About

Percy's research spans both machine learning and natural language processing. His goal is to bridge theory and practice by discovering new insights on machine learning and developing new methods with practical impact. In particular, he developed a unified theoretical analysis of commonly-used machine learning methods (which won a best student paper award) and also improved the state-of-the-art on various problems in natural language processing (e.g., word alignment for machine translation). Percy has also taken initiatives in strengthening the research community. For example, he was involved in running a new practical machine learning class, starting a weekly machine learning tea to increase collaboration and discussion between research groups, and organizing various group social events. He is also working on a website which allows for an objective and collaborative way of comparing different machine learning methods.

Education

UC Berkeley

Ph.D. , Computer Science

Class of 2010

Awards & Publications