Is there any example code for attacking the speech recognition models (DeepSpeech/Lingvo)?

[xinghua-qu]

I see the estimator provides some speech recognition models (e.g., DeepSpeech and Lingvo). May I ask is there any example code for attacking those models?

[beat-buesser]

Hi @QuXinghuaNTU Thank you very much for your interest in ART!

Currently we have multiple notebook related to ART and audio/ASR:

• adversarial_audio_examples.ipynb
• asr_deepspeech_examples.ipynb

Btw, ART 1.7 also provides an estimator for Espresso.

[xinghua-qu]

Thanks for the quick reply. That greatly helps:)