Improving Discourse Relation Projection to Build Discourse Annotated Corpora

TL;DR

This paper presents a novel intersection-based method to improve discourse annotation projection across languages, resulting in a French discourse corpus and enhanced classifier performance.

Contribution

It introduces an intersection-based filtering approach to identify unsupported discourse annotations, enabling the creation of a French discourse corpus and improving classification accuracy.

Findings

65% of unsupported annotations identified in Europarl

First PDTB-style French discourse corpus created

15% F1-score improvement in discourse connective classification

Abstract

The naive approach to annotation projection is not effective to project discourse annotations from one language to another because implicit discourse relations are often changed to explicit ones and vice-versa in the translation. In this paper, we propose a novel approach based on the intersection between statistical word-alignment models to identify unsupported discourse annotations. This approach identified 65% of the unsupported annotations in the English-French parallel sentences from Europarl. By filtering out these unsupported annotations, we induced the first PDTB-style discourse annotated corpus for French from Europarl. We then used this corpus to train a classifier to identify the discourse-usage of French discourse connectives and show a 15% improvement of F1-score compared to the classifier trained on the non-filtered annotations.