mlsanityData to 14 Oct 2026

Preserving Semantics in Textual Adversarial Attacks

arXiv 2211.04205 ↗First posted 8 Nov 2022cs.CL cs.AIcited by 5

Not on the map. Why?

Abstract (arXiv)

The growth of hateful online content, or hate speech, has been associated with a global increase in violent crimes against minorities [23]. Harmful online content can be produced easily, automatically and anonymously. Even though, some form of auto-detection is already achieved through text classifiers in NLP, they can be fooled by adversarial attacks. To strengthen existing systems and stay ahead of attackers, we need better adversarial attacks. In this paper, we show that up to 70% of adversarial examples generated by adversarial attacks should be discarded because they do not preserve semantics. We address this core weakness and propose a new, fully supervised sentence embedding technique called Semantics-Preserving-Encoder (SPE). Our method outperforms existing sentence encoders used in adversarial attacks by achieving 1.2x - 5.1x better real attack success rate. We release our code as a plugin that can be used in any existing adversarial attack to improve its quality and speed up its execution.

Key facts. Cited by 5 papers in the map (data through 14 Oct 2026). Builds on 30 of its 46 references that are in the map. First posted on arXiv 8 Nov 2022.

Builds on

30 of 46 references.

References · 30

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  3. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,372
  4. 2014Explaining and Harnessing Adversarial Examplescited by 7,364
  5. 2019BERTScore: Evaluating Text Generation with BERTcited by 5,512
  6. 2013Intriguing properties of neural networkscited by 5,155
  7. 2018GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understandingcited by 4,899
  8. 2019DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lightercited by 4,307
  9. 2013Distributed Representations of Words and Phrases and their Compositionalitycited by 3,474
  10. 2015Character-level Convolutional Networks for Text Classificationcited by 2,330
  11. 2015A large annotated corpus for learning natural language inferencecited by 1,847
  12. 2016Adversarial examples in the physical worldcited by 1,769
18 more
  1. 2016Adversarial Machine Learning at Scalecited by 936
  2. 2016Bag of Tricks for Efficient Text Classificationcited by 876
  3. 2019Is BERT Really Robust? A Strong Baseline for Natural Language Attack on Text Classification and Entailmentcited by 671
  4. 2018Universal Sentence Encodercited by 489
  5. 2018Generating Natural Language Adversarial Examplescited by 424
  6. 2020TweetEval: Unified Benchmark and Comparative Evaluation for Tweet Classificationcited by 417
  7. 2018TextBugger: Generating Adversarial Text Against Real-world Applicationscited by 381
  8. 2018Black-box Generation of Adversarial Text Sequences to Evade Deep Learning Classifierscited by 345
  9. 2015Skip-Thought Vectorscited by 322
  10. 2020BAE: BERT-based Adversarial Examples for Text Classificationcited by 316
  11. 2017Efficient Natural Language Response Suggestion for Smart Replycited by 229
  12. 2017Generating Natural Adversarial Examplescited by 183
  13. 2019Combating Adversarial Misspellings with Robust Word Recognitioncited by 163
  14. 2020Contextualized Perturbation for Textual Adversarial Attackcited by 136
  15. 2020Reevaluating Adversarial Examples in Natural Languagecited by 85
  16. 2021Towards Robustness Against Natural Language Word Substitutionscited by 69
  17. 2020Robust Encodings: A Framework for Combating Adversarial Typoscited by 58
  18. 2021Model Extraction and Adversarial Transferability, Your BERT is Vulnerable!cited by 53

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 2; 2024: 1; 2025: 1; 2026: 1.

5 papers.

  1. 2026Prompt Overflow: What the Guardrail Inspects Is Not What the Model Inferscited by 3
  2. 2023Sample Attackability in Natural Language Adversarial Attackscited by 2
  3. 2023Alquist 5.0: Dialogue Trees Meet Generative Models. A Novel Approach for Enhancing SocialBot Conversationscited by 1
  4. 2024Adversarial Attack for Explanation Robustness of Rationalization Modelscited by 1
  5. 2025Adversarially Robust Detection of Harmful Online Content: A Computational Design Science Approachcited by 0