Towards a More Generalized Approach in Open Relation Extraction
This work addresses a more realistic scenario in OpenRE for applications like information extraction, though it appears incremental by extending existing methods to handle mixed data.
The paper tackles the problem of Open Relation Extraction (OpenRE) by proposing a generalized setting where unlabeled data contains a mixture of known and novel relations, and introduces MixORE, a two-phase framework that outperforms baselines in classification and clustering tasks on three benchmark datasets.
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.