Representation Learning for Audio Privacy Preservation using Source Separation and Robust Adversarial Learning

TL;DR

This paper introduces a novel audio privacy preservation method combining source separation and adversarial learning to protect speech data in acoustic monitoring systems while maintaining task performance.

Contribution

The study proposes an integrated approach that enhances speech privacy preservation by combining source separation with adversarial representation learning, a novel combination in this context.

Findings

Significantly improves speech privacy preservation over baseline methods.

Maintains high performance in acoustic monitoring tasks.

Effective in preventing differentiation between speech and non-speech recordings.

Abstract

Privacy preservation has long been a concern in smart acoustic monitoring systems, where speech can be passively recorded along with a target signal in the system's operating environment. In this study, we propose the integration of two commonly used approaches in privacy preservation: source separation and adversarial representation learning. The proposed system learns the latent representation of audio recordings such that it prevents differentiating between speech and non-speech recordings. Initially, the source separation network filters out some of the privacy-sensitive data, and during the adversarial learning process, the system will learn privacy-preserving representation on the filtered signal. We demonstrate the effectiveness of our proposed method by comparing our method against systems without source separation, without adversarial learning, and without both. Overall, our results suggest that the proposed system can significantly improve speech privacy preservation compared to that of using source separation or adversarial learning solely while maintaining good performance in the acoustic monitoring task.