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

TL;DR

This paper analyzes the robustness of pre-trained Seq2Seq models like BART, finds vulnerabilities affecting faithfulness and informativeness, and introduces AdvSeq, an adversarial augmentation framework that significantly enhances these qualities.

Contribution

It provides the first quantitative analysis of Seq2Seq robustness and proposes a novel adversarial augmentation method, AdvSeq, to improve faithfulness and informativeness.

Findings

AdvSeq improves faithfulness in text generation.

AdvSeq enhances informativeness of Seq2Seq models.

Experimental results show significant gains in robustness.

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.