Rewarding Coreference Resolvers for Being Consistent with World Knowledge
Computation and Language
2019-11-13 v2
Abstract
Unresolved coreference is a bottleneck for relation extraction, and high-quality coreference resolvers may produce an output that makes it a lot easier to extract knowledge triples. We show how to improve coreference resolvers by forwarding their input to a relation extraction system and reward the resolvers for producing triples that are found in knowledge bases. Since relation extraction systems can rely on different forms of supervision and be biased in different ways, we obtain the best performance, improving over the state of the art, using multi-task reinforcement learning.
Cite
@article{arxiv.1909.02392,
title = {Rewarding Coreference Resolvers for Being Consistent with World Knowledge},
author = {Rahul Aralikatte and Heather Lent and Ana Valeria Gonzalez and Daniel Hershcovich and Chen Qiu and Anders Sandholm and Michael Ringaard and Anders Søgaard},
journal= {arXiv preprint arXiv:1909.02392},
year = {2019}
}
Comments
To appear in EMNLP 2019 (with corrected Fig. 2)