Adversarial Generative NMF for Single Channel Source Separation
This addresses source separation for audio/image processing, but is incremental as it adapts an existing adversarial learning approach to NMF.
The paper tackles single-channel source separation by applying adversarial learning to train non-negative matrix factorization (NMF) bases, showing in numerical experiments on image and audio data that this leads to clear improvements in reconstructed signals, especially with limited supervision.
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 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.