2012 Poster Sessions : Event Extraction as Dependency Parsing

Student Name : David McClosky
Advisor : Chris Manning
Research Areas: Artificial Intelligence
Information extraction is the task of distilling raw text into machine readable propositions. We focus on nested event extraction, a complex form of information extraction where the propositions are events consisting of a predicate and its labeled arguments. Nesting allows the labeled arguments to be other events, indicating causality. However, most current approaches address event extraction with highly local models that extract each predicate and argument independently. We propose a simple approach for the extraction of such structures by taking the tree of event-argument relations and using them directly as the representation in a reranking dependency parser. This provides a simple framework that captures global properties of both nested and flat event structures. Our approach obtains competitive results in the extraction of biomedical events in the BioNLP'09 and BioNLP'11 shared tasks.

David McClosky is a postdoctoral researcher in the Natural Language Processing Group in the Stanford Computer Science Department. He obtained his Ph.D. in Computer Science at Brown University in 2009. His research interests include syntactic parsing, information extraction, and the interchange between them.