Representation Learning for Audio Privacy Preservation using Source Separation and Robust Adversarial Learning
This addresses privacy concerns in smart acoustic monitoring systems, but it is incremental as it combines existing approaches.
The study tackled the problem of preserving speech privacy in smart acoustic monitoring systems by integrating source separation and adversarial representation learning to prevent differentiation between speech and non-speech recordings, resulting in significantly improved privacy preservation compared to using either method alone while maintaining good monitoring performance.
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.