Blind Source Separation of Single-Channel Mixtures via Multi-Encoder Autoencoders

TL;DR

This paper introduces a novel multi-encoder autoencoder approach for blind source separation of single-channel non-linear mixtures, demonstrating effective source extraction in both synthetic and real-world biosignal data.

Contribution

It proposes a new method utilizing multi-encoder autoencoders with encoding masking and sparse mixing loss for improved source separation in challenging single-channel non-linear mixtures.

Findings

Successful separation of sources in toy dataset demonstrating feature subspace specialization

Effective extraction of respiration signals from biosignals in sleep study data

Demonstrated robustness of the method on real-world biosignal recordings

Abstract

The task of blind source separation (BSS) involves separating sources from a mixture without prior knowledge of the sources or the mixing system. Single-channel mixtures and non-linear mixtures are a particularly challenging problem in BSS. In this paper, we propose a novel method for addressing BSS with single-channel non-linear mixtures by leveraging the natural feature subspace specialization ability of multi-encoder autoencoders. During the training phase, our method unmixes the input into the separate encoding spaces of the multi-encoder network and then remixes these representations within the decoder for a reconstruction of the input. Then to perform source inference, we introduce a novel encoding masking technique whereby masking out all but one of the encodings enables the decoder to estimate a source signal. To this end, we also introduce a sparse mixing loss that encourages sparse remixing of source encodings throughout the decoder and a so-called zero reconstruction loss on the decoder for coherent source estimations. To analyze and evaluate our method, we conduct experiments on a toy dataset, designed to demonstrate this property of feature subspace specialization, and with real-world biosignal recordings from a polysomnography sleep study for extracting respiration from electrocardiogram and photoplethysmography signals.