach a stationary status consisting of 4-8 graduate students and 0-2 postdocs, who have diverse backgrounds that cover multiple (and ideally all) disciplines listed above.

Research Style. Research is a process of both learning and discovery, and paper is the primary way to communicate the produced insights rigorously. I am personally most interested in new discoveries about language and language processing systems and their theoretical explanation; however, I value all types of scientific contributions---I'm open to other topics that students would like to investigate, as long as they can convince me that these topics are sufficiently interesting and there is enough scientific contribution. In terms of theme, I'm happy to work on anything that generates insights about human language, including but not limited to (1) how to build models that assist humans in language processing and (2) how to model the language processing system of humans computationally.

Advising Style.

• Getting Started: I expect to work with every student to develop a short-term research plan and help execute it, i.e., help students find a particular topic of interest if they still need to get one. For students who already have good enough research experience before entering the program, I'm happy to follow their own research agenda and keep them on the right track.
• Meetings: I have one-on-one meetings with all students for at least one hour per week, and the schedule is open to change. These meetings can cover project planning, research details, debugging together, administrative issues, career plans or general chatting. We will have (1.5-)weekly, 1.5-hour group meetings to read papers together, present work, and discuss research ideas.
• Coding: Students are expected to have basic coding skills and code individually most of the time; however, if a student is not from a computer science background, is entering a new research area, or simply needs some input, I'm happy to help with the details.
• Work Time: Everyone is, of course, expected to work ~40 hours per week. Students may have their own agenda, as long as they show up at most of our group meetings (some time between 10am--5pm Monday to Friday by voting for each term) and keep working on research. For general advice, I find this document on how to become a happier PhD student (by my Ph.D. advisor Dr. Kevin Gimpel) particularly helpful.
  Maintaining work-life balance is particularly valued and encouraged. Note to my present and future students: I will probably send and/or reply to emails outside of working hours. In case that happens, do not feel obligated to reply immediately unless you feel it is urgent to yourself (e.g., requesting my last-minute signature for some important milestone documents).

While I am still learning how to become a good advisor, I look forward to working to build a healthy and productive research lab with my students!