Adversarial examples training input placeholders

[JRMeyer]

>>> dijksterhuis
[May 26, 2019, 6:09pm]

I'm a PhD student researching audio adversarial examples. So I'll be
using DeepSpeech to generate some attacks for ASR (also looking at
verification models).

There are existing attacks against DeepSpeech, but only work against
0.1. Github issues for those attacks: slash

github.com/mozilla/DeepSpeech

.onebox-avatar width='96'

#### Issue: Did you hear that? Adversarial Examples Against Automatic Speech Recognition

opened by kdavis-mozilla on
2018-01-17

{.content style='white-space: pre-wrap;'}<br/> Address adversarial attack of Did you hear that? Adversarial Examples Against Automatic Speech Recognition<br/>

slash

github.com/mozilla/DeepSpeech

.onebox-avatar width='96'

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

opened by kdavis-mozilla on
2018-01-17

{.content style='white-space: pre-wrap;'}<br/> Address adversarial attack of Audio Adversarial Examples: Targeted Attacks on Speech-to-Text<br/>

Would devs (or anyone) be interested in opening up the source version
with an option argument to create_inference_model to modify the input
placeholders for optimising our attacks? N.B. This is not regarding
training DeepSpeech, rather optimising a specific attack method against
a pre-trained model checkpoint.

This would be particularly useful in the 0.5 release, as it looks like
the mfcc & windowing will be handled natively in DeepSpeech and won't
require any additional code for attacks (current blocker).

Currently building own fork to do this, but figured I'd ask and see if
there's any interest in doing this on the project side.

Thanks, Dx

[This is an archived TTS discussion thread from discourse.mozilla.org/t/adversarial-examples-training-input-placeholders]

[JRMeyer]

>>> dijksterhuis
[May 26, 2019, 6:09pm]

I'm a PhD student researching audio adversarial examples. So I'll be
using DeepSpeech to generate some attacks for ASR (also looking at
verification models).

There are existing attacks against DeepSpeech, but only work against
0.1. Github issues for those attacks: slash

github.com/mozilla/DeepSpeech

.onebox-avatar width='96'

#### Issue: Did you hear that? Adversarial Examples Against Automatic Speech Recognition

opened by kdavis-mozilla on
2018-01-17

{.content style='white-space: pre-wrap;'}<br/> Address adversarial attack of Did you hear that? Adversarial Examples Against Automatic Speech Recognition<br/>

slash

github.com/mozilla/DeepSpeech

.onebox-avatar width='96'

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

opened by kdavis-mozilla on
2018-01-17

{.content style='white-space: pre-wrap;'}<br/> Address adversarial attack of Audio Adversarial Examples: Targeted Attacks on Speech-to-Text<br/>

Would devs (or anyone) be interested in opening up the source version
with an option argument to create_inference_model to modify the input
placeholders for optimising our attacks? N.B. This is not regarding
training DeepSpeech, rather optimising a specific attack method against
a pre-trained model checkpoint.

This would be particularly useful in the 0.5 release, as it looks like
the mfcc & windowing will be handled natively in DeepSpeech and won't
require any additional code for attacks (current blocker).

Currently building own fork to do this, but figured I'd ask and see if
there's any interest in doing this on the project side.

Thanks, Dx

[This is an archived TTS discussion thread from discourse.mozilla.org/t/adversarial-examples-training-input-placeholders]