Momentum Contrastive Pre-training for Question Answering

Minda Hu, Muzhi Li, Yasheng Wang, Irwin King


Abstract
Existing pre-training methods for extractive Question Answering (QA) generate cloze-like queries different from natural questions in syntax structure, which could overfit pre-trained models to simple keyword matching. In order to address this problem, we propose a novel Momentum Contrastive pRe-training fOr queStion anSwering (MCROSS) method for extractive QA. Specifically, MCROSS introduces a momentum contrastive learning framework to align the answer probability between cloze-like and natural query-passage sample pairs. Hence, the pre-trained models can better transfer the knowledge learned in cloze-like samples to answering natural questions. Experimental results on three benchmarking QA datasets show that our method achieves noticeable improvement compared with all baselines in both supervised and zero-shot scenarios.
Anthology ID:
2022.emnlp-main.291
Original:
2022.emnlp-main.291v1
Version 2:
2022.emnlp-main.291v2
Volume:
Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing
Month:
December
Year:
2022
Address:
Abu Dhabi, United Arab Emirates
Editors:
Yoav Goldberg, Zornitsa Kozareva, Yue Zhang
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
4324–4330
Language:
URL:
https://aclanthology.org/2022.emnlp-main.291
DOI:
10.18653/v1/2022.emnlp-main.291
Bibkey:
Cite (ACL):
Minda Hu, Muzhi Li, Yasheng Wang, and Irwin King. 2022. Momentum Contrastive Pre-training for Question Answering. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 4324–4330, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics.
Cite (Informal):
Momentum Contrastive Pre-training for Question Answering (Hu et al., EMNLP 2022)
Copy Citation:
PDF:
https://aclanthology.org/2022.emnlp-main.291.pdf