I agree that this and the 'adversarial road signs' work indicate a growing interest in security attacks on ML. https://medium.com/self-driving-cars/adversarial

I agree that this and the 'adversarial road signs' work indicate a growing interest in security attacks on ML. https://medium.com/self-driving-cars/adversarial-traffic-signs-fd16b7171906 https://medium.com/self-driving-cars/adversarial-traffic-sig... However, in this case, it's not just machine learning. It will affect non-ML audio-processing systems just as much.