Improving Factual Consistency of Abstractive Summarization via Question Answering
Feng Nan, Cicero Nogueira dos Santos, Henghui Zhu, Patrick Ng, Kathleen McKeown, Ramesh Nallapati, Dejiao Zhang, Zhiguo Wang
Abstract
A commonly observed problem with the state-of-the art abstractive summarization models is that the generated summaries can be factually inconsistent with the input documents. The fact that automatic summarization may produce plausible-sounding yet inaccurate summaries is a major concern that limits its wide application. In this paper we present an approach to address factual consistency in summarization. We first propose an efficient automatic evaluation metric to measure factual consistency; next, we propose a novel learning algorithm that maximizes the proposed metric during model training. Through extensive experiments, we confirm that our method is effective in improving factual consistency and even overall quality of the summaries, as judged by both automatic metrics and human evaluation.
BibTeX
@inproceedings{nan-etal-2021-improving,
title = "Improving Factual Consistency of Abstractive Summarization via Question Answering",
author = "Nan, Feng and
Nogueira dos Santos, Cicero and
Zhu, Henghui and
Ng, Patrick and
McKeown, Kathleen and
Nallapati, Ramesh and
Zhang, Dejiao and
Wang, Zhiguo and
Arnold, Andrew O. and
Xiang, Bing",
editor = "Zong, Chengqing and
Xia, Fei and
Li, Wenjie and
Navigli, Roberto",
booktitle = "Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)",
month = aug,
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2021.acl-long.536/",
doi = "10.18653/v1/2021.acl-long.536",
pages = "6881--6894"
}