Picture for Joseph Szurley

Joseph Szurley

Mitigating Closed-model Adversarial Examples with Bayesian Neural Modeling for Enhanced End-to-End Speech Recognition

Feb 17, 2022
Figure 1 for Mitigating Closed-model Adversarial Examples with Bayesian Neural Modeling for Enhanced End-to-End Speech Recognition
Figure 2 for Mitigating Closed-model Adversarial Examples with Bayesian Neural Modeling for Enhanced End-to-End Speech Recognition
Figure 3 for Mitigating Closed-model Adversarial Examples with Bayesian Neural Modeling for Enhanced End-to-End Speech Recognition
Figure 4 for Mitigating Closed-model Adversarial Examples with Bayesian Neural Modeling for Enhanced End-to-End Speech Recognition
Viaarxiv icon

Adversarial Music: Real World Audio Adversary Against Wake-word Detection System

Dec 06, 2019
Figure 1 for Adversarial Music: Real World Audio Adversary Against Wake-word Detection System
Figure 2 for Adversarial Music: Real World Audio Adversary Against Wake-word Detection System
Figure 3 for Adversarial Music: Real World Audio Adversary Against Wake-word Detection System
Figure 4 for Adversarial Music: Real World Audio Adversary Against Wake-word Detection System
Viaarxiv icon