Barrage of Random Transforms for Adversarially Robust Defense
Authors:
- 1 June 2019
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 6521-6530
Abstract
Defenses against adversarial examples, when using the ImageNet dataset, are historically easy to defeat. The common understanding is that a combination of simple image transformations and other various defenses are insufficient to provide the necessary protection when the obfuscated gradient is taken into account. In this paper, we explore the idea of stochastically combining a large number of individually weak defenses into a single barrage of randomized transformations to build a strong defense against adversarial attacks. We show that, even after accounting for obfuscated gradients, the Barrage of Random Transforms (BaRT) is a resilient defense against even the most difficult attacks, such as PGD. BaRT achieves up to a 24x improvement in accuracy compared to previous work, and has even extended effectiveness out to a previously untested maximum adversarial perturbation of ε=32.Keywords
This publication has 16 references indexed in Scilit:
- Adversarial Examples Detection in Deep Networks with Convolutional Filter Statistics2017
- Exploring the space of adversarial images2016
- ImageNet Large Scale Visual Recognition ChallengeInternational Journal of Computer Vision, 2015
- Security Evaluation of Pattern Classifiers under AttackIEEE Transactions on Knowledge and Data Engineering, 2014
- Fast nonlocal filtering applied to electron cryomicroscopy2008
- Seam carving for content-aware image resizing2007
- Fast median and bilateral filtering2006
- Affine Invariant Flows in the Beltrami FrameworkJournal of Mathematical Imaging and Vision, 2004
- Adaptive wavelet thresholding for image denoising and compressionIEEE Transactions on Image Processing, 2000
- Contrast Limited Adaptive Histogram Equalization1994