Domain Adaptation via Teacher-Student Learning for End-to-End Speech Recognition

January 06, 2020 Β· Declared Dead Β· πŸ› Automatic Speech Recognition & Understanding

πŸ‘» CAUSE OF DEATH: Ghosted
No code link whatsoever

"No code URL or promise found in abstract"

Evidence collected by the PWNC Scanner

Authors Zhong Meng, Jinyu Li, Yashesh Gaur, Yifan Gong arXiv ID 2001.01798 Category eess.AS: Audio & Speech Cross-listed cs.CL, cs.LG, cs.NE, cs.SD Citations 54 Venue Automatic Speech Recognition & Understanding Last Checked 5 months ago
Abstract
Teacher-student (T/S) has shown to be effective for domain adaptation of deep neural network acoustic models in hybrid speech recognition systems. In this work, we extend the T/S learning to large-scale unsupervised domain adaptation of an attention-based end-to-end (E2E) model through two levels of knowledge transfer: teacher's token posteriors as soft labels and one-best predictions as decoder guidance. To further improve T/S learning with the help of ground-truth labels, we propose adaptive T/S (AT/S) learning. Instead of conditionally choosing from either the teacher's soft token posteriors or the one-hot ground-truth label, in AT/S, the student always learns from both the teacher and the ground truth with a pair of adaptive weights assigned to the soft and one-hot labels quantifying the confidence on each of the knowledge sources. The confidence scores are dynamically estimated at each decoder step as a function of the soft and one-hot labels. With 3400 hours parallel close-talk and far-field Microsoft Cortana data for domain adaptation, T/S and AT/S achieve 6.3% and 10.3% relative word error rate improvement over a strong E2E model trained with the same amount of far-field data.
Community shame:
Not yet rated
Community Contributions

Found the code? Know the venue? Think something is wrong? Let us know!

πŸ“œ Similar Papers

In the same crypt β€” Audio & Speech

Died the same way β€” πŸ‘» Ghosted