Guest Speaker: Luke Zettlemoyer
Scaling Semantic Parsers to Large and Varied Domains
Semantic parsers map natural language sentences to formal representations of their meaning. Recently, algorithms have been developed to learn such parsers for many applications, including question answering, information extraction, and robot control. However, each application requires innovations in modeling and learning and we lack semantic parsers that generalize across domains.
In this talk, I present a new approach to learning domain independent semantic parsers from indirect supervision, with application to question answering and information extraction for a wide variety of topics, ranging from astronomy to celebrity trivia. The key challenge is to develop new parsing and learning techniques that require very little in-domain data. I will describe a novel two-stage parser that separates (1) a domain-independent, linguistically motivated, parse step from (2) a domain-dependent ontology matching process. This allows a single grammar to be learnt for multiple different domains while also supporting reasoning about meaning in very large ontologies. We show that the two stage parsing approach significantly outperforms competing systems when answering questions with Freebase. We also show that it can be adapted to applications in information extraction, where the key challenge is to automatically determine which subparts of a text actually describe concepts in the target ontology. Finally, I will also give a brief overview of our ongoing work towards even wider scale natural language interfaces and information extraction systems.
Luke Zettlemoyer is an Associate Professor in the Computer Science & Engineering department at the University of Washington. He does research in natural language processing, with a focus on empirical computational semantics. Honors include paper awards at UAI, ACL, and EMNLP, selection to the DARPA CSSG, an NSF CAREER Award, and an Allen Distinguished Investigator Award. Luke received his PhD from MIT and was a postdoc at the University of Edinburgh.