Document-level relation extraction with entity type constraints

doi:10.1016/j.neunet.2025.108469

Document-level relation extraction with entity type constraints

2025-12-11

0

PRE

AI

DOI：10.1016/j.neunet.2025.108469

Original

Original request for help

Save

Abstract

En 中文

• Using relation correlations to solve both long-tail and multi-label problems. • Capturing relation correlations with entity type constraints from the global and local perspectives. • Type-constrained graph is constructed in the global perspective. • Type-constrained loss is designed in the local perspective. • Experiments confirm the effectiveness of proposed model in solving long-tail problem and multi-label problem.

Document-level relation extraction with entity type constraints

Abstract

Journal

Organization