Unveiling the Role of Randomization in Multiclass Adversarial Classification: Insights from Graph Theory

Lucas Gnecco Heredia, Matteo Sammut, Muni Sreenivas Pydi, Rafael Pinot, benjamin negrevergne, Yann Chevaleyre

TL;DR: This paper explores how randomization can improve adversarial robustness in multiclass classification by using graph theory.

Abstract: Randomization as a mean to improve the adversarial robustness of machine learning models has recently attracted significant attention. Unfortunately, much of the theoretical analysis so far has focused on binary classification, providing only limited insights into the more complex multiclass setting. In this paper, we take a step toward closing this gap by drawing inspiration from the field of graph theory. Our analysis focuses on discrete data distributions, allowing us to cast the adversarial risk minimization problems within the well-established framework of set packing problems. By doing so, we are able to identify three structural conditions on the support of the data distribution that are necessary for randomization to improve robustness. Furthermore, we are able to construct several data distributions where (contrarily to binary classification) switching from a deterministic to a randomized solution significantly reduces the optimal adversarial risk. These findings highlight the crucial role randomization can play in enhancing robustness to adversarial attacks in multiclass classification.

Submission Number: 1373