Global Adversarial Attacks for Assessing Deep Learning Robustness

Abstract

It has been shown that deep neural networks (DNNs) may be vulnerable to adversarial attacks, raising the concern on their robustness particularly for safety-critical applications. Recognizing the local nature and limitations of existing adversarial attacks, we present a new type of global adversarial attacks for assessing global DNN robustness. More specifically, we propose a novel concept of global adversarial example pairs in which each pair of two examples are close to each other but have different class labels predicted by the DNN. We further propose two families of global attack methods and show that our methods are able to generate diverse and intriguing adversarial example pairs at locations far from the training or testing data. Moreover, we demonstrate that DNNs hardened using the strong projected gradient descent (PGD) based (local) adversarial training are vulnerable to the proposed global adversarial example pairs, suggesting that global robustness must be considered while training robust deep learning networks.Preprint. Under review.

Exploiting Self-Supervised and Semi-Supervised Learning for Facial Landmark Tracking with Unlabeled Data

“…Once the GEV distribution p y * is fitted, we adopt a Markov Chain Monte Carlo (MCMC) method (Hu et al, 2019) to sample from it. First, we collect samples from a proposal distribution, a surrogate distribution that is easy to sample.…”

“…The size of the validation data is 10 in the topology and expression tasks and 30 in the chemical design task, respectively. The hyperparameters of MCMC sampling for the GEV distribution are set as: the number of warm-up rounds is 10 and the parameters of the proposed distribution are λ 0 = 0.01, λ m = 0.04 (Hu et al, 2019).…”