Unrestricted Adversarial Attacks on ImageNet Competition

TL;DR

This paper discusses a competition focused on developing and evaluating unrestricted adversarial attacks on ImageNet, where attackers make large, visible modifications to images to fool models, highlighting the need for more robust defenses.

Contribution

It organizes a competition to foster research on effective unrestricted adversarial attack algorithms, advancing understanding of model robustness against unbounded attacks.

Findings

Participants developed novel attack algorithms.

The competition revealed vulnerabilities in current models.

Results highlight the importance of robust defense strategies.

Abstract

Many works have investigated the adversarial attacks or defenses under the settings where a bounded and imperceptible perturbation can be added to the input. However in the real-world, the attacker does not need to comply with this restriction. In fact, more threats to the deep model come from unrestricted adversarial examples, that is, the attacker makes large and visible modifications on the image, which causes the model classifying mistakenly, but does not affect the normal observation in human perspective. Unrestricted adversarial attack is a popular and practical direction but has not been studied thoroughly. We organize this competition with the purpose of exploring more effective unrestricted adversarial attack algorithm, so as to accelerate the academical research on the model robustness under stronger unbounded attacks. The competition is held on the TianChi platform (\url{https://tianchi.aliyun.com/competition/entrance/531853/introduction}) as one of the series of AI Security Challengers Program.