Assessing Robustness via Score-Based Adversarial Image Generation

TL;DR

This paper introduces ScoreAG, a novel framework using score-based generative models to create unrestricted, semantics-preserving adversarial examples, enhancing robustness evaluation beyond traditional $\\ell_p$-norm constraints.

Contribution

ScoreAG leverages score-based generative models for unrestricted adversarial example generation, improving robustness assessments and surpassing existing attack and defense methods.

Findings

ScoreAG outperforms many state-of-the-art attacks and defenses.

It generates realistic, semantics-preserving adversarial examples.

Purification with ScoreAG enhances classifier robustness.

Abstract

Most adversarial attacks and defenses focus on perturbations within small $\ell_p$-norm constraints. However, $\ell_p$ threat models cannot capture all relevant semantics-preserving perturbations, and hence, the scope of robustness evaluations is limited. In this work, we introduce Score-Based Adversarial Generation (ScoreAG), a novel framework that leverages the advancements in score-based generative models to generate unrestricted adversarial examples that overcome the limitations of $\ell_p$-norm constraints. Unlike traditional methods, ScoreAG maintains the core semantics of images while generating adversarial examples, either by transforming existing images or synthesizing new ones entirely from scratch. We further exploit the generative capability of ScoreAG to purify images, empirically enhancing the robustness of classifiers. Our extensive empirical evaluation demonstrates that ScoreAG improves upon the majority of state-of-the-art attacks and defenses across multiple benchmarks. This work highlights the importance of investigating adversarial examples bounded by semantics rather than $\ell_p$-norm constraints. ScoreAG represents an important step towards more encompassing robustness assessments.