Robust Audio Adversarial Example for a Physical Attack

Hiromu Yakura, Jun Sakuma

We propose a method to generate audio adversarial examples that can attack a state-of-the-art speech recognition model in the physical world. Previous work assumes that generated a…