Towards a More Generalized Approach in Open Relation Extraction

TL;DR

This paper introduces MixORE, a two-phase framework for generalized open relation extraction that effectively handles mixed data containing both known and novel relations, outperforming existing methods.

Contribution

Proposes MixORE, a novel two-phase approach that jointly classifies known relations and clusters novel ones in a generalized OpenRE setting.

Findings

MixORE outperforms baselines in relation classification.

MixORE effectively clusters novel relations.

The approach advances open relation extraction in real-world scenarios.

Abstract

Open Relation Extraction (OpenRE) seeks to identify and extract novel relational facts between named entities from unlabeled data without pre-defined relation schemas. Traditional OpenRE methods typically assume that the unlabeled data consists solely of novel relations or is pre-divided into known and novel instances. However, in real-world scenarios, novel relations are arbitrarily distributed. In this paper, we propose a generalized OpenRE setting that considers unlabeled data as a mixture of both known and novel instances. To address this, we propose MixORE, a two-phase framework that integrates relation classification and clustering to jointly learn known and novel relations. Experiments on three benchmark datasets demonstrate that MixORE consistently outperforms competitive baselines in known relation classification and novel relation clustering. Our findings contribute to the advancement of generalized OpenRE research and real-world applications.