Adversarial Attack
Implementing adversarial attack using TensorFlow 2.0
-
In deepfool/ folder, deepfool attack is written using TensorFlow 2.0 according to DeepFool: a simple and accurate method to fool deep neural networks.
-
In l2_attack/ folder, carlini_wagnerL2 attack is written using TensorFlow 2.0 according to Towards Evaluating the Robustness of Neural Networks.
-
In fgsm/ folder, fast gradient sign method attack is written using TensorFlow 2.0 according to Explaining and Harnessing Adversarial Examples.