adversarial_ml

Adversarial Machine Learning Implementation for existing attacks and defenses:

  1. ZOO: ZOO: Zeroth Order Optimization based Black-box Attacks to Deep Neural Networks without Training Substitute Models. Note: this repository contains CW attacks.
  2. CLEVER: Evaluating the Robustness of Neural Networks: An Extreme Value Theory Approach
  3. Feature Squeezing: Feature Squeezing: Detecting Adversarial Examples in Deep Neural Networks
  4. MagNet: MagNet: a Two-Pronged Defense against Adversarial Examples
  5. PGD: Towards Deep Learning Models Resistant to Adversarial Attacks
  6. Convex adversarial: Provable defenses against adversarial examples via the convex outer adversarial polytope
  7. AutoZoom AutoZOOM: Autoencoder-based Zeroth Order Optimization Method