English

Sharp Statistical Guarantees for Adversarially Robust Gaussian Classification

Machine Learning 2020-07-01 v1 Machine Learning

Abstract

Adversarial robustness has become a fundamental requirement in modern machine learning applications. Yet, there has been surprisingly little statistical understanding so far. In this paper, we provide the first result of the optimal minimax guarantees for the excess risk for adversarially robust classification, under Gaussian mixture model proposed by \cite{schmidt2018adversarially}. The results are stated in terms of the Adversarial Signal-to-Noise Ratio (AdvSNR), which generalizes a similar notion for standard linear classification to the adversarial setting. For the Gaussian mixtures with AdvSNR value of rr, we establish an excess risk lower bound of order Θ(e(18+o(1))r2dn)\Theta(e^{-(\frac{1}{8}+o(1)) r^2} \frac{d}{n}) and design a computationally efficient estimator that achieves this optimal rate. Our results built upon minimal set of assumptions while cover a wide spectrum of adversarial perturbations including p\ell_p balls for any p1p \ge 1.

Keywords

Cite

@article{arxiv.2006.16384,
  title  = {Sharp Statistical Guarantees for Adversarially Robust Gaussian Classification},
  author = {Chen Dan and Yuting Wei and Pradeep Ravikumar},
  journal= {arXiv preprint arXiv:2006.16384},
  year   = {2020}
}

Comments

25 pages, 1 figure. Accepted by ICML 2020

R2 v1 2026-06-23T16:43:01.513Z