mlsanityData to 7 Oct 2026

Combating Adversarial Misspellings with Robust Word Recognition

arXiv 1905.11268 ↗First posted 27 May 2019cs.CL cs.CR cs.LGcited by 162

Not on the map. Why?

Abstract (arXiv)

To combat adversarial spelling mistakes, we propose placing a word recognition model in front of the downstream classifier. Our word recognition models build upon the RNN semi-character architecture, introducing several new backoff strategies for handling rare and unseen words. Trained to recognize words corrupted by random adds, drops, swaps, and keyboard mistakes, our method achieves 32% relative (and 3.3% absolute) error reduction over the vanilla semi-character model. Notably, our pipeline confers robustness on the downstream classifier, outperforming both adversarial training and off-the-shelf spell checkers. Against a BERT model fine-tuned for sentiment analysis, a single adversarially-chosen character attack lowers accuracy from 90.3% to 45.8%. Our defense restores accuracy to 75%. Surprisingly, better word recognition does not always entail greater robustness. Our analysis reveals that robustness also depends upon a quantity that we denote the sensitivity.

Key facts. Cited by 162 papers in the map (data through 7 Oct 2026). First posted on arXiv 27 May 2019.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 34; 2022: 41; 2023: 30; 2024: 21; 2025: 17; 2026: 19.

162 papers.

Sort:

By area:

  1. 2021ByT5: Towards a token-free future with pre-trained byte-to-byte modelscited by 504
  2. 2023SmoothLLM: Defending Large Language Models Against Jailbreaking Attackscited by 434
  3. 2021CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representationcited by 229
  4. 2023Do LLMs exhibit human-like response biases? A case study in survey designcited by 176
  5. 2021Charformer: Fast Character Transformers via Gradient-based Subword Tokenizationcited by 158
  6. 2021AMMUS : A Survey of Transformer-based Pretrained Models in Natural Language Processingcited by 131
  7. 2021Measure and Improve Robustness in NLP Models: A Surveycited by 107
  8. 2022Why Should Adversarial Perturbations be Imperceptible? Rethink the Research Paradigm in Adversarial NLPcited by 99
  9. 2022The Fallacy of AI Functionalitycited by 88
  10. 2021Evaluating the Robustness of Neural Language Models to Input Perturbationscited by 87
  11. 2021Towards Robustness Against Natural Language Word Substitutionscited by 68
  12. 2021Triggerless Backdoor Attack for NLP Tasks with Clean Labelscited by 61
38 more
  1. 2022Rickrolling the Artist: Injecting Backdoors into Text Encoders for Text-to-Image Synthesiscited by 59
  2. 2023Adversarial Demonstration Attacks on Large Language Modelscited by 58
  3. 2022Language Modelling with Pixelscited by 53
  4. 2021How Should Pre-Trained Language Models Be Fine-Tuned Towards Adversarial Robustness?cited by 49
  5. 2022CodeAttack: Code-Based Adversarial Attacks for Pre-trained Programming Language Modelscited by 48
  6. 2022Constrained Optimization with Dynamic Bound-scaling for Effective NLPBackdoor Defensecited by 44
  7. 2021LM-Critic: Language Models for Unsupervised Grammatical Error Correctioncited by 33
  8. 2021AMMU : A Survey of Transformer-based Biomedical Pretrained Language Modelscited by 32
  9. 2022Detection of Word Adversarial Examples in Text Classification: Benchmark and Baseline via Robust Density Estimationcited by 32
  10. 2023Text-CRS: A Generalized Certified Robustness Framework against Textual Adversarial Attackscited by 32
  11. 2021Defending Pre-trained Language Models from Adversarial Word Substitutions Without Performance Sacrificecited by 30
  12. 2022Analogy Generation by Prompting Large Language Models: A Case Study of InstructGPTcited by 27
  13. 2024B-AVIBench: Towards Evaluating the Robustness of Large Vision-Language Model on Black-box Adversarial Visual-Instructionscited by 26
  14. 2022"That Is a Suspicious Reaction!": Interpreting Logits Variation to Detect NLP Adversarial Attackscited by 24
  15. 2023Evaluating the Robustness of Text-to-image Diffusion Models against Real-world Attackscited by 24
  16. 2025Bypassing LLM Guardrails: An Empirical Analysis of Evasion Attacks against Prompt Injection and Jailbreak Detection Systemscited by 24
  17. 2022Generalized but not Robust? Comparing the Effects of Data Modification Methods on Out-of-Domain Generalization and Adversarial Robustnesscited by 23
  18. 2022Phrase-level Textual Adversarial Attack with Label Preservationcited by 23
  19. 2022Certified Robustness Against Natural Language Attacks by Causal Interventioncited by 23
  20. 2021Understanding the Generalization Benefit of Model Invariance from a Data Perspectivecited by 22
  21. 2022A Survey of Adversarial Defences and Robustness in NLPcited by 22
  22. 2024A Trembling House of Cards? Mapping Adversarial Attacks against Language Agentscited by 22
  23. 2021Multi-granularity Textual Adversarial Attack with Behavior Cloningcited by 21
  24. 2023How do humans perceive adversarial text? A reality check on the validity and naturalness of word-based adversarial attackscited by 18
  25. 2021BERT-Defense: A Probabilistic Model Based on BERT to Combat Cognitively Inspired Orthographic Adversarial Attackscited by 17
  26. 2021Efficient Combinatorial Optimization for Word-level Adversarial Textual Attackcited by 17
  27. 2021PETGEN: Personalized Text Generation Attack on Deep Sequence Embedding-based Classification Modelscited by 17
  28. 2021On the Robustness of Reading Comprehension Models to Entity Renamingcited by 17
  29. 2021Synthesizing Adversarial Negative Responses for Robust Response Ranking and Evaluationcited by 16
  30. 2024Are AI-Generated Text Detectors Robust to Adversarial Perturbations?cited by 16
  31. 2022On The Empirical Effectiveness of Unrealistic Adversarial Hardening Against Realistic Adversarial Attackscited by 15
  32. 2022On Robust Prefix-Tuning for Text Classificationcited by 15
  33. 2022Towards Explainable Evaluation Metrics for Natural Language Generationcited by 15
  34. 2022TASA: Deceiving Question Answering Models by Twin Answer Sentences Attackcited by 15
  35. 2024LLM Whisperer: An Inconspicuous Attack to Bias LLM Responsescited by 15
  36. 2024Multi-granular Adversarial Attacks against Black-box Neural Ranking Modelscited by 14
  37. 2022Robust Textual Embedding against Word-level Adversarial Attackscited by 12
  38. 2022Perturbations in the Wild: Leveraging Human-Written Text Perturbations for Realistic Adversarial Attack and Defensecited by 12

50 most cited of 162 All 162 (JSON)