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.
@article{arxiv.2212.05762,
title = {Momentum Contrastive Pre-training for Question Answering},
author = {Minda Hu and Muzhi Li and Yasheng Wang and Irwin King},
journal= {arXiv preprint arXiv:2212.05762},
year = {2023}
}
Comments
This work has been accepted by EMNLP 2022. Reference to ACL Anthology: https://aclanthology.org/2022.emnlp-main.291.pdf