NIPS 2017: Non-targeted Adversarial Attack

Imperceptibly transform images in ways that fool classification models

Actions