Square Attack: a query-efficient black-box adversarial attack via random search

TL;DR

Square Attack is a query-efficient black-box adversarial attack that uses randomized square-shaped updates, outperforming existing methods in success rate and query efficiency on ImageNet, even surpassing some white-box attacks.

Contribution

Introduces the Square Attack, a novel black-box attack method that does not rely on gradients and achieves higher success rates with fewer queries.

Findings

Improves query efficiency by a factor of 1.8 to 3 on ImageNet.

Achieves higher success rates than state-of-the-art black-box attacks.

Outperforms some white-box attacks in success rate.

Abstract

We propose the Square Attack, a score-based black-box $l_2$- and $l_\infty$-adversarial attack that does not rely on local gradient information and thus is not affected by gradient masking. Square Attack is based on a randomized search scheme which selects localized square-shaped updates at random positions so that at each iteration the perturbation is situated approximately at the boundary of the feasible set. Our method is significantly more query efficient and achieves a higher success rate compared to the state-of-the-art methods, especially in the untargeted setting. In particular, on ImageNet we improve the average query efficiency in the untargeted setting for various deep networks by a factor of at least $1.8$ and up to $3$ compared to the recent state-of-the-art $l_\infty$-attack of Al-Dujaili & O'Reilly. Moreover, although our attack is black-box, it can also outperform gradient-based white-box attacks on the standard benchmarks achieving a new state-of-the-art in terms of the success rate. The code of our attack is available at https://github.com/max-andr/square-attack.