mlsanityData to 7 Oct 2026

The Best Defense is Attack: Repairing Semantics in Textual Adversarial Examples

arXiv 2305.04067 ↗First posted 6 May 2023cs.CLcited by 3

Not on the map. Why?

Abstract (arXiv)

Recent studies have revealed the vulnerability of pre-trained language models to adversarial attacks. Existing adversarial defense techniques attempt to reconstruct adversarial examples within feature or text spaces. However, these methods struggle to effectively repair the semantics in adversarial examples, resulting in unsatisfactory performance and limiting their practical utility. To repair the semantics in adversarial examples, we introduce a novel approach named Reactive Perturbation Defocusing (Rapid). Rapid employs an adversarial detector to identify fake labels of adversarial examples and leverage adversarial attackers to repair the semantics in adversarial examples. Our extensive experimental results conducted on four public datasets, convincingly demonstrate the effectiveness of Rapid in various adversarial attack scenarios. To address the problem of defense performance validation in previous works, we provide a demonstration of adversarial detection and repair based on our work, which can be easily evaluated at https://tinyurl.com/22ercuf8.

Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Builds on 38 of its 62 references that are in the map. First posted on arXiv 6 May 2023.

Builds on

38 of 62 references.

References · 38

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  2. 2018GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understandingcited by 4,854
  3. 2015Character-level Convolutional Networks for Text Classificationcited by 2,293
  4. 2019SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systemscited by 1,718
  5. 2018Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examplescited by 1,259
  6. 2021DeBERTaV3: Improving DeBERTa using ELECTRA-Style Pre-Training with Gradient-Disentangled Embedding Sharingcited by 1,219
  7. 2019Is BERT Really Robust? A Strong Baseline for Natural Language Attack on Text Classification and Entailmentcited by 663
  8. 2020BERT-ATTACK: Adversarial Attack Against BERT Using BERTcited by 488
  9. 2018Generating Natural Language Adversarial Examplescited by 422
  10. 2018TextBugger: Generating Adversarial Text Against Real-world Applicationscited by 376
  11. 2018Black-box Generation of Adversarial Text Sequences to Evade Deep Learning Classifierscited by 341
  12. 2020BAE: BERT-based Adversarial Examples for Text Classificationcited by 316
26 more
  1. 2016Adversarial Training Methods for Semi-Supervised Text Classificationcited by 280
  2. 2019FreeLB: Enhanced Adversarial Training for Natural Language Understandingcited by 261
  3. 2021Gradient-based Adversarial Attacks against Text Transformerscited by 233
  4. 2021Improving Robustness using Generated Datacited by 233
  5. 2023Better Diffusion Models Further Improve Adversarial Trainingcited by 207
  6. 2017Generating Natural Adversarial Examplescited by 182
  7. 2019Combating Adversarial Misspellings with Robust Word Recognitioncited by 162
  8. 2020Contextualized Perturbation for Textual Adversarial Attackcited by 136
  9. 2018Decoupling Direction and Norm for Efficient Gradient-Based L2 Adversarial Attacks and Defensescited by 109
  10. 2019Robust Neural Machine Translation with Doubly Adversarial Inputscited by 105
  11. 2020OpenAttack: An Open-source Textual Adversarial Attack Toolkitcited by 84
  12. 2021Towards Robustness Against Natural Language Word Substitutionscited by 68
  13. 2020Robust Encodings: A Framework for Combating Adversarial Typoscited by 58
  14. 2019Learning to Discriminate Perturbations for Blocking Adversarial Attacks in Text Classificationcited by 55
  15. 2020Frequency-Guided Word Substitutions for Detecting Textual Adversarial Examplescited by 54
  16. 2021How Should Pre-Trained Language Models Be Fine-Tuned Towards Adversarial Robustness?cited by 49
  17. 2022SemAttack: Natural Textual Attacks via Different Semantic Spacescited by 42
  18. 2019Natural Language Adversarial Defense through Synonym Encodingcited by 39
  19. 2018Greedy Attack and Gumbel Attack: Generating Adversarial Examples for Discrete Datacited by 37
  20. 2023Exploring and Exploiting Decision Boundary Dynamics for Adversarial Robustnesscited by 35
  21. 2021Defending Pre-trained Language Models from Adversarial Word Substitutions Without Performance Sacrificecited by 30
  22. 2021Achieving Model Robustness through Discrete Adversarial Trainingcited by 23
  23. 2020Mind Your Inflections! Improving NLP for Non-Standard Englishes with Base-Inflection Encodingcited by 22
  24. 2021BERT-Defense: A Probabilistic Model Based on BERT to Combat Cognitively Inspired Orthographic Adversarial Attackscited by 17
  25. 2021Detecting Textual Adversarial Examples through Randomized Substitution and Votecited by 8
  26. 2023TextShield: Beyond Successfully Detecting Adversarial Sentences in Text Classificationcited by 8

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 1; 2025: 1; 2026: 1.

3 papers.

  1. 2024Enhancing adversarial robustness in Natural Language Inference using explanationscited by 2
  2. 2025Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attackcited by 0
  3. 2026Rectifying Adversarial Examples Using Their Vulnerabilitiescited by 0