Anaphora and Coreference Resolution, Statistical

Abstract

Anaphora resolution is an important component of natural language processing applications such as information extraction or automatic summarization. Therefore, anaphora have to be resolved in unrestricted input, which can be done with statistical anaphora resolution algorithms. Those algorithms are mostly implemented as binary classification. While this makes the task accessible to standard machine learning techniques, it has the drawback that knowledge about the context is lost. In this article, the generation of data for machine learning–based anaphora resolution algorithms and their evaluation are discussed. The state-of-the-art in statistical anaphora resolution is reviewed. Based on a critical assessment, future directions are given.