Incorporating Singletons and Mention-based Features in Coreference Resolution via Multi-task Learning for Better Generalization

TL;DR

This paper introduces a multi-task learning approach for coreference resolution that incorporates singleton mentions and mention-based features, improving generalization and achieving state-of-the-art results on multiple datasets.

Contribution

It presents a novel multi-task learning model that learns singletons and features like entity type, enhancing coreference resolution performance and robustness.

Findings

Achieves +2.7 points on OntoGUM benchmark.

Increases robustness by +2.3 points on out-of-domain datasets.

Utilizes singleton mention data for better generalization.

Abstract

Previous attempts to incorporate a mention detection step into end-to-end neural coreference resolution for English have been hampered by the lack of singleton mention span data as well as other entity information. This paper presents a coreference model that learns singletons as well as features such as entity type and information status via a multi-task learning-based approach. This approach achieves new state-of-the-art scores on the OntoGUM benchmark (+2.7 points) and increases robustness on multiple out-of-domain datasets (+2.3 points on average), likely due to greater generalizability for mention detection and utilization of more data from singletons when compared to only coreferent mention pair matching.