Don’t understand a measure? Learn it: Structured Prediction for Coreference Resolution optimizing its measures

Iryna Haponchyk, Alessandro Moschitti


Abstract
An interesting aspect of structured prediction is the evaluation of an output structure against the gold standard. Especially in the loss-augmented setting, the need of finding the max-violating constraint has severely limited the expressivity of effective loss functions. In this paper, we trade off exact computation for enabling the use and study of more complex loss functions for coreference resolution. Most interestingly, we show that such functions can be (i) automatically learned also from controversial but commonly accepted coreference measures, e.g., MELA, and (ii) successfully used in learning algorithms. The accurate model comparison on the standard CoNLL-2012 setting shows the benefit of more expressive loss functions.
Anthology ID:
P17-1094
Volume:
Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
Month:
July
Year:
2017
Address:
Vancouver, Canada
Editors:
Regina Barzilay, Min-Yen Kan
Venue:
ACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
1018–1028
Language:
URL:
https://aclanthology.org/P17-1094
DOI:
10.18653/v1/P17-1094
Bibkey:
Cite (ACL):
Iryna Haponchyk and Alessandro Moschitti. 2017. Don’t understand a measure? Learn it: Structured Prediction for Coreference Resolution optimizing its measures. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 1018–1028, Vancouver, Canada. Association for Computational Linguistics.
Cite (Informal):
Don’t understand a measure? Learn it: Structured Prediction for Coreference Resolution optimizing its measures (Haponchyk & Moschitti, ACL 2017)
Copy Citation:
PDF:
https://aclanthology.org/P17-1094.pdf
Video:
 https://aclanthology.org/P17-1094.mp4
Data
CoNLL-2012