RSD-GAN: Regularized Sobolev Defense GAN Against Speech-to-Text Adversarial Attacks
This addresses security vulnerabilities in speech-to-text systems, which is critical for applications like voice assistants and transcription services, but it appears incremental as it builds upon existing GAN-based defense methods.
The paper tackles the problem of defending speech-to-text transcription systems against adversarial attacks by introducing a synthesis-based defense algorithm using a Sobolev-based GAN with a novel regularizer, achieving remarkable performance on systems like DeepSpeech, Kaldi, and Lingvo.
This paper introduces a new synthesis-based defense algorithm for counteracting with a varieties of adversarial attacks developed for challenging the performance of the cutting-edge speech-to-text transcription systems. Our algorithm implements a Sobolev-based GAN and proposes a novel regularizer for effectively controlling over the functionality of the entire generative model, particularly the discriminator network during training. Our achieved results upon carrying out numerous experiments on the victim DeepSpeech, Kaldi, and Lingvo speech transcription systems corroborate the remarkable performance of our defense approach against a comprehensive range of targeted and non-targeted adversarial attacks.