mlsanityData to 14 Oct 2026

Data-Dependent Stability Analysis of Adversarial Training

arXiv 2401.03156 ↗First posted 6 Jan 2024cs.LGcited by 4

Not on the map. Why?

Abstract (arXiv)

Stability analysis is an essential aspect of studying the generalization ability of deep learning, as it involves deriving generalization bounds for stochastic gradient descent-based training algorithms. Adversarial training is the most widely used defense against adversarial example attacks. However, previous generalization bounds for adversarial training have not included information regarding the data distribution. In this paper, we fill this gap by providing generalization bounds for stochastic gradient descent-based adversarial training that incorporate data distribution information. We utilize the concepts of on-average stability and high-order approximate Lipschitz conditions to examine how changes in data distribution and adversarial budget can affect robust generalization gaps. Our derived generalization bounds for both convex and non-convex losses are at least as good as the uniform stability-based counterparts which do not include data distribution information. Furthermore, our findings demonstrate how distribution shifts from data poisoning attacks can impact robust generalization.

Key facts. Cited by 4 papers in the map (data through 14 Oct 2026). Builds on 32 of its 50 references that are in the map. First posted on arXiv 6 Jan 2024.

Builds on

32 of 50 references.

References · 32

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2014Explaining and Harnessing Adversarial Examplescited by 7,364
  3. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,415
  4. 2013Intriguing properties of neural networkscited by 5,155
  5. 2019Theoretically Principled Trade-off between Robustness and Accuracycited by 1,517
  6. 2019Adversarial Examples Are Not Bugs, They Are Featurescited by 953
  7. 2014Deep Neural Networks are Easily Fooled: High Confidence Predictions for Unrecognizable Imagescited by 776
  8. 2017Evasion Attacks against Machine Learning at Test Timecited by 689
  9. 2015Train faster, generalize better: Stability of stochastic gradient descentcited by 552
  10. 2020Overfitting in adversarially robust deep learningcited by 506
  11. 2017Certifying Some Distributional Robustness with Principled Adversarial Trainingcited by 337
  12. 2018Adversarially Robust Generalization Requires More Datacited by 318
20 more
  1. 2015Escaping From Saddle Points --- Online Stochastic Gradient for Tensor Decompositioncited by 243
  2. 2021Improving Robustness using Generated Datacited by 234
  3. 2023Better Diffusion Models Further Improve Adversarial Trainingcited by 208
  4. 2021Unlearnable Examples: Making Personal Data Unexploitablecited by 193
  5. 2020Stability of Stochastic Gradient Descent on Nonsmooth Convex Lossescited by 173
  6. 2021Adversarial Examples Make Strong Poisonscited by 119
  7. 2017Data-Dependent Stability of Stochastic Gradient Descentcited by 92
  8. 2019Learning to Confuse: Generating Training Time Adversarial Data with Auto-Encodercited by 70
  9. 2021Availability Attacks Create Shortcutscited by 62
  10. 2020On the Loss Landscape of Adversarial Training: Identifying Challenges and How to Overcome Themcited by 59
  11. 2021Better Safe Than Sorry: Preventing Delusive Adversaries with Adversarial Trainingcited by 59
  12. 2021Probabilistic Margins for Instance Reweighting in Adversarial Trainingcited by 57
  13. 2020Train simultaneously, generalize better: Stability of gradient-based minimax learnerscited by 44
  14. 2022Robust Unlearnable Examples: Protecting Data Against Adversarial Learningcited by 35
  15. 2022Stability Analysis and Generalization Bounds of Adversarial Trainingcited by 30
  16. 2022Robust Weight Perturbation for Adversarial Trainingcited by 25
  17. 2022Why Robust Generalization in Deep Learning is Difficult: Perspective of Expressive Powercited by 23
  18. 2021Fooling Adversarial Training with Inducing Noisecited by 13
  19. 2022Can Adversarial Training Be Manipulated By Non-Robust Features?cited by 12
  20. 2021Going Grayscale: The Road to Understanding and Improving Unlearnable Examplescited by 10

Built on by

Cited by year

Citing papers per year: 2024: 2; 2025: 2.

4 papers.

  1. 2024Generalization Bound and New Algorithm for Clean-Label Backdoor Attackcited by 11
  2. 2024Generalizability of Memorization Neural Networkscited by 1
  3. 2025Generalizability of Neural Networks Minimizing Empirical Risk Based on Expressive Abilitycited by 0
  4. 2025Adversarial generalization of unfolding (model-based) networkscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Abstract Gradient Training: A Unified Certification Framework for Data Poisoning, Unlearning, and Differential Privacy
  2. 2024Regularization for Adversarial Robust Learning
  3. 2022Towards Alternative Techniques for Improving Adversarial Robustness: Analysis of Adversarial Training at a Spectrum of Perturbations
  4. 2026Closing the Distribution Gap in Adversarial Training for LLMs
  5. 2021Robust Learning Meets Generative Models: Can Proxy Distributions Improve Adversarial Robustness?
  6. 2025Tight Stability Bounds for Robust Distributed Learning: Byzantine Failures Hurt Generalization More than Data Poisoning
  7. 2023Raising the Bar for Certified Adversarial Robustness with Diffusion Models
  8. 2025DARD: Dice Adversarial Robustness Distillation against Adversarial Attacks