BioEv: A system for learning biological event extraction

Abstract

Previous research in information extraction from biological texts has focused intensively on the recognition of named entities, such as gene, protein or disease names and on the extraction of simple relations of these entities, such as protein-protein interactions. Recently, the focus of research has been moving to higher levels of information extraction such as co-reference resolution and event extraction. In our work, we are interested in event extraction task which involves the filling of an event template. For each event, we extract its trigger expression, class and arguments. In this paper we describe a system that uses a kernel-based-method for the extraction of biological event templates from literature.