Defense against Adversarial Attacks Using High-Level Representation Guided Denoiser

Abstract

Neural networks are vulnerable to adversarial examples, which poses a threat to their application in security sensitive systems. We propose high-level representation guided denoiser (HGD) as a defense for image classification. Standard denoiser suffers from the error amplification effect, in which small residual adversarial noise is progressively amplified and leads to wrong classifications. HGD overcomes this problem by using a loss function defined as the difference between the target model's outputs activated by the clean image and denoised image. Compared with ensemble adversarial training which is the state-of-the-art defending method on large images, HGD has three advantages. First, with HGD as a defense, the target model is more robust to either white-box or black-box adversarial attacks. Second, HGD can be trained on a small subset of the images and generalizes well to other images and unseen classes. Third, HGD can be transferred to defend models other than the one guiding it. In NIPS competition on defense against adversarial attacks, our HGD solution won the first place and outperformed other models by a large margin.

References in corpus (11)

• Adam: A Method for Stochastic Optimization
• Distilling the Knowledge in a Neural Network
• Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising
• Ensemble Adversarial Training: Attacks and Defenses
• Transferability in Machine Learning: from Phenomena to Black-Box Attacks using Adversarial Samples
• Towards Deep Neural Network Architectures Robust to Adversarial Examples
• Aggregated Residual Transformations for Deep Neural Networks
• Keeping the Bad Guys Out: Protecting and Vaccinating Deep Learning with JPEG Compression
• MagNet: a Two-Pronged Defense against Adversarial Examples
• Biologically inspired protection of deep networks from adversarial attacks
• Assessing Threat of Adversarial Examples on Deep Neural Networks

Cited by in corpus (16)

• Adversarial Risk and the Dangers of Evaluating Against Weak Attacks
• Divide, Denoise, and Defend against Adversarial Attacks
• Towards Understanding Adversarial Examples Systematically: Exploring Data Size, Task and Model Factors
• Playing the Game of Universal Adversarial Perturbations
• Detecting Adversarial Examples via Neural Fingerprinting
• Detecting Overfitting via Adversarial Examples
• Feature Prioritization and Regularization Improve Standard Accuracy and Adversarial Robustness
• Strength in Numbers: Trading-off Robustness and Computation via Adversarially-Trained Ensembles
• Generating Out of Distribution Adversarial Attack using Latent Space Poisoning
• A Self-supervised Approach for Adversarial Robustness
• Mitigating Gradient-based Adversarial Attacks via Denoising and Compression
• Once a MAN: Towards Multi-Target Attack via Learning Multi-Target Adversarial Network Once
• Training Artificial Neural Networks by Generalized Likelihood Ratio Method: Exploring Brain-like Learning to Improve Robustness
• Defective Convolutional Networks
• A Computationally Efficient Method for Defending Adversarial Deep Learning Attacks
• A Hierarchical Feature Constraint to Camouflage Medical Adversarial Attacks