arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2109.07403cs.CL
收藏

BERT是强大的!文本分类中基于同义词的对抗性实例

BERT is Robust! A Case Against Synonym-Based Adversarial Examples in Text Classification

ETH Zurich(苏黎世联邦理工学院)

Jens Hauser, Zhao Meng, Damián Pascual, Roger Wattenhofer

摘要:

深度神经网络已经席卷了自然语言处理。虽然这导致了许多任务的惊人改进,但它也开创了一个新的研究领域,通过攻击这些神经网络来质疑它们的鲁棒性。在本文中,我们研究了四个字替换为基础的攻击BERT。我们结合了对单个单词替换的人工评估和概率分析,结果表明,96%到99%的分析攻击不保留语义,这表明它们的成功主要是基于向模型提供不良数据。为了进一步证实这一点,我们引入了一个有效的数据增强过程,并表明可以通过在训练过程中包含类似于攻击的数据来防止许多对抗性示例。额外的后处理步骤将最先进攻击的成功率降低到5%以下。最后,通过寻找更合理的阈值对词替换的约束,我们得出结论,BERT是一个更强大的攻击研究表明。

英文摘要:

Deep Neural Networks have taken Natural Language Processing by storm. While this led to incredible improvements across many tasks, it also initiated a new research field, questioning the robustness of these neural networks by attacking them. In this paper, we investigate four word substitution-based attacks on BERT. We combine a human evaluation of individual word substitutions and a probabilistic analysis to show that between 96% and 99% of the analyzed attacks do not preserve semantics, indicating that their success is mainly based on feeding poor data to the model. To further confirm that, we introduce an efficient data augmentation procedure and show that many adversarial examples can be prevented by including data similar to the attacks during training. An additional post-processing step reduces the success rates of state-of-the-art attacks below 5%. Finally, by looking at more reasonable thresholds on constraints for word substitutions, we conclude that BERT is a lot more robust than research on attacks suggests.

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。

↑