Generating Minimal Adversarial Perturbations with Integrated Adaptive Gradients

Abstract

Deep neural networks are easily fooled high confidence predictions for adversarial samples

arXiv comment8 pages, 7 figures. arXiv admin note: text overlap with arXiv:1902.01220 The formula in Algorithm 1 lacks important representations

References in corpus (3)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Explaining and Harnessing Adversarial Examples
• Feature Squeezing: Detecting Adversarial Examples in Deep Neural Networks