Role Semantics for Better Models of Implicit Discourse Relations

TL;DR

This paper introduces role-based semantic features to improve the classification of implicit discourse relations, demonstrating competitive results and providing insights into the effectiveness of role semantics.

Contribution

It presents a novel set of semantic role features for implicit discourse relation classification and analyzes their impact, advancing understanding of role semantics in discourse modeling.

Findings

Role features improve classification accuracy

Results are competitive with existing approaches

Analysis explains when role semantics are most effective

Abstract

Predicting the structure of a discourse is challenging because relations between discourse segments are often implicit and thus hard to distinguish computationally. I extend previous work to classify implicit discourse relations by introducing a novel set of features on the level of semantic roles. My results demonstrate that such features are helpful, yielding results competitive with other feature-rich approaches on the PDTB. My main contribution is an analysis of improvements that can be traced back to role-based features, providing insights into why and when role semantics is helpful.