RemixIT: Continual self-training of speech enhancement models via bootstrapped remixing

Abstract

We present RemixIT, a simple yet effective self-supervised method for training speech enhancement without the need of a single isolated in-domain speech nor a noise waveform. Our approach overcomes limitations of previous methods which make them dependent on clean in-domain target signals and thus, sensitive to any domain mismatch between train and test samples. RemixIT is based on a continuous self-training scheme in which a pre-trained teacher model on out-of-domain data infers estimated pseudo-target signals for in-domain mixtures. Then, by permuting the estimated clean and noise signals and remixing them together, we generate a new set of bootstrapped mixtures and corresponding pseudo-targets which are used to train the student network. Vice-versa, the teacher periodically refines its estimates using the updated parameters of the latest student models. Experimental results on multiple speech enhancement datasets and tasks not only show the superiority of our method over prior approaches but also showcase that RemixIT can be combined with any separation model as well as be applied towards any semi-supervised and unsupervised domain adaptation task. Our analysis, paired with empirical evidence, sheds light on the inside functioning of our self-training scheme wherein the student model keeps obtaining better performance while observing severely degraded pseudo-targets.

To appear in IEEE Journal of Selected Topics in Signal Processing

References in corpus (6)

• TERA: Self-Supervised Learning of Transformer Encoder Representation for Speech
• Sudo rm -rf: Efficient Networks for Universal Audio Source Separation
• Two-Step Sound Source Separation: Training on Learned Latent Targets
• Unsupervised Speech Enhancement Based on Multichannel NMF-Informed Beamforming for Noise-Robust Automatic Speech Recognition
• Compute and memory efficient universal sound source separation
• Separate but Together: Unsupervised Federated Learning for Speech Enhancement from Non-IID Data

Cited by in corpus (1)

• The CHiME-7 UDASE task: Unsupervised domain adaptation for conversational speech enhancement