Audio Adversarial Examples: Targeted Attacks on Speech-to-Text

Code

• github.com/carlini/audio_adversarial_examples
  tf★ 0
• github.com/tom-doerr/audio_adversarial_examples
  tf★ 0
• github.com/etapiahe/adversarial_audio_attack
  pytorch★ 0
• github.com/vishKurama/audio-adversarial-attack
  tf★ 0

Abstract

We construct targeted audio adversarial examples on automatic speech recognition. Given any audio waveform, we can produce another that is over 99.9% similar, but transcribes as any phrase we choose (recognizing up to 50 characters per second of audio). We apply our white-box iterative optimization-based attack to Mozilla's implementation DeepSpeech end-to-end, and show it has a 100% success rate. The feasibility of this attack introduce a new domain to study adversarial examples.

Tasks

Reproductions