A Minimax Approach Against Multi-Armed Adversarial Attacks Detection

TL;DR

This paper introduces a minimax-based aggregation framework for multiple adversarial example detectors, significantly enhancing detection robustness against multi-armed adversarial attacks on datasets like CIFAR10 and SVHN.

Contribution

It proposes a novel, mathematically sound, and modular aggregation method that improves detection performance against complex multi-armed adversarial attacks.

Findings

Outperforms individual detectors on CIFAR10 and SVHN datasets.

Consistently improves robustness against multi-armed adversarial attacks.

Framework is easy to implement and integrate with existing detectors.

Abstract

Multi-armed adversarial attacks, in which multiple algorithms and objective loss functions are simultaneously used at evaluation time, have been shown to be highly successful in fooling state-of-the-art adversarial examples detectors while requiring no specific side information about the detection mechanism. By formalizing the problem at hand, we can propose a solution that aggregates the soft-probability outputs of multiple pre-trained detectors according to a minimax approach. The proposed framework is mathematically sound, easy to implement, and modular, allowing for integrating existing or future detectors. Through extensive evaluation on popular datasets (e.g., CIFAR10 and SVHN), we show that our aggregation consistently outperforms individual state-of-the-art detectors against multi-armed adversarial attacks, making it an effective solution to improve the resilience of available methods.