Semi-supervised anaphora resolution in biomedical texts

Abstract

Resolving anaphora is an important step in the identification of named entities such as genes and proteins in biomedical scientific articles. The goal of this work is to resolve associative and coreferential anaphoric expressions making use of the rich domain resources (such as databases and ontologies) available for the biomedical area, instead of annotated training data. The results are comparable to extant state-of-the-art supervised methods in the same domain. The system is integrated into an interactive tool designed to assist FlyBase curators by aiding the identification of the salient entities in a given paper as a first step in the aggregation of information about them.