mlsanityData to 7 Oct 2026

Group-Structured Adversarial Training

arXiv 2106.10324 ↗First posted 18 Jun 2021cs.LG stat.MLcited by 0

Not on the map. Why?

Abstract (arXiv)

Robust training methods against perturbations to the input data have received great attention in the machine learning literature. A standard approach in this direction is adversarial training which learns a model using adversarially-perturbed training samples. However, adversarial training performs suboptimally against perturbations structured across samples such as universal and group-sparse shifts that are commonly present in biological data such as gene expression levels of different tissues. In this work, we seek to close this optimality gap and introduce Group-Structured Adversarial Training (GSAT) which learns a model robust to perturbations structured across samples. We formulate GSAT as a non-convex concave minimax optimization problem which minimizes a group-structured optimal transport cost. Specifically, we focus on the applications of GSAT for group-sparse and rank-constrained perturbations modeled using group and nuclear norm penalties. In order to solve GSAT's non-smooth optimization problem in those cases, we propose a new minimax optimization algorithm called GDADMM by combining Gradient Descent Ascent (GDA) and Alternating Direction Method of Multipliers (ADMM). We present several applications of the GSAT framework to gain robustness against structured perturbations for image recognition and computational biology datasets.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 38 of its 59 references that are in the map. First posted on arXiv 18 Jun 2021.

Builds on

38 of 59 references.

References · 38

  1. 2014Explaining and Harnessing Adversarial Examplescited by 7,287
  2. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,326
  3. 2013Intriguing properties of neural networkscited by 5,114
  4. 2016Towards Evaluating the Robustness of Neural Networkscited by 2,976
  5. 2021Convex optimizationcited by 2,479
  6. 2015DeepFool: a simple and accurate method to fool deep neural networkscited by 1,315
  7. 2015Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs)cited by 1,061
  8. 2016Adversarial Machine Learning at Scalecited by 934
  9. 2017Ensemble Adversarial Training: Attacks and Defensescited by 833
  10. 2016Universal adversarial perturbationscited by 774
  11. 2017Evasion Attacks against Machine Learning at Test Timecited by 686
  12. 2016TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systemscited by 660
26 more
  1. 2017Provable defenses against adversarial examples via the convex outer adversarial polytopecited by 564
  2. 2020On Adaptive Attacks to Adversarial Example Defensescited by 547
  3. 2018Certified Defenses against Adversarial Examplescited by 305
  4. 2015Data-driven Distributionally Robust Optimization Using the Wasserstein Metric: Performance Guarantees and Tractable Reformulationscited by 285
  5. 2019On Gradient Descent Ascent for Nonconvex-Concave Minimax Problemscited by 266
  6. 2017EAD: Elastic-Net Attacks to Deep Neural Networks via Adversarial Examplescited by 154
  7. 2015Distributionally Robust Logistic Regressioncited by 124
  8. 2018Are adversarial examples inevitable?cited by 113
  9. 2020Certified Defenses for Adversarial Patchescited by 100
  10. 2019Wasserstein Adversarial Examples via Projected Sinkhorn Iterationscited by 89
  11. 2018Is feature selection secure against training data poisoning?cited by 88
  12. 2018Universal Adversarial Trainingcited by 85
  13. 2019Functional Adversarial Attackscited by 82
  14. 2017Regularization via Mass Transportationcited by 79
  15. 2018Structured Adversarial Attack: Towards General Implementation and Better Interpretabilitycited by 63
  16. 2016Defensive Distillation is Not Robust to Adversarial Examplescited by 60
  17. 2019Robustness Certificates for Sparse Adversarial Attacks by Randomized Ablationcited by 59
  18. 2017Defense against Universal Adversarial Perturbationscited by 48
  19. 2020Second-Order Provable Defenses against Adversarial Attackscited by 39
  20. 2019Wasserstein Smoothing: Certified Robustness against Wasserstein Adversarial Attackscited by 25
  21. 2017Attacking the Madry Defense Model with L1-based Adversarial Examplescited by 21
  22. 2020Breaking certified defenses: Semantic adversarial examples with spoofed robustness certificatescited by 17
  23. 2017Robustness of classifiers to universal perturbations: a geometric perspectivecited by 14
  24. 2020Adversarial Feature Selection against Evasion Attackscited by 13
  25. 2018Sparsity-based Defense against Adversarial Attacks on Linear Classifierscited by 8
  26. 2020Improved Image Wasserstein Attacks and Defensescited by 1

Built on by

Not cited yet.