On The Utility of Conditional Generation Based Mutual Information for Characterizing Adversarial Subspaces

TL;DR

This paper introduces a novel approach using mutual information approximated by conditional generation to characterize adversarial subspaces, enhancing neural network robustness and detection of adversarial examples.

Contribution

It proposes a new MI-based metric for adversarial subspace characterization, improving adversary detection and robustness in neural networks.

Findings

MI detector enhances MagNet defense against attacks

Mutual information effectively characterizes adversarial subspaces

Improved robustness in adversary detection methods

Abstract

Recent studies have found that deep learning systems are vulnerable to adversarial examples; e.g., visually unrecognizable adversarial images can easily be crafted to result in misclassification. The robustness of neural networks has been studied extensively in the context of adversary detection, which compares a metric that exhibits strong discriminate power between natural and adversarial examples. In this paper, we propose to characterize the adversarial subspaces through the lens of mutual information (MI) approximated by conditional generation methods. We use MI as an information-theoretic metric to strengthen existing defenses and improve the performance of adversary detection. Experimental results on MagNet defense demonstrate that our proposed MI detector can strengthen its robustness against powerful adversarial attacks.