Comment by 0xNotMyAccount

Has anyone tried the same adversarial examples against many different DNNs? I would think these are fairly brittle attacks in reality and only effective with some amount of inside knowledge.