Random Spiking and Systematic Evaluation of Defenses Against Adversarial Examples

Huangyi Ge, Sze Yiu Chau, Bruno Ribeiro +1

Image classifiers often suffer from adversarial examples, which are generated by strategically adding a small amount of noise to input images to trick classifiers into misclassific…