Unsupervised Noise adaptation using Data Simulation
This addresses the domain mismatch issue in speech enhancement for real-world applications, but it is incremental as it builds on existing GAN and adaptation techniques.
The paper tackles the problem of speech enhancement models failing on unseen noises by proposing an unsupervised noise adaptation method using a generative adversarial network to simulate target domain data, achieving a large margin improvement over baselines.
Deep neural network based speech enhancement approaches aim to learn a noisy-to-clean transformation using a supervised learning paradigm. However, such a trained-well transformation is vulnerable to unseen noises that are not included in training set. In this work, we focus on the unsupervised noise adaptation problem in speech enhancement, where the ground truth of target domain data is completely unavailable. Specifically, we propose a generative adversarial network based method to efficiently learn a converse clean-to-noisy transformation using a few minutes of unpaired target domain data. Then this transformation is utilized to generate sufficient simulated data for domain adaptation of the enhancement model. Experimental results show that our method effectively mitigates the domain mismatch between training and test sets, and surpasses the best baseline by a large margin.