A PDTB-Styled End-to-End Discourse Parser

TL;DR

This paper introduces a comprehensive end-to-end discourse parser based on the PDTB style, capable of identifying, locating, and classifying discourse relations and their arguments, with detailed evaluation.

Contribution

It presents the first full PDTB-style discourse parser that performs relation detection, argument labeling, and relation classification in an integrated manner.

Findings

High accuracy in relation identification and classification

Effective argument span detection for discourse relations

Robust performance demonstrated through comprehensive evaluation

Abstract

We have developed a full discourse parser in the Penn Discourse Treebank (PDTB) style. Our trained parser first identifies all discourse and non-discourse relations, locates and labels their arguments, and then classifies their relation types. When appropriate, the attribution spans to these relations are also determined. We present a comprehensive evaluation from both component-wise and error-cascading perspectives.