TEACHER-STUDENT MIXIT FOR UNSUPERVISED AND SEMI-SUPERVISED SPEECH SEPARATION
(3 minutes introduction)
| Jisi Zhang (University of Sheffield, UK), Cătălin Zorilă (Toshiba, UK), Rama Doddipatla (Toshiba, UK), Jon Barker (University of Sheffield, UK) |
|---|
In this paper, we introduce a novel semi-supervised learning framework for end-to-end speech separation. The proposed method first uses mixtures of unseparated sources and the mixture invariant training (MixIT) criterion to train a teacher model. The teacher model then estimates separated sources that are used to train a student model with standard permutation invariant training (PIT). The student model can be fine-tuned with supervised data, i.e., paired artificial mixtures and clean speech sources, and further improved via model distillation. Experiments with single and multi channel mixtures show that the teacher-student training resolves the over-separation problem observed in the original MixIT method. Further, the semi-supervised performance is comparable to a fully-supervised separation system trained using ten times the amount of supervised data.
