Entity-Augmented Distributional Semantics for Discourse Relations

TL;DR

This paper introduces an entity-augmented compositional distributional semantics model that improves the automatic prediction of discourse relations by incorporating entity mention representations, leading to state-of-the-art results.

Contribution

It proposes a novel downward compositional pass to compute entity mention representations, enhancing discourse relation prediction beyond sentence-level semantics.

Findings

Significant improvement over previous models in predicting implicit discourse relations.

Effective integration of entity mentions improves semantic understanding of discourse.

Achieved state-of-the-art performance on the Penn Discourse Treebank.

Abstract

Discourse relations bind smaller linguistic elements into coherent texts. However, automatically identifying discourse relations is difficult, because it requires understanding the semantics of the linked sentences. A more subtle challenge is that it is not enough to represent the meaning of each sentence of a discourse relation, because the relation may depend on links between lower-level elements, such as entity mentions. Our solution computes distributional meaning representations by composition up the syntactic parse tree. A key difference from previous work on compositional distributional semantics is that we also compute representations for entity mentions, using a novel downward compositional pass. Discourse relations are predicted not only from the distributional representations of the sentences, but also of their coreferent entity mentions. The resulting system obtains substantial improvements over the previous state-of-the-art in predicting implicit discourse relations in the Penn Discourse Treebank.