A Survey of Black-Box Adversarial Attacks on Computer Vision Models

TL;DR

This paper provides a comprehensive survey of black-box adversarial attacks and defenses on computer vision models, highlighting the challenges and current research directions in enhancing model robustness against such attacks.

Contribution

It offers a detailed comparative analysis of existing black-box attack methods and defense strategies, filling a gap in understanding their effectiveness and limitations.

Findings

Black-box attacks pose significant threats to model reliability.

Defense techniques vary widely in effectiveness.

The survey identifies key challenges and future research directions.

Abstract

Machine learning has seen tremendous advances in the past few years, which has lead to deep learning models being deployed in varied applications of day-to-day life. Attacks on such models using perturbations, particularly in real-life scenarios, pose a severe challenge to their applicability, pushing research into the direction which aims to enhance the robustness of these models. After the introduction of these perturbations by Szegedy et al. [1], significant amount of research has focused on the reliability of such models, primarily in two aspects - white-box, where the adversary has access to the targeted model and related parameters; and the black-box, which resembles a real-life scenario with the adversary having almost no knowledge of the model to be attacked. To provide a comprehensive security cover, it is essential to identify, study, and build defenses against such attacks. Hence, in this paper, we propose to present a comprehensive comparative study of various black-box adversarial attacks and defense techniques.