Attackar: Attack of the Evolutionary Adversary

Deep neural networks (DNNs) are sensitive to adversarial data in a variety of scenarios, including the black-box scenario, where the attacker is only allowed to query the trained model and receive an output. Existing black-box methods for creating adversarial instances are costly, often using gradient estimation or training a replacement network. This paper introduces \textit{Attackar}, an evolutionary, score-based, black-box attack. Attackar is based on a novel objective function that can be used in gradient-free optimization problems. The attack only requires access to the output logits of the classifier and is thus not affected by gradient masking. No additional information is needed, rendering our method more suitable to real-life situations. We test its performance with three different state-of-the-art models -- Inception-v3, ResNet-50, and VGG-16-BN -- against three benchmark datasets: MNIST, CIFAR10 and ImageNet. Furthermore, we evaluate Attackar's performance on non-differential transformation defenses and state-of-the-art robust models. Our results demonstrate the superior performance of Attackar, both in terms of accuracy score and query efficiency.