paper-with-me

홈 › Papers

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

2021-09-15 · Jens Hauser, Zhao Meng, Damián Pascual, Roger Wattenhofer

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.

📄 PDF Abstract BibTeX arXiv:2109.07403

Code (0)

등록된 구현이 없습니다.

Tasks

Data Augmentationtext-classificationText Classification

Methods 이 논문이 사용한 방법론

Attention 설명 없음
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
Layer Normalization Unlike batch normalization, Layer Normalization directly estimates the normalization statistics from the summed inputs…
Linear Warmup With Linear Decay Linear Warmup With Linear Decay is a learning rate schedule in which we increase the learning rate linearly for $n$ updates and then linearly decay afterwards.
Weight Decay 설명 없음
Refunds@Expedia|||How do I get a full refund from Expedia? “How do I get a full refund from Expedia? How do I get a full refund from Expedia? – Call ☎️ +1-(888) 829 (0881) or +1-805-330-4056 or +1-805-330-4056 for Quick Help &…
Adam 설명 없음
Residual Connection 설명 없음

Similar Papers 제목 키워드 기반

Arabic Synonym BERT-based Adversarial Examples for Text Classification

2024-02-05 · Norah Alshahrani, Saied Alshahrani, Esma Wali, Jeanna Matthews

Text classification systems have been proven vulnerable to adversarial text examples, modified versions of the original text examples that are often unnoticed by human eyes, yet can force text classification models to al…

Adversarial TextLanguage ModelingLanguage ModellingMasked Language Modeling+2

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

2021-11-16 · ACL ARR November 2021 11 · Anonymous

In this work, we investigate the robustness of BERT using four word substitution-based attacks. We combine a human evaluation of individual word substitutions and a probabilistic analysis to show that between 96% and 99%…

Data Augmentationtext-classificationText Classification

Defense against Synonym Substitution-based Adversarial Attacks via Dirichlet Neighborhood Ensemble

2021-08-01 · ACL 2021 5 · Yi Zhou, Xiaoqing Zheng, Cho-Jui Hsieh, Kai-Wei Chang 외

Although deep neural networks have achieved prominent performance on many NLP tasks, they are vulnerable to adversarial examples. We propose Dirichlet Neighborhood Ensemble (DNE), a randomized method for training a robus…

Sentence

Fast Adversarial Training against Textual Adversarial Attacks

2024-01-23 · Yichen Yang, Xin Liu, Kun He

Many adversarial defense methods have been proposed to enhance the adversarial robustness of natural language processing models. However, most of them introduce additional pre-set linguistic knowledge and assume that the…

Adversarial DefenseAdversarial Robustness

BAE: BERT-based Adversarial Examples for Text Classification

2020-04-04 · EMNLP 2020 11 · Siddhant Garg, Goutham Ramakrishnan

Modern text classification models are susceptible to adversarial examples, perturbed versions of the original text indiscernible by humans which get misclassified by the model. Recent works in NLP use rule-based synonym …

Adversarial AttackAdversarial TextClassificationGeneral Classification+3