Heat and Blur: An Effective and Fast Defense Against Adversarial Examples
arXiv:2003.07573
Abstract
The growing incorporation of artificial neural networks (NNs) into many fields, and especially into life-critical systems, is restrained by their vulnerability to adversarial examples (AEs). Some existing defense methods can increase NNs' robustness, but they often require special architecture or training procedures and are irrelevant to already trained models. In this paper, we propose a simple defense that combines feature visualization with input modification, and can, therefore, be applicable to various pre-trained networks. By reviewing several interpretability methods, we gain new insights regarding the influence of AEs on NNs' computation. Based on that, we hypothesize that information about the "true" object is preserved within the NN's activity, even when the input is adversarial, and present a feature visualization version that can extract that information in the form of relevance heatmaps. We then use these heatmaps as a basis for our defense, in which the adversarial effects are corrupted by massive blurring. We also provide a new evaluation metric that can capture the effects of both attacks and defenses more thoroughly and descriptively, and demonstrate the effectiveness of the defense and the utility of the suggested evaluation measurement with VGG19 results on the ImageNet dataset.
arXiv commentSubmitted to IJCAI 2020 conference
References in corpus (15)
- Explaining and Harnessing Adversarial Examples
- Intriguing properties of neural networks
- Feature Squeezing: Detecting Adversarial Examples in Deep Neural Networks
- Explaining NonLinear Classification Decisions with Deep Taylor Decomposition
- Adversarial Examples Are Not Bugs, They Are Features
- Spatially Transformed Adversarial Examples
- The (Un)reliability of saliency methods
- A Boundary Tilting Persepective on the Phenomenon of Adversarial Examples
- Adversarial Examples Are a Natural Consequence of Test Error in Noise
- Towards Interpretable Deep Neural Networks by Leveraging Adversarial Examples
- Interpreting Adversarial Examples by Activation Promotion and Suppression
- Explaining Convolutional Neural Networks using Softmax Gradient Layer-wise Relevance Propagation
- Saliency Methods for Explaining Adversarial Attacks
- Interpretable Deep Learning under Fire
- Concise Explanations of Neural Networks using Adversarial Training