Robustness Of Saak Transform Against Adversarial Attacks
作者： Suya You , C-C Jay Kuo , Abinaya Manimaran , Thiyagarajan Ramanathan
DOI:
关键词:
摘要: Image classification is vulnerable to adversarial attacks. This work investigates the robustness of Saak transform against attacks towards high performance image classification. We develop a complete system based on multi-stage transform. In domain, clean and images demonstrate different distributions at spectral dimensions. Selection dimensions every stage can be viewed as an automatic denoising process. Motivated by this observation, we carefully design strategies feature extraction, representation that increase robustness. The performances with well-known datasets are demonstrated extensive experimental evaluations.
arxiv.org参考文章(15)
Ian Goodfellow, Samy Bengio, Alexey Kurakin, Adversarial examples in the physical world arXiv: Computer Vision and Pattern Recognition. ,(2016)
Ian Goodfellow, Samy Bengio, Alexey Kurakin, Adversarial Machine Learning at Scale arXiv: Computer Vision and Pattern Recognition. ,(2016)
Nicolas Papernot, Patrick McDaniel, Arunesh Sinha, Michael P. Wellman, Towards the Science of Security and Privacy in Machine Learning arXiv: Cryptography and Security. pp. 1- ,(2016) , 10.1109/EUROSP.2018.00035
Nicolas Papernot, Patrick D. McDaniel, Kathrin Grosse, Praveen Manoharan, Michael Backes, On the (Statistical) Detection of Adversarial Examples arXiv: Cryptography and Security. ,(2017)
Saurabh Shintre, Andrew B. Gardner, Ryan R. Curtin, Reuben Feinman, Detecting Adversarial Samples from Artifacts. arXiv: Machine Learning. ,(2017)
Aran Nayebi, Surya Ganguli, Biologically inspired protection of deep networks from adversarial attacks arXiv: Machine Learning. ,(2017)
Yanjun Qi, Weilin Xu, David Evans, Feature Squeezing: Detecting Adversarial Examples in Deep Neural Networks arXiv: Computer Vision and Pattern Recognition. ,(2017) , 10.14722/NDSS.2018.23198
C. C. Jay Kuo, Ngai-Man Cheung, Sibo Song, Yueru Chen, Defense Against Adversarial Attacks with Saak Transform arXiv: Computer Vision and Pattern Recognition. ,(2018)
Ali Shafahi, Larry S. Davis, John Dickerson, Tom Goldstein, Mahyar Najibi, Zheng Xu, Universal Adversarial Training arXiv: Computer Vision and Pattern Recognition. ,(2018)
Laurens van der Maaten, Alan Yuille, Cihang Xie, Kaiming He, Yuxin Wu, Feature Denoising for Improving Adversarial Robustness arXiv: Computer Vision and Pattern Recognition. ,(2018)