Houdini: Fooling Deep Structured Visual and Speech Recognition Models with Adversarial Examples

Open in new window