Simple and Efficient Hard Label Black-box Adversarial Attacks in Low Query Budget Regimes
Abstract
We focus on the problem of black-box adversarial attacks, where the aim is to generate adversarial examples for deep learning models solely based on information limited to output label~(hard label) to a queried data input. We propose a simple and efficient Bayesian Optimization~(BO) based approach for developing black-box adversarial attacks. Issues with BO's performance in high dimensions are avoided by searching for adversarial examples in a structured low-dimensional subspace. We demonstrate the efficacy of our proposed attack method by evaluating both and norm constrained untargeted and targeted hard label black-box attacks on three standard datasets - MNIST, CIFAR-10 and ImageNet. Our proposed approach consistently achieves 2x to 10x higher attack success rate while requiring 10x to 20x fewer queries compared to the current state-of-the-art black-box adversarial attacks.
Cite
@article{arxiv.2007.07210,
title = {Simple and Efficient Hard Label Black-box Adversarial Attacks in Low Query Budget Regimes},
author = {Satya Narayan Shukla and Anit Kumar Sahu and Devin Willmott and J. Zico Kolter},
journal= {arXiv preprint arXiv:2007.07210},
year = {2021}
}
Comments
Accepted at KDD 2021. arXiv admin note: substantial text overlap with arXiv:1909.13857