A Geometry-Inspired Attack for Generating Natural Language Adversarial Examples
Not on the map. Why?
Abstract (arXiv)
Generating adversarial examples for natural language is hard, as natural language consists of discrete symbols, and examples are often of variable lengths. In this paper, we propose a geometry-inspired attack for generating natural language adversarial examples. Our attack generates adversarial examples by iteratively approximating the decision boundary of Deep Neural Networks (DNNs). Experiments on two datasets with two different models show that our attack fools natural language models with high success rates, while only replacing a few words. Human evaluation shows that adversarial examples generated by our attack are hard for humans to recognize. Further experiments show that adversarial training can improve model robustness against our attack.
Key facts. Cited by 20 papers in the map (data through 7 Oct 2026). First posted on arXiv 3 Oct 2020.
Builds on
No reference list yet.
Built on by
Citing papers per year: 2021: 6; 2022: 6; 2023: 2; 2024: 2; 2025: 1; 2026: 3.
20 papers.
Sort:
By area:
- 2023On Evaluating Adversarial Robustness of Large Vision-Language Modelscited by 281
- 2023Evaluating the Robustness of Text-to-image Diffusion Models against Real-world Attackscited by 24
- 2022TextGrad: Advancing Robustness Evaluation in NLP by Gradient-Driven Optimizationcited by 21
- 2024Assessing Adversarial Robustness of Large Language Models: An Empirical Studycited by 21
- 2024HQA-Attack: Toward High Quality Black-Box Hard-Label Adversarial Attack on Textcited by 16
- 2021Token-Modification Adversarial Attacks for Natural Language Processing: A Surveycited by 12
- 2022Robust Textual Embedding against Word-level Adversarial Attackscited by 12
- 2022TextHacker: Learning based Hybrid Local Search Algorithm for Text Hard-label Adversarial Attackcited by 10
- 2021A Differentiable Language Model Adversarial Attack on Text Classifierscited by 9
- 2021BERT is Robust! A Case Against Synonym-Based Adversarial Examples in Text Classificationcited by 9
- 2022Rethinking Textual Adversarial Defense for Pre-trained Language Modelscited by 5
- 2021Self-Supervised Contrastive Learning with Adversarial Perturbations for Defending Word Substitution-based Attackscited by 4
8 more
- 2022Distinguishing Non-natural from Natural Adversarial Samples for More Robust Pre-trained Language Modelcited by 3
- 2022A Context-Aware Approach for Textual Adversarial Attack through Probability Difference Guided Beam Searchcited by 1
- 2025Multi-task Adversarial Attacks against Black-box Model with Few-shot Queriescited by 1
- 2021SEPP: Similarity Estimation of Predicted Probabilities for Defending and Detecting Adversarial Textcited by 0
- 2021Generating Watermarked Adversarial Textscited by 0
- 2026PivotAttack: Rethinking the Search Trajectory in Hard-Label Text Attacks via Pivot Wordscited by 0
- 2026SEP-Attack: A Simple and Effective Paradigm for Transfer-Based Textual Adversarial Attackcited by 0
- 2026OASIS: Optimizing Attacker Sequences for Hard-Label Black-Box Text Attackscited by 0
- 2026OASIS: Optimizing Attacker Sequences for Hard-Label Black-Box Text Attacks30 Aug 2026
- 2026SEP-Attack: A Simple and Effective Paradigm for Transfer-Based Textual Adversarial Attack24 May 2026
- 2026PivotAttack: Rethinking the Search Trajectory in Hard-Label Text Attacks via Pivot Words11 Mar 2026
- 2025Multi-task Adversarial Attacks against Black-box Model with Few-shot Queries10 Aug 2025
- 2024Assessing Adversarial Robustness of Large Language Models: An Empirical Study4 May 2024
- 2024HQA-Attack: Toward High Quality Black-Box Hard-Label Adversarial Attack on Text2 Feb 2024
- 2023Evaluating the Robustness of Text-to-image Diffusion Models against Real-world Attacks16 Jun 2023
- 2023On Evaluating Adversarial Robustness of Large Vision-Language Models26 May 2023
- 2022TextGrad: Advancing Robustness Evaluation in NLP by Gradient-Driven Optimization19 Dec 2022
- 2022A Context-Aware Approach for Textual Adversarial Attack through Probability Difference Guided Beam Search17 Aug 2022
- 2022Rethinking Textual Adversarial Defense for Pre-trained Language Models21 Jul 2022
- 2022Distinguishing Non-natural from Natural Adversarial Samples for More Robust Pre-trained Language Model19 Mar 2022
8 more
- 2022Robust Textual Embedding against Word-level Adversarial Attacks28 Feb 2022
- 2022TextHacker: Learning based Hybrid Local Search Algorithm for Text Hard-label Adversarial Attack20 Jan 2022
- 2021Generating Watermarked Adversarial Texts25 Oct 2021
- 2021SEPP: Similarity Estimation of Predicted Probabilities for Defending and Detecting Adversarial Text12 Oct 2021
- 2021BERT is Robust! A Case Against Synonym-Based Adversarial Examples in Text Classification15 Sep 2021
- 2021A Differentiable Language Model Adversarial Attack on Text Classifiers23 Jul 2021
- 2021Self-Supervised Contrastive Learning with Adversarial Perturbations for Defending Word Substitution-based Attacks15 Jul 2021
- 2021Token-Modification Adversarial Attacks for Natural Language Processing: A Survey1 Mar 2021
- 2021Generating Watermarked Adversarial Textscited by 0
- 2021Generating Watermarked Adversarial Texts25 Oct 2021
- 2022TextGrad: Advancing Robustness Evaluation in NLP by Gradient-Driven Optimizationcited by 21
- 2024Assessing Adversarial Robustness of Large Language Models: An Empirical Studycited by 21
- 2024HQA-Attack: Toward High Quality Black-Box Hard-Label Adversarial Attack on Textcited by 16
- 2021Token-Modification Adversarial Attacks for Natural Language Processing: A Surveycited by 12
- 2022Robust Textual Embedding against Word-level Adversarial Attackscited by 12
- 2022TextHacker: Learning based Hybrid Local Search Algorithm for Text Hard-label Adversarial Attackcited by 10
- 2021A Differentiable Language Model Adversarial Attack on Text Classifierscited by 9
- 2021BERT is Robust! A Case Against Synonym-Based Adversarial Examples in Text Classificationcited by 9
- 2022Rethinking Textual Adversarial Defense for Pre-trained Language Modelscited by 5
- 2021Self-Supervised Contrastive Learning with Adversarial Perturbations for Defending Word Substitution-based Attackscited by 4
- 2022Distinguishing Non-natural from Natural Adversarial Samples for More Robust Pre-trained Language Modelcited by 3
- 2022A Context-Aware Approach for Textual Adversarial Attack through Probability Difference Guided Beam Searchcited by 1
12 most cited of 17 All 17 (JSON)
- 2026OASIS: Optimizing Attacker Sequences for Hard-Label Black-Box Text Attacks30 Aug 2026
- 2026SEP-Attack: A Simple and Effective Paradigm for Transfer-Based Textual Adversarial Attack24 May 2026
- 2026PivotAttack: Rethinking the Search Trajectory in Hard-Label Text Attacks via Pivot Words11 Mar 2026
- 2025Multi-task Adversarial Attacks against Black-box Model with Few-shot Queries10 Aug 2025
- 2024Assessing Adversarial Robustness of Large Language Models: An Empirical Study4 May 2024
- 2024HQA-Attack: Toward High Quality Black-Box Hard-Label Adversarial Attack on Text2 Feb 2024
- 2022TextGrad: Advancing Robustness Evaluation in NLP by Gradient-Driven Optimization19 Dec 2022
- 2022A Context-Aware Approach for Textual Adversarial Attack through Probability Difference Guided Beam Search17 Aug 2022
- 2022Rethinking Textual Adversarial Defense for Pre-trained Language Models21 Jul 2022
- 2022Distinguishing Non-natural from Natural Adversarial Samples for More Robust Pre-trained Language Model19 Mar 2022
- 2022Robust Textual Embedding against Word-level Adversarial Attacks28 Feb 2022
- 2022TextHacker: Learning based Hybrid Local Search Algorithm for Text Hard-label Adversarial Attack20 Jan 2022
12 newest of 17 All 17 (JSON)