Sparse Representations Improve Adversarial Robustness of Neural Network Classifiers

Open in new window