ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 2012.13573
20
10

Robustness, Privacy, and Generalization of Adversarial Training

25 December 2020
Fengxiang He
Shaopeng Fu
Bohan Wang
Dacheng Tao
ArXivPDFHTML
Abstract

Adversarial training can considerably robustify deep neural networks to resist adversarial attacks. However, some works suggested that adversarial training might comprise the privacy-preserving and generalization abilities. This paper establishes and quantifies the privacy-robustness trade-off and generalization-robustness trade-off in adversarial training from both theoretical and empirical aspects. We first define a notion, {\it robustified intensity} to measure the robustness of an adversarial training algorithm. This measure can be approximate empirically by an asymptotically consistent empirical estimator, {\it empirical robustified intensity}. Based on the robustified intensity, we prove that (1) adversarial training is (ε,δ)(\varepsilon, \delta)(ε,δ)-differentially private, where the magnitude of the differential privacy has a positive correlation with the robustified intensity; and (2) the generalization error of adversarial training can be upper bounded by an O(log⁡N/N)\mathcal O(\sqrt{\log N}/N)O(logN​/N) on-average bound and an O(1/N)\mathcal O(1/\sqrt{N})O(1/N​) high-probability bound, both of which have positive correlations with the robustified intensity. Additionally, our generalization bounds do not explicitly rely on the parameter size which would be prohibitively large in deep learning. Systematic experiments on standard datasets, CIFAR-10 and CIFAR-100, are in full agreement with our theories. The source code package is available at \url{https://github.com/fshp971/RPG}.

View on arXiv
Comments on this paper