Improving Robustness to Adversarial Examples by Encouraging Discriminative Features | IEEE Conference Publication | IEEE Xplore