mlsanityData to 14 Oct 2026

The King is Naked: on the Notion of Robustness for Natural Language Processing

arXiv 2112.07605 ↗First posted 13 Dec 2021cs.CL cs.AIcited by 16

Not on the map. Why?

Abstract (arXiv)

There is growing evidence that the classical notion of adversarial robustness originally introduced for images has been adopted as a de facto standard by a large part of the NLP research community. We show that this notion is problematic in the context of NLP as it considers a narrow spectrum of linguistic phenomena. In this paper, we argue for semantic robustness, which is better aligned with the human concept of linguistic fidelity. We characterize semantic robustness in terms of biases that it is expected to induce in a model. We study semantic robustness of a range of vanilla and robustly trained architectures using a template-based generative test bed. We complement the analysis with empirical evidence that, despite being harder to implement, semantic robustness can improve performance %gives guarantees for on complex linguistic phenomena where models robust in the classical sense fail.

Key facts. Cited by 16 papers in the map (data through 14 Oct 2026). Builds on 48 of its 67 references that are in the map. First posted on arXiv 13 Dec 2021.

Builds on

48 of 67 references.

References · 48

  1. 2017Attention Is All You Needcited by 60,361
  2. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  3. 2020Language Models are Few-Shot Learnerscited by 32,419
  4. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,305
  5. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,372
  6. 2013Intriguing properties of neural networkscited by 5,155
  7. 2021On the Opportunities and Risks of Foundation Modelscited by 4,033
  8. 2013Distributed Representations of Words and Phrases and their Compositionalitycited by 3,474
  9. 2015Character-level Convolutional Networks for Text Classificationcited by 2,330
  10. 2021What Makes Good In-Context Examples for GPT-3?cited by 1,195
  11. 2020Beyond Accuracy: Behavioral Testing of NLP models with CheckListcited by 1,006
  12. 2018Robustness May Be at Odds with Accuracycited by 880
36 more
  1. 2020A Primer in BERTology: What we know about how BERT workscited by 815
  2. 2017Evasion Attacks against Machine Learning at Test Timecited by 689
  3. 2017Adversarial Examples for Evaluating Reading Comprehension Systemscited by 571
  4. 2020BERT-ATTACK: Adversarial Attack Against BERT Using BERTcited by 491
  5. 2018Threat of Adversarial Attacks on Deep Learning in Computer Vision: A Surveycited by 460
  6. 2018Generating Natural Language Adversarial Examplescited by 424
  7. 2021A Survey of Data Augmentation Approaches for NLPcited by 409
  8. 2021Pay Attention to MLPscited by 368
  9. 2018Black-box Generation of Adversarial Text Sequences to Evade Deep Learning Classifierscited by 345
  10. 2018On the Effectiveness of Interval Bound Propagation for Training Verifiably Robust Modelscited by 328
  11. 2020DCN V2: Improved Deep & Cross Network and Practical Lessons for Web-scale Learning to Rank Systemscited by 327
  12. 2017Synthetic and Natural Noise Both Break Neural Machine Translationcited by 306
  13. 2020Pretrained Transformers Improve Out-of-Distribution Robustnesscited by 294
  14. 2016Safety Verification of Deep Neural Networkscited by 263
  15. 2021Masked Language Modeling and the Distributional Hypothesis: Order Word Matters Pre-training for Littlecited by 177
  16. 2019Certified Robustness to Adversarial Word Substitutionscited by 168
  17. 2019Reducing Sentiment Bias in Language Models via Counterfactual Evaluationcited by 148
  18. 2017Deep Text Classification Can be Fooledcited by 106
  19. 2017Long Text Generation via Adversarial Training with Leaked Informationcited by 100
  20. 2018Seq2Sick: Evaluating the Robustness of Sequence-to-Sequence Models with Adversarial Examplescited by 94
  21. 2019Achieving Verified Robustness to Symbol Substitutions via Interval Bound Propagationcited by 88
  22. 2020Reevaluating Adversarial Examples in Natural Languagecited by 85
  23. 2020Generate Your Counterfactuals: Towards Controlled Counterfactual Generation for Textcited by 75
  24. 2020Adv-BERT: BERT is not robust on misspellings! Generating nature adversarial samples on BERTcited by 69
  25. 2021Towards Robustness Against Natural Language Word Substitutionscited by 69
  26. 2021On Guaranteed Optimal Robust Explanations for NLP Modelscited by 54
  27. 2020Universal Adversarial Attacks with Natural Triggers for Text Classificationcited by 51
  28. 2020Very Deep Transformers for Neural Machine Translationcited by 50
  29. 2017Towards Linguistically Generalizable NLP Systems: A Workshop and Shared Taskcited by 24
  30. 2020What they do when in doubt: a study of inductive biases in seq2seq learnerscited by 20
  31. 2018Ranking Robustness Under Adversarial Document Manipulationscited by 15
  32. 2019Sentiment analysis is not solved! Assessing and probing sentiment classificationcited by 14
  33. 2020Assessing Robustness of Text Classification through Maximal Safe Radius Computationcited by 11
  34. 2019Improving Sentiment Analysis with Multi-task Learning of Negationcited by 8
  35. 2013The Expressive Power of Word Embeddingscited by 7
  36. 2020Elephant in the Room: An Evaluation Framework for Assessing Adversarial Examples in NLPcited by 7

Built on by

Cited by year

Citing papers per year: 2022: 3; 2023: 5; 2024: 2; 2025: 6.

16 papers.

Sort:

By area:

  1. 2022A Survey of Adversarial Defences and Robustness in NLPcited by 23
  2. 2025Robustness and Cybersecurity in the EU Artificial Intelligence Actcited by 11
  3. 2023Language Models as a Service: Overview of a New Paradigm and its Challengescited by 10
  4. 2022A.I. Robustness: a Human-Centered Perspective on Technological Challenges and Opportunitiescited by 9
  5. 2024A Modified Word Saliency-Based Adversarial Attack on Text Classification Modelscited by 9
  6. 2023When to Trust AI: Advances and Challenges for Certification of Neural Networkscited by 6
  7. 2023Break it, Imitate it, Fix it: Robustness by Generating Human-Like Attackscited by 5
  8. 2024PID Control-Based Self-Healing to Improve the Robustness of Large Language Modelscited by 3
  9. 2022Emergent Linguistic Structures in Neural Networks are Fragilecited by 2
  10. 2025Jailbreaking Large Language Models in Infinitely Many Wayscited by 2
  11. 2025Model Hemorrhage and the Robustness Limits of Large Language Modelscited by 1
  12. 2025Bounded PCTL Model Checking of Large Language Model Outputscited by 1
4 more
  1. 2025Verifying Memoryless Sequential Decision-making of Large Language Modelscited by 1
  2. 2023Continuously Reliable Detection of New-Normal Misinformation: Semantic Masking and Contrastive Smoothing in High-Density Latent Regionscited by 0
  3. 2023Perturbed examples reveal invariances shared by language modelscited by 0
  4. 2025Feedback Loops and Code Perturbations in LLM-based Software Engineering: A Case Study on a C-to-Rust Translation Systemcited by 0