The vulnerability of deep networks to adversarial attacks is a central problem for deep learning from the perspective of both cognition and security. The current most successful defense method is to train a classifier using adversarial images created during learning. Another defense approach involves transformation or purification of the original input to remove adversarial signals before the image is classified. We focus on defending naturally-trained classifiers using Markov Chain Monte Carlo (MCMC) sampling with an Energy-Based Model (EBM) for adversarial purification. In contrast to adversarial training, our approach is intended to secure pre-existing and highly vulnerable classifiers. The memoryless behavior of long-run MCMC sampling will eventually remove adversarial signals, while metastable behavior preserves consistent appearance of MCMC samples after many steps to allow accurate long-run prediction. Balancing these factors can lead to effective purification and robust classification. We evaluate adversarial defense with an EBM using the strongest known attacks against purification. Our contributions are 1) an improved method for training EBM's with realistic long-run MCMC samples, 2) an Expectation-Over-Transformation (EOT) defense that resolves theoretical ambiguities for stochastic defenses and from which the EOT attack naturally follows, and 3) state-of-the-art adversarial defense for naturally-trained classifiers and competitive defense compared to adversarially-trained classifiers on Cifar-10, SVHN, and Cifar-100. Code and pre-trained models are available at https://github.com/point0bar1/ebm-defense.
翻译:从认知和安全的角度来看,深层次网络对对抗性攻击的脆弱性是深层次学习的中心问题。目前最成功的国防方法是使用学习期间产生的对抗性图像对分类人员进行培训。另一种防御方法是改造或净化原始输入,以便在图像分类之前去除对抗性信号。我们的重点是利用以能源为基础的模型对以能源为基础的蒙特卡洛(MCMCM)进行抽样来保护自然训练的分类人员,以对抗性净化。与对抗性培训相比,我们的方法旨在保护原有的和高度脆弱的分类人员。长期MCMC取样的无记忆行为最终将消除对抗性信号,而元化行为则在采取许多步骤进行准确的长期预测后保持了MCMC样本的一贯外观。平衡这些因素可导致有效的净化和稳健的分类。我们利用已知最强的对净化的以能源为基础的模型来评估对抗性防御性防御性辩护。我们的贡献是:(1)用现实的长效的MC样品来培训EBMM公司;(2) 期待性-彻底的国防(EOT)防御(E-OT)行为最终消除对抗性信号信号,而使S-RO-RO-RO-I-I-S-S-S-S-S-IAR-S-S-S-IAR-S-S-S-IAR-S-S-IAR-S-S-S-S-S-S-IAR-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-IAR-IAR-S-S-S-S-S-S-S-S-IAR-IAR-S-S-S-S-S-S-S-S-IAR-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-I