一句话：将深度学习变异体作为集成模型增强梯度对抗攻击，提升成功率并突破多种防御。
深度学习中的变异测试虽受关注，但变异体如何真正帮助深度学习尚不明确。本文提出 MUTEN，将变异体视为一种集成模型，并利用集成模型可增强对抗攻击的特性，对变异体实施攻击以提升对梯度掩蔽模型的攻击成功率。作者在 MNIST、SVHN 和 CIFAR-10 上评估，四种主流攻击的成功率最高提升 45%；在 bit-depth reduction、JPEG compression、Defensive distillation 和 Label smoothing 四种防御上，MUTEN 也能有效突破，攻击成功率最高达 96%。
Mutation testing (MT) for deep learning (DL) has gained huge attention in the past few years. However, how MT can really help DL is still unclear. In this paper, we introduce one promising direction for the usage of mutants. Specifically, since mutants can be seen as one kind of ensemble model and ensemble model can be used to boost the adversarial attack, we propose MUTEN, which applies the attack on mutants to improve the success rate of well-known attacks against gradient-masking models. Experimental results on MNIST, SVHN, and CIFAR-10 show that MUTEN can increase the success rate of four attacks by up to 45%. Furthermore, experiments on four defense approaches, bit-depth reduction, JPEG compression, Defensive distillation, and Label smoothing, demonstrate that MUTEN can break the defense models effectively by enhancing the attacks with the success rate of up to 96%.
@inproceedings{HuGCPT23,
title = {MUTEN: Mutant-Based Ensembles for Boosting Gradient-Based Adversarial Attack},
author = {Qiang Hu and Yuejun Guo and Maxime Cordy and Mike Papadakis and Yves Le Traon},
booktitle = {ASE},
year = {2023},
doi = {10.1109/ase56229.2023.00042}
}