HiddenSpeaker: Generate Imperceptible Unlearnable Audios for Speaker Verification System

TL;DR

HiddenSpeaker introduces a method to embed imperceptible perturbations in speech data, making it unlearnable for speaker verification systems while maintaining human perceptual quality, thus protecting privacy and preventing unauthorized model training.

Contribution

The paper proposes HiddenSpeaker, a novel framework that generates imperceptible, unlearnable audio samples using a simplified error-minimizing method and a hybrid perceptual optimization, enhancing privacy protection.

Findings

Successfully deceives state-of-the-art speaker verification models

Perturbations are highly imperceptible to human listeners

Demonstrates strong transferability across different models

Abstract

In recent years, the remarkable advancements in deep neural networks have brought tremendous convenience. However, the training process of a highly effective model necessitates a substantial quantity of samples, which brings huge potential threats, like unauthorized exploitation with privacy leakage. In response, we propose a framework named HiddenSpeaker, embedding imperceptible perturbations within the training speech samples and rendering them unlearnable for deep-learning-based speaker verification systems that employ large-scale speakers for efficient training. The HiddenSpeaker utilizes a simplified error-minimizing method named Single-Level Error-Minimizing (SLEM) to generate specific and effective perturbations. Additionally, a hybrid objective function is employed for human perceptual optimization, ensuring the perturbation is indistinguishable from human listeners. We conduct extensive experiments on multiple state-of-the-art (SOTA) models in the speaker verification domain to evaluate HiddenSpeaker. Our results demonstrate that HiddenSpeaker not only deceives the model with unlearnable samples but also enhances the imperceptibility of the perturbations, showcasing strong transferability across different models.