arXiv (Cornell University) · 2019 · 136 citations · 24 references
ComputingComputer ScienceComputer SecurityComputer VisionMachine LearningAdversary TechniquesOverfittingAdditive NoiseAdversarial ExamplesAdversarial RobustnessHuman Performance LimitationsRobustness
Over the last few years, the phenomenon of adversarial examples --- maliciously constructed inputs that fool trained machine learning models --- has captured the attention of the research community, especially when the adversary is restricted to small modifications of a correctly handled input. Less surprisingly, image classifiers also lack human-level performance on randomly corrupted images, such as images with additive Gaussian noise. In this paper we provide both empirical and theoretical evidence that these are two manifestations of the same underlying phenomenon, establishing close connections between the adversarial robustness and corruption robustness research programs. This suggests that improving adversarial robustness should go hand in hand with improving performance in the presence of more general and realistic image corruptions. Based on our results we recommend that future adversarial defenses consider evaluating the robustness of their methods to distributional shift with benchmarks such as Imagenet-C.
24
Sergey Zagoruyko, Nikos Komodakis · arXiv (Cornell University) · 2016 · 1.9K citations · Full text
Towards Deep Learning Models Resistant to Adversarial Attacks
Aleksander Mądry, Aleksandar Makelov, Ludwig Schmidt et al. · arXiv (Cornell University) · 2017 · 1.5K citations · Full text
AutoAugment: Learning Augmentation Policies from Data
Ekin D. Cubuk, Barret Zoph, Dandelion Mané et al. · arXiv (Cornell University) · 2018 · 733 citations · Full text
The Brunn-Minkowski inequality in Gauss space
Christer Borell · Inventiones mathematicae · 1975 · 725 citations