9 papers
Towards Robust Speech-to-Text Adversarial Attack
Mohammad Esmaeilpour, Patrick Cardinal, Alessandro Lameiras Koerich
This paper introduces a novel adversarial algorithm for attacking the state-of-the-art speech-to-text systems, namely DeepSpeech, Kaldi, and Lingvo. Our approach is based on develo…
Multi-Discriminator Sobolev Defense-GAN Against Adversarial Attacks for End-to-End Speech Systems
Mohammad Esmaeilpour, Patrick Cardinal, Alessandro Lameiras Koerich
This paper introduces a defense approach against end-to-end adversarial attacks developed for cutting-edge speech-to-text systems. The proposed defense algorithm has four major ste…
Conditioning Trick for Training Stable GANs
Mohammad Esmaeilpour, Raymel Alfonso Sallo, Olivier St-Georges +2
In this paper we propose a conditioning trick, called difference departure from normality, applied on the generator network in response to instability issues during GAN training. W…
Class-Conditional Defense GAN Against End-to-End Speech Attacks
Mohammad Esmaeilpour, Patrick Cardinal, Alessandro Lameiras Koerich
In this paper we propose a novel defense approach against end-to-end adversarial attacks developed to fool advanced speech-to-text systems such as DeepSpeech and Lingvo. Unlike con…
Improving Stability of LS-GANs for Audio and Speech Signals
Mohammad Esmaeilpour, Raymel Alfonso Sallo, Olivier St-Georges +2
In this paper we address the instability issue of generative adversarial network (GAN) by proposing a new similarity metric in unitary space of Schur decomposition for 2D represent…
Adversarially Training for Audio Classifiers
Raymel Alfonso Sallo, Mohammad Esmaeilpour, Patrick Cardinal
In this paper, we investigate the potential effect of the adversarially training on the robustness of six advanced deep neural networks against a variety of targeted and non-target…