Certified Patch Robustness via Smoothed Vision Transformers
Authors:
- Hadi SalmanHSHadi Salman
- Saachi JainSJSaachi Jain
- Eric WongEWEric Wong
- Aleksander MadryAMAleksander Madry
- 1 June 2022
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 15116-15126
Abstract
Certified patch defenses can guarantee robustness of an image classifier to arbitrary changes within a bounded contiguous region. But, currently, this robustness comes at a cost of degraded standard accuracies and slower inference times. We demonstrate how using vision transformers enables significantly better certified patch robustness that is also more computationally efficient and does not incur a substantial drop in standard accuracy. These improvements stem from the inherent ability of the vision transformer to gracefully handle largely masked images. 1 1 Our code is available at https://github.com/MadryLab/smoothed-vit..Keywords
All Related Versions
Funding Information
- NSF (CCF-1553428,CNS-1815221)
- Defense Advanced Research Projects Agency (DARPA) (HR001120C0015)
This publication has 20 references indexed in Scilit:
- ShapeShifter: Robust Physical Adversarial Attack on Faster R-CNN Object Detector2019
- Local Gradients Smoothing: Defense Against Localized Adversarial Attacks2019
- Adversarial Attacks on Face Detectors Using Neural Net Based Constrained Optimization2018
- On Visible Adversarial Perturbations & Digital Watermarking2018
- Formal Verification of Piece-Wise Linear Feed-Forward Neural Networks2017
- Reluplex: An Efficient SMT Solver for Verifying Deep Neural Networks2017
- Accessorize to a Crime2016
- Deep Residual Learning for Image Recognition2016
- ImageNet Large Scale Visual Recognition ChallengeInternational Journal of Computer Vision, 2015
- ImageNet: A large-scale hierarchical image database2009