A Variational Approach for Mitigating Entity Bias in Relation Extraction

TL;DR

This paper introduces a variational information bottleneck method to reduce entity bias in relation extraction models, improving their generalization and robustness across multiple domains and settings.

Contribution

It proposes a novel VIB-based framework that compresses entity information while maintaining task-relevant features, enhancing relation extraction performance.

Findings

Achieves state-of-the-art results across diverse datasets

Improves out-of-domain generalization

Provides a robust and interpretable approach

Abstract

Mitigating entity bias is a critical challenge in Relation Extraction (RE), where models often rely excessively on entities, resulting in poor generalization. This paper presents a novel approach to address this issue by adapting a Variational Information Bottleneck (VIB) framework. Our method compresses entity-specific information while preserving task-relevant features. It achieves state-of-the-art performance on relation extraction datasets across general, financial, and biomedical domains, in both indomain (original test sets) and out-of-domain (modified test sets with type-constrained entity replacements) settings. Our approach offers a robust, interpretable, and theoretically grounded methodology.