Abstract: In this paper, we propose a method for defending against audio adversarial examples that operates by applying audio style transfer learning. The proposed method has the effect of maintaining ...