Learning to Separate Clusters of Adversarial Representations for Robust Adversarial Detection

Abstract

Although deep neural networks have shown promising performances on various tasks, they are susceptible to incorrect predictions induced by imperceptibly small perturbations in inputs. A large number of previous works proposed to detect adversarial attacks. Yet, most of them cannot effectively detect them against adaptive whitebox attacks where an adversary has the knowledge of the model and the defense method. In this paper, we propose a new probabilistic adversarial detector motivated by a recently introduced non-robust feature. We consider the non-robust features as a common property of adversarial examples, and we deduce it is possible to find a cluster in representation space corresponding to the property. This idea leads us to probability estimate distribution of adversarial representations in a separate cluster, and leverage the distribution for a likelihood based adversarial detector.

References in corpus (5)

• PyTorch: An Imperative Style, High-Performance Deep Learning Library
• Certified Adversarial Robustness via Randomized Smoothing
• Countering Adversarial Images using Input Transformations
• An approach to reachability analysis for feed-forward ReLU neural networks
• Output Range Analysis for Deep Neural Networks