Skip to yearly menu bar Skip to main content


Poster

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

Moustapha Cisse · Yossi Adi · Natalia Neverova · Joseph Keshet

Pacific Ballroom #137

Keywords: [ Deep Learning ] [ Adversarial Networks ] [ Privacy, Anonymity, and Security ]


Abstract:

Generating adversarial examples is a critical step for evaluating and improving the robustness of learning machines. So far, most existing methods only work for classification and are not designed to alter the true performance measure of the problem at hand. We introduce a novel flexible approach named Houdini for generating adversarial examples specifically tailored for the final performance measure of the task considered, be it combinatorial and non-decomposable. We successfully apply Houdini to a range of applications such as speech recognition, pose estimation and semantic segmentation. In all cases, the attacks based on Houdini achieve higher success rate than those based on the traditional surrogates used to train the models while using a less perceptible adversarial perturbation.

Live content is unavailable. Log in and register to view live content