Publication Details
Improving Noise Robustness of Automatic Speech Recognition via Parallel Data and Teacher-student Learning
Wu Minhua (AmazonCom)
Raju Anirudh (AmazonCom)
Parthasarathi Sree Hari Krishnan (AmazonCom)
Kumatani Kenichi (AmazonCom)
Sundaram Shiva (AmazonCom)
Maas Roland (AmazonCom)
Hoffmeister Björn (AmazonCom)
automatic speech recognition, noise robustness, teacher-student training, domain adaptation
For real-world speech recognition applications, noise robustness is still a challenge. In this work, we adopt the teacherstudent (T/S) learning technique using a parallel clean and noisy corpus for improving automatic speech recognition (ASR) performance under multimedia noise. On top of that, we apply a logits selection method which only preserves the k highest values to prevent wrong emphasis of knowledge from the teacher and to reduce bandwidth needed for transferring data. We incorporate up to 8000 hours of untranscribed data for training and present our results on sequence trained models apart from cross entropy trained ones. The best sequence trained student model yields relative word error rate (WER) reductions of approximately 10.1%, 28.7% and 19.6% on our clean, simulated noisy and real test sets respectively comparing to a sequence trained teacher.
@INPROCEEDINGS{FITPUB12098,
author = "Ladislav Mo\v{s}ner and Minhua Wu and Anirudh Raju and Krishnan Hari Sree Parthasarathi and Kenichi Kumatani and Shiva Sundaram and Roland Maas and Bj{\"{o}}rn Hoffmeister",
title = "Improving Noise Robustness of Automatic Speech Recognition via Parallel Data and Teacher-student Learning",
pages = "6475--6479",
booktitle = "Proceedings of ICASSP",
year = 2019,
location = "Brighton, GB",
publisher = "IEEE Signal Processing Society",
ISBN = "978-1-5386-4658-8",
doi = "10.1109/ICASSP.2019.8683422",
language = "english",
url = "https://www.fit.vut.cz/research/publication/12098"
}