Audio Adversarial Examples: Attacks Using Vocal Masks

2021/02/04 by Tay, Kai Yuan, Ng, Lynnette, Chua, Wei Han +3
#Artificial Intelligence (cs.AI) #Audio and Speech Processing (eess.AS) #FOS: Computer and information sciences #FOS: Electrical engineering #Sound (cs.SD) #electronic engineering #information engineering

paper · doi:10.48550/arxiv.2102.02417

Abstract

We construct audio adversarial examples on automatic Speech-To-Text systems . Given any audio waveform, we produce an another by overlaying an audio vocal mask generated from the original audio. We apply our audio adversarial attack to five SOTA STT systems: DeepSpeech, Julius, Kaldi, wav2letter@anywhere and CMUSphinx. In addition, we engaged human annotators to transcribe the adversarial audio. Our experiments show that these adversarial examples fool State-Of-The-Art Speech-To-Text systems, yet humans are able to consistently pick out the speech. The feasibility of this attack introduces a new domain to study machine and human perception of speech.

Related

• Audio Adversarial Examples: Targeted Attacks on Speech-to-Text
• Isolated and Ensemble Audio Preprocessing Methods for Detecting Adversarial Examples against Automatic Speech Recognition
• Did you hear that? Adversarial Examples Against Automatic Speech Recognition
• SirenAttack: Generating Adversarial Audio for End-to-End Acoustic Systems
• Imperceptible, Robust, and Targeted Adversarial Examples for Automatic Speech Recognition