SparseFool: a few pixels make a big difference
Preprint
Abstract
Deep Neural Networks have achieved extraordinary results on image classification tasks, but have been shown to be vulnerable to attacks with carefully crafted perturbations of the input data. Although most attacks usually change values of many image's pixels, it has been shown that deep networks are also vulnerable to sparse alterations of the input. However, no computationally efficient method has been proposed to compute sparse perturbations. In this paper, we exploit the low mean curvature of the decision boundary, and propose SparseFool, a geometry inspired sparse attack that controls the sparsity of the perturbations. Extensive evaluations show that our approach computes sparse perturbations very fast, and scales efficiently to high dimensional data. We further analyze the transferability and the visual effects of the perturbations, and show the existence of shared semantic information across the images and the networks. Finally, we show that adversarial training can only slightly improve the robustness against sparse additive perturbations computed with SparseFool.Keywords
All Related Versions
This publication has 12 references indexed in Scilit:
- Facial attributes: Accuracy and adversarial robustnessPattern Recognition Letters, 2019
- One Pixel Attack for Fooling Deep Neural NetworksIEEE Transactions on Evolutionary Computation, 2019
- Audio Adversarial Examples: Targeted Attacks on Speech-to-Text2018
- The Robustness of Deep Networks: A Geometrical PerspectiveIEEE Signal Processing Magazine, 2017
- Random Coordinate Descent Methods for $\ell_{0}$ Regularized Convex OptimizationIEEE Transactions on Automatic Control, 2015
- Sparse Packetized Predictive Control for Networked Control Over Erasure ChannelsIEEE Transactions on Automatic Control, 2013
- Description of the Minimizers of Least Squares Regularized with $\ell_0$-norm. Uniqueness of the Global MinimizerSIAM Journal on Imaging Sciences, 2013
- Iterative Thresholding for Sparse ApproximationsJournal of Fourier Analysis and Applications, 2008
- Decoding by Linear ProgrammingIEEE Transactions on Information Theory, 2005
- Sparse Approximate Solutions to Linear SystemsSIAM Journal on Computing, 1995