mlsanityData to 14 Oct 2026

The Price of Implicit Bias in Adversarially Robust Generalization

arXiv 2406.04981 ↗First posted 7 Jun 2024cs.LG stat.MLcited by 5

Not on the map. Why?

Abstract (arXiv)

We study the implicit bias of optimization in robust empirical risk minimization (robust ERM) and its connection with robust generalization. In classification settings under adversarial perturbations with linear models, we study what type of regularization should ideally be applied for a given perturbation set to improve (robust) generalization. We then show that the implicit bias of optimization in robust ERM can significantly affect the robustness of the model and identify two ways this can happen; either through the optimization algorithm or the architecture. We verify our predictions in simulations with synthetic data and experimentally study the importance of implicit bias in robust ERM with deep neural networks.

Key facts. Cited by 5 papers in the map (data through 14 Oct 2026). Builds on 53 of its 77 references that are in the map. First posted on arXiv 7 Jun 2024.

Builds on

53 of 77 references.

References · 53

  1. 2014Explaining and Harnessing Adversarial Examplescited by 7,364
  2. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,415
  3. 2013Intriguing properties of neural networkscited by 5,155
  4. 2023Universal and Transferable Adversarial Attacks on Aligned Language Modelscited by 3,104
  5. 2016Towards Evaluating the Robustness of Neural Networkscited by 2,998
  6. 2021Convex optimizationcited by 2,515
  7. 2016Adversarial examples in the physical worldcited by 1,769
  8. 2016Understanding deep learning requires rethinking generalizationcited by 1,674
  9. 2019Theoretically Principled Trade-off between Robustness and Accuracycited by 1,517
  10. 2018Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examplescited by 1,268
  11. 2019Adversarial Examples Are Not Bugs, They Are Featurescited by 953
  12. 2016Adversarial Machine Learning at Scalecited by 936
38 more
  1. 2018Robustness May Be at Odds with Accuracycited by 880
  2. 2015The Limitations of Deep Learning in Adversarial Settingscited by 790
  3. 2020Fast is better than free: Revisiting adversarial trainingcited by 727
  4. 2017Evasion Attacks against Machine Learning at Test Timecited by 689
  5. 2017Spectrally-normalized margin bounds for neural networkscited by 686
  6. 2020RobustBench: a standardized adversarial robustness benchmarkcited by 661
  7. 2017The Implicit Bias of Gradient Descent on Separable Datacited by 655
  8. 2019Adversarial Training for Free!cited by 643
  9. 2017Adversarial Examples for Evaluating Reading Comprehension Systemscited by 571
  10. 2020Overfitting in adversarially robust deep learningcited by 506
  11. 2019Implicit Regularization in Deep Matrix Factorizationcited by 405
  12. 2014In Search of the Real Inductive Bias: On the Role of Implicit Regularization in Deep Learningcited by 342
  13. 2017Implicit Regularization in Matrix Factorizationcited by 322
  14. 2019Gradient Descent Maximizes the Margin of Homogeneous Neural Networkscited by 320
  15. 2017A PAC-Bayesian Approach to Spectrally-Normalized Margin Bounds for Neural Networkscited by 314
  16. 2020Kernel and Rich Regimes in Overparametrized Modelscited by 306
  17. 2018Characterizing Implicit Bias in Terms of Optimization Geometrycited by 266
  18. 2018Implicit Bias of Gradient Descent on Linear Convolutional Networkscited by 251
  19. 2023Better Diffusion Models Further Improve Adversarial Trainingcited by 208
  20. 2019You Only Propagate Once: Accelerating Adversarial Training via Maximal Principlecited by 183
  21. 2018Gradient descent aligns the layers of deep linear networkscited by 182
  22. 2018Rademacher Complexity for Adversarially Robust Generalizationcited by 132
  23. 2021Fit without fear: remarkable mathematical phenomena of deep learning through the prism of interpolationcited by 132
  24. 2018Are adversarial examples inevitable?cited by 113
  25. 2022On the Implicit Bias in Deep-Learning Algorithmscited by 86
  26. 2018Adversarial examples from computational constraintscited by 80
  27. 2018Adversarial vulnerability for any classifiercited by 78
  28. 2020Implicit Bias in Deep Linear Classification: Initialization Scale vs Training Accuracycited by 71
  29. 2019Generalization bounds for deep convolutional neural networkscited by 59
  30. 2019Lexicographic and Depth-Sensitive Margins in Homogeneous and Non-Homogeneous Deep Modelscited by 53
  31. 2020Adversarial Learning Guarantees for Linear Hypotheses and Neural Networkscited by 47
  32. 2019Adversarial Robustness May Be at Odds With Simplicitycited by 39
  33. 2013Margins, Shrinkage, and Boostingcited by 35
  34. 2015Understanding Adversarial Training: Increasing Local Stability of Neural Nets through Robust Optimizationcited by 32
  35. 2022Gradient Methods Provably Converge to Non-Robust Networkscited by 27
  36. 2019Hypothesis Set Stability and Generalizationcited by 23
  37. 2023The Double-Edged Sword of Implicit Bias: Generalization vs. Robustness in ReLU Networkscited by 23
  38. 2022What Can the Neural Tangent Kernel Tell Us About Adversarial Robustness?cited by 21

50 most cited of 53 All 53 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

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

5 papers.

  1. 2024Flavors of Margin: Implicit Bias of Steepest Descent in Homogeneous Neural Networkscited by 18
  2. 2024On the Geometry of Regularization in Adversarial Training: High-Dimensional Asymptotics and Generalization Boundscited by 4
  3. 2026Never Saddle for Reparameterized Steepest Descent as Mirror Flowcited by 4
  4. 2025On the Existence of Consistent Adversarial Attacks in High-Dimensional Linear Classificationcited by 2
  5. 2026The Fourth Quadrant: A Stylized View of Benign Misfittingcited by 0