Precisely the Point: Adversarial Augmentations for Faithful and Informative Text Generation

Wenhao Wu, Wei Li, Jiachen Liu, Xinyan Xiao, Sujian Li, Yajuan Lyu


Abstract
Though model robustness has been extensively studied in language understanding, the robustness of Seq2Seq generation remains understudied. In this paper, we conduct the first quantitative analysis on the robustness of pre-trained Seq2Seq models. We find that even current SOTA pre-trained Seq2Seq model (BART) is still vulnerable, which leads to significant degeneration in faithfulness and informativeness for text generation tasks. This motivated us to further propose a novel adversarial augmentation framework, namely AdvSeq, for generally improving faithfulness and informativeness of Seq2Seq models via enhancing their robustness. AdvSeq automatically constructs two types of adversarial augmentations during training, including implicit adversarial samples by perturbing word representations and explicit adversarial samples by word swapping, both of which effectively improve Seq2Seq robustness. Extensive experiments on three popular text generation tasks demonstrate that AdvSeq significantly improves both the faithfulness and informativeness of Seq2Seq generation under both automatic and human evaluation settings.
Anthology ID:
2022.emnlp-main.482
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:
7160–7176
Language:
URL:
https://aclanthology.org/2022.emnlp-main.482
DOI:
10.18653/v1/2022.emnlp-main.482
Bibkey:
Cite (ACL):
Wenhao Wu, Wei Li, Jiachen Liu, Xinyan Xiao, Sujian Li, and Yajuan Lyu. 2022. Precisely the Point: Adversarial Augmentations for Faithful and Informative Text Generation. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 7160–7176, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics.
Cite (Informal):
Precisely the Point: Adversarial Augmentations for Faithful and Informative Text Generation (Wu et al., EMNLP 2022)
Copy Citation:
PDF:
https://preview.aclanthology.org/emnlp22-frontmatter/2022.emnlp-main.482.pdf