arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2201.02009cs.CL
收藏

PAEG:面向神经机器翻译的短语级对抗性实例生成

PAEG: Phrase-level Adversarial Example Generation for Neural Machine Translation

Microsoft Research Asia(微软亚洲研究院);Shanghai Jiao Tong University(上海交通大学);Beihang University(北京航空航天大学)

Juncheng Wan, Jian Yang, Shuming Ma, Dongdong Zhang, Weinan Zhang, Yong Yu, Zhoujun Li

摘要:

虽然端到端神经机器翻译(NMT)已经取得了令人印象深刻的进展,但噪声输入通常会导致模型变得脆弱和不稳定。生成对抗性示例作为增强数据已被证明有助于缓解这个问题。现有的对抗性示例生成方法都是基于词或字符的,忽略了普遍存在的短语结构。在本文中,我们提出了一个短语级对抗性示例生成(PAEG)框架,以提高翻译模型的鲁棒性。我们的方法通过采用短语级替换策略进一步改进了基于梯度的词级AEG方法。我们验证了我们的方法在三个基准,包括LDC中文-英文,IWSLT 14德语-英语,和WMT 14英语-德语的任务。实验结果表明,我们的方法显着提高翻译性能和鲁棒性噪声相比,以前的强基线。

英文摘要:

While end-to-end neural machine translation (NMT) has achieved impressive progress, noisy input usually leads models to become fragile and unstable. Generating adversarial examples as the augmented data has been proved to be useful to alleviate this problem. Existing methods for adversarial example generation (AEG) are word-level or character-level, which ignore the ubiquitous phrase structure. In this paper, we propose a Phrase-level Adversarial Example Generation (PAEG) framework to enhance the robustness of the translation model. Our method further improves the gradient-based word-level AEG method by adopting a phrase-level substitution strategy. We verify our method on three benchmarks, including LDC Chinese-English, IWSLT14 German-English, and WMT14 English-German tasks. Experimental results demonstrate that our approach significantly improves translation performance and robustness to noise compared to previous strong baselines.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑