2003 Reports

Improving the Coherence of Multi-document Summaries: A Corpus Study for Modeling the Syntactic Realization of Entities

Nenkova, Ani; McKeown, Kathleen

References included in multi-document summaries are often problematic. In this paper, we present a corpus study performed to derive statistical models for the syntactic realization of referential expressions. Our work shows how the syntactic realization of entities can influence the coherence of the text and provides a model for rewriting references in multi-document summaries to smooth disfluencies. It shows how the syntactic realization of entities can influence the coherence of the text and how rewrite change s can smooth the disfluencies. A large corpus study is conducted in order to derive initial models for syntactic realization.