Maximum Discrepancy Generative Regularization and Non-Negative Matrix Factorization for Single Channel Source Separation

TL;DR

This paper introduces a novel adversarial regularization approach called Maximum Discrepancy Generative Regularization, applied to Non-negative Matrix Factorization for single channel source separation, improving signal reconstruction especially with limited supervision.

Contribution

It presents a new adversarial training method for NMF bases, enhancing source separation performance in low-supervision scenarios.

Findings

Improved signal reconstruction in image and audio separation

Effective with limited supervision data

Demonstrated superiority over traditional NMF methods

Abstract

The idea of adversarial learning of regularization functionals has recently been introduced in the wider context of inverse problems. The intuition behind this method is the realization that it is not only necessary to learn the basic features that make up a class of signals one wants to represent, but also, or even more so, which features to avoid in the representation. In this paper, we will apply this approach to the training of generative models, leading to what we call Maximum Discrepancy Generative Regularization. In particular, we apply this to problem of source separation by means of Non-negative Matrix Factorization (NMF) and present a new method for the adversarial training of NMF bases. We show in numerical experiments, both for image and audio separation, that this leads to a clear improvement of the reconstructed signals, in particular in the case where little or no strong supervision data is available.