MOS-Attack: A Scalable Multi-objective Adversarial Attack Framework

TL;DR

MOS-Attack introduces a multi-objective adversarial attack framework that leverages multiple loss functions and uncovers their interrelations, resulting in more effective adversarial examples for evaluating DNN robustness.

Contribution

The paper presents a novel set-based multi-objective optimization framework for adversarial attacks that automatically discovers and exploits synergistic relationships among multiple loss functions.

Findings

Outperforms single-objective attacks in effectiveness.

Automatically uncovers synergistic patterns among loss functions.

Maintains high attack potency with fewer loss functions.

Abstract

Crafting adversarial examples is crucial for evaluating and enhancing the robustness of Deep Neural Networks (DNNs), presenting a challenge equivalent to maximizing a non-differentiable 0-1 loss function. However, existing single objective methods, namely adversarial attacks focus on a surrogate loss function, do not fully harness the benefits of engaging multiple loss functions, as a result of insufficient understanding of their synergistic and conflicting nature. To overcome these limitations, we propose the Multi-Objective Set-based Attack (MOS Attack), a novel adversarial attack framework leveraging multiple loss functions and automatically uncovering their interrelations. The MOS Attack adopts a set-based multi-objective optimization strategy, enabling the incorporation of numerous loss functions without additional parameters. It also automatically mines synergistic patterns among various losses, facilitating the generation of potent adversarial attacks with fewer objectives. Extensive experiments have shown that our MOS Attack outperforms single-objective attacks. Furthermore, by harnessing the identified synergistic patterns, MOS Attack continues to show superior results with a reduced number of loss functions. Our code is available at https://github.com/pgg3/MOS-Attack.