Query-Efficient Black-box Adversarial Attacks Guided by a Transfer-based Prior

TL;DR

This paper introduces two transfer-guided, query-efficient black-box adversarial attack algorithms that leverage surrogate model gradients and query feedback to improve success rates with fewer queries.

Contribution

The paper proposes two novel prior-guided gradient-free algorithms that effectively combine transfer-based priors with query information for black-box attacks.

Findings

Require fewer queries than existing methods

Achieve higher attack success rates

Theoretically optimize the integration of priors and queries

Abstract

Adversarial attacks have been extensively studied in recent years since they can identify the vulnerability of deep learning models before deployed. In this paper, we consider the black-box adversarial setting, where the adversary needs to craft adversarial examples without access to the gradients of a target model. Previous methods attempted to approximate the true gradient either by using the transfer gradient of a surrogate white-box model or based on the feedback of model queries. However, the existing methods inevitably suffer from low attack success rates or poor query efficiency since it is difficult to estimate the gradient in a high-dimensional input space with limited information. To address these problems and improve black-box attacks, we propose two prior-guided random gradient-free (PRGF) algorithms based on biased sampling and gradient averaging, respectively. Our methods can take the advantage of a transfer-based prior given by the gradient of a surrogate model and the query information simultaneously. Through theoretical analyses, the transfer-based prior is appropriately integrated with model queries by an optimal coefficient in each method. Extensive experiments demonstrate that, in comparison with the alternative state-of-the-arts, both of our methods require much fewer queries to attack black-box models with higher success rates.