Percy Liang : 2013 Plenary Session

Developing intelligent natural language interfaces is becoming an increasingly important problem given the rise of mobile devices and structured data on the web. At the core is semantic parsing, the task of mapping natural language utterances to executable programs (e.g., database queries, API calls). Tackling this problem in the open-domain setting involves negotiating the tension between the desire to learn the semantic parser from tons of shallow supervision (language + desired behavior) and the necessity to represent the deep semantics of utterances. We present a latent-variable model and promising experiments on question answering given a large knowledge base.

Percy Liang is an Assistant Professor of Computer Science at Stanford University (B.S. from MIT, 2004; Ph.D. from UC Berkeley, 2011). His research focuses on methods for learning richly-structured statistical models from limited supervision, most recently in the context of semantic parsing in natural language processing. He won a best student paper at the International Conference on Machine Learning in 2008, received the NSF, GAANN, and NDSEG fellowships, and is also a 2010 Siebel Scholar.