U-DREAM: Unsupervised Dereverberation guided by a Reverberation Model
This addresses the challenge of obtaining paired data for dereverberation, offering a more practical solution for audio processing applications, though it is incremental in its approach.
The paper tackles the problem of dereverberation without requiring paired dry and reverberant data by proposing a sequential learning strategy guided by a reverberation matching loss. The most data-efficient variant uses only 100 labelled samples to outperform an unsupervised baseline, showing effectiveness in low-resource scenarios.
This paper explores the outcome of training state-ofthe-art dereverberation models with supervision settings ranging from weakly-supervised to fully unsupervised, relying solely on reverberant signals and an acoustic model for training. Most of the existing deep learning approaches typically require paired dry and reverberant data, which are difficult to obtain in practice. We develop instead a sequential learning strategy motivated by a bayesian formulation of the dereverberation problem, wherein acoustic parameters and dry signals are estimated from reverberant inputs using deep neural networks, guided by a reverberation matching loss. Our most data-efficient variant requires only 100 reverberation-parameter-labelled samples to outperform an unsupervised baseline, demonstrating the effectiveness and practicality of the proposed method in low-resource scenarios.