Detecting Adversarial Perturbations with Saliency
arXiv:1803.08773
Abstract
In this paper we propose a novel method for detecting adversarial examples by training a binary classifier with both origin data and saliency data. In the case of image classification model, saliency simply explain how the model make decisions by identifying significant pixels for prediction. A model shows wrong classification output always learns wrong features and shows wrong saliency as well. Our approach shows good performance on detecting adversarial perturbations. We quantitatively evaluate generalization ability of the detector, showing that detectors trained with strong adversaries perform well on weak adversaries.
References in corpus (14)
- Adam: A Method for Stochastic Optimization
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems
- Striving for Simplicity: The All Convolutional Net
- Towards Deep Neural Network Architectures Robust to Adversarial Examples
- Understanding Adversarial Training: Increasing Local Stability of Neural Nets through Robust Optimization
- On the (Statistical) Detection of Adversarial Examples
- Grad-CAM: Why did you say that?
- Learning with a Strong Adversary
- Early Methods for Detecting Adversarial Images
- Improving the Robustness of Deep Neural Networks via Stability Training
- The Limitations of Deep Learning in Adversarial Settings
- Robust Convolutional Neural Networks under Adversarial Noise
- A New Method to Visualize Deep Neural Networks