mlsanityData to 7 Oct 2026

DiffuseDef: Improved Robustness to Adversarial Attacks via Iterative Denoising

arXiv 2407.00248 ↗First posted 28 Jun 2024cs.CLcited by 1

Not on the map. Why?

Abstract (arXiv)

Pretrained language models have significantly advanced performance across various natural language processing tasks. However, adversarial attacks continue to pose a critical challenge to systems built using these models, as they can be exploited with carefully crafted adversarial texts. Inspired by the ability of diffusion models to predict and reduce noise in computer vision, we propose a novel and flexible adversarial defense method for language classification tasks, DiffuseDef, which incorporates a diffusion layer as a denoiser between the encoder and the classifier. The diffusion layer is trained on top of the existing classifier, ensuring seamless integration with any model in a plug-and-play manner. During inference, the adversarial hidden state is first combined with sampled noise, then denoised iteratively and finally ensembled to produce a robust text representation. By integrating adversarial training, denoising, and ensembling techniques, we show that DiffuseDef improves over existing adversarial defense methods and achieves state-of-the-art performance against common black-box and white-box adversarial attacks.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 38 of its 57 references that are in the map. First posted on arXiv 28 Jun 2024.

Builds on

38 of 57 references.

References · 38

  1. 2017Attention Is All You Needcited by 59,932
  2. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  3. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  4. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,266
  5. 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,005
  6. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,326
  7. 2018GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understandingcited by 4,854
  8. 2016Towards Evaluating the Robustness of Neural Networkscited by 2,976
  9. 2015Character-level Convolutional Networks for Text Classificationcited by 2,293
  10. 2019Is BERT Really Robust? A Strong Baseline for Natural Language Attack on Text Classification and Entailmentcited by 663
  11. 2020BERT-ATTACK: Adversarial Attack Against BERT Using BERTcited by 488
  12. 2018Generating Natural Language Adversarial Examplescited by 422
26 more
  1. 2018TextBugger: Generating Adversarial Text Against Real-world Applicationscited by 376
  2. 2018Black-box Generation of Adversarial Text Sequences to Evade Deep Learning Classifierscited by 341
  3. 2020BAE: BERT-based Adversarial Examples for Text Classificationcited by 316
  4. 2020InfoBERT: Improving Robustness of Language Models from An Information Theoretic Perspectivecited by 92
  5. 2018On Adversarial Examples for Character-Level Neural Machine Translationcited by 85
  6. 2021Towards Improving Adversarial Training of NLP Modelscited by 79
  7. 2020SAFER: A Structure-free Approach for Certified Robustness to Adversarial Word Substitutionscited by 72
  8. 2021Certified Robustness to Text Adversarial Attacks by Randomized [MASK]cited by 57
  9. 2021Searching for an Effective Defender: Benchmarking Defense against Adversarial Word Substitutioncited by 55
  10. 2021Generating Syntactically Controlled Paraphrases without Using Annotated Parallel Pairscited by 45
  11. 2015Denoising Criterion for Variational Auto-Encoding Frameworkcited by 43
  12. 2022SemAttack: Natural Textual Attacks via Different Semantic Spacescited by 42
  13. 2020Better Robustness by More Coverage: Adversarial Training with Mixup Augmentation for Robust Fine-tuningcited by 39
  14. 2021Defending Pre-trained Language Models from Adversarial Word Substitutions Without Performance Sacrificecited by 30
  15. 2021Grey-box Adversarial Attack And Defence For Sentiment Classificationcited by 28
  16. 2022Character-level White-Box Adversarial Attacks against Transformers via Attachable Subwords Substitutioncited by 28
  17. 2021Bridge the Gap Between CV and NLP! A Gradient-based Textual Adversarial Attack Frameworkcited by 27
  18. 2022Textual Manifold-based Defense Against Natural Language Adversarial Examplescited by 19
  19. 2024Advancing the Robustness of Large Language Models through Self-Denoised Smoothingcited by 15
  20. 2023Randomized Smoothing with Masked Inference for Adversarially Robust Text Classificationscited by 11
  21. 2023Don't Retrain, Just Rewrite: Countering Adversarial Perturbations by Rewriting Textcited by 8
  22. 2023DSRM: Boost Textual Adversarial Training with Distribution Shift Risk Minimizationcited by 7
  23. 2022Efficient Adversarial Training with Robust Early-Bird Ticketscited by 5
  24. 2024ROIC-DM: Robust Text Inference and Classification via Diffusion Modelcited by 5
  25. 2024SemRoDe: Macro Adversarial Training to Learn Representations That are Robust to Word-Level Attackscited by 3
  26. 2024PCLD: Point Cloud Layerwise Diffusion for Adversarial Purificationcited by 2

Built on by

Cited by year

Citing papers per year: 2024: 1.

1 paper.

  1. 2024Defensive Dual Masking for Robust Adversarial Defensecited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Beyond Classification: Evaluating Diffusion Denoised Smoothing for Security-Utility Trade off
  2. 2024MaskPure: Improving Defense Against Text Adversaries with Stochastic Purification
  3. 2024Adversarial Robustification via Text-to-Image Diffusion Models
  4. 2026Adversarial Diffusion Across Modalities: A Fusion Survey of Attacks, Defenses, and Evaluation for Text, Vision, and Vision-Language Models
  5. 2025Diffusion-based Cumulative Adversarial Purification for Vision Language Models
  6. 2025CoDefend: Cross-Modal Collaborative Defense via Diffusion Purification and Prompt Optimization
  7. 2023Intriguing Properties of Diffusion Models: An Empirical Study of the Natural Attack Capability in Text-to-Image Generative Models
  8. 2026NADD: Amplifying Noise for Effective Diffusion-based Adversarial Purification