Percy Liang Office: Soda 529 [  Resume  ] [  Publications  ] [  Code  ] [  Housing  ]

News

• I'm co-organizing a NIPS workshop on natural language processing.

Research
Hello! I am currently a fourth-year computer science Ph.D. student at UC Berkeley. My advisors are Michael Jordan and Dan Klein, putting me in both the statistical AI lab (SAIL) group and the Berkeley Natural Language Processing (NLP) group. I am interested in machine learning and NLP, particularly unsupervised learning and learning theory. More specifically, I haved worked on nonparametric Bayesian methods, asymptotic statistical analyses, parsing, word alignment, machine translation, and historical linguistics. For more details, look at my publications. I have also taken some classes.

I graduated in June 2005 with my masters degree (MEng) from MIT, where I worked on statistical NLP with my advisor Michael Collins in CSAIL. I investigated ways of using unlabeled data to improve performance on named-entity recognition and Chinese word segmentation.

As an undergraduate at MIT, I worked on hypertree algorithms for finding maximum likelihood bounded tree-width Markov random fields with Nati Srebro and David Karger.

Teaching
Fall 2008: Artificial Intelligence (CS188)
Spring 2008: Practical Machine Learning (CS294-34)
Fall 2007: Statistical Learning Theory (CS281A)

Miscellaneous
Looking for housing? Try the Craigslist-Google Maps mashup that I created.
Tired of Powerpoint? Try rfig, a package I wrote for creating figures and slide presentations in Ruby.

Links
MIT ACM Programming Contest, which I used to organize
USA Computing Olympiad, a programming contest for high school students that I coach
Wikipedia, a free encyclopedia