English

Rademacher Complexity for Adversarially Robust Generalization

Machine Learning 2020-07-30 v4 Cryptography and Security Neural and Evolutionary Computing Machine Learning

Abstract

Many machine learning models are vulnerable to adversarial attacks; for example, adding adversarial perturbations that are imperceptible to humans can often make machine learning models produce wrong predictions with high confidence. Moreover, although we may obtain robust models on the training dataset via adversarial training, in some problems the learned models cannot generalize well to the test data. In this paper, we focus on \ell_\infty attacks, and study the adversarially robust generalization problem through the lens of Rademacher complexity. For binary linear classifiers, we prove tight bounds for the adversarial Rademacher complexity, and show that the adversarial Rademacher complexity is never smaller than its natural counterpart, and it has an unavoidable dimension dependence, unless the weight vector has bounded 1\ell_1 norm. The results also extend to multi-class linear classifiers. For (nonlinear) neural networks, we show that the dimension dependence in the adversarial Rademacher complexity also exists. We further consider a surrogate adversarial loss for one-hidden layer ReLU network and prove margin bounds for this setting. Our results indicate that having 1\ell_1 norm constraints on the weight matrices might be a potential way to improve generalization in the adversarial setting. We demonstrate experimental results that validate our theoretical findings.

Keywords

Cite

@article{arxiv.1810.11914,
  title  = {Rademacher Complexity for Adversarially Robust Generalization},
  author = {Dong Yin and Kannan Ramchandran and Peter Bartlett},
  journal= {arXiv preprint arXiv:1810.11914},
  year   = {2020}
}

Comments

ICML 2019

R2 v1 2026-06-23T04:55:13.921Z