Facilitating Contrastive Learning of Discourse Relational Senses by Exploiting the Hierarchy of Sense Relations

TL;DR

This paper introduces a novel contrastive learning approach that leverages the hierarchical structure of discourse senses to improve implicit discourse relation recognition, achieving state-of-the-art results.

Contribution

It integrates the sense hierarchy into the recognition process and negative sampling, enhancing contrastive learning for discourse relation classification.

Findings

Achieved state-of-the-art performance on PDTB-2 and PDTB-3 datasets.

Effectively utilizes sense hierarchy for negative example selection.

Improves implicit discourse relation recognition accuracy.

Abstract

Implicit discourse relation recognition is a challenging task that involves identifying the sense or senses that hold between two adjacent spans of text, in the absence of an explicit connective between them. In both PDTB-2 and PDTB-3, discourse relational senses are organized into a three-level hierarchy ranging from four broad top-level senses, to more specific senses below them. Most previous work on implicit discourse relation recognition have used the sense hierarchy simply to indicate what sense labels were available. Here we do more -- incorporating the sense hierarchy into the recognition process itself and using it to select the negative examples used in contrastive learning. With no additional effort, the approach achieves state-of-the-art performance on the task.