FCDS: Fusing Constituency and Dependency Syntax into Document-Level Relation Extraction

TL;DR

This paper introduces FCDS, a novel method that fuses constituency and dependency syntax to improve document-level relation extraction by better capturing syntactic information and reasoning across sentences.

Contribution

It proposes a new approach combining constituency and dependency syntax for enhanced document-level relation extraction, addressing limitations of previous graph-based methods.

Findings

Significant performance improvements on multiple datasets

Effective integration of constituency and dependency syntax

Code publicly available for reproducibility

Abstract

Document-level Relation Extraction (DocRE) aims to identify relation labels between entities within a single document. It requires handling several sentences and reasoning over them. State-of-the-art DocRE methods use a graph structure to connect entities across the document to capture dependency syntax information. However, this is insufficient to fully exploit the rich syntax information in the document. In this work, we propose to fuse constituency and dependency syntax into DocRE. It uses constituency syntax to aggregate the whole sentence information and select the instructive sentences for the pairs of targets. It exploits the dependency syntax in a graph structure with constituency syntax enhancement and chooses the path between entity pairs based on the dependency graph. The experimental results on datasets from various domains demonstrate the effectiveness of the proposed method. The code is publicly available at this url.