Improving Noise Robustness of Automatic Speech Recognition via Parallel Data and Teacher-student Learning

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216305%3A26230%2F19%3APU134189" target="_blank" >RIV/00216305:26230/19:PU134189 - isvavai.cz</a>

Výsledek na webu

<a href="https://ieeexplore.ieee.org/document/8683422" target="_blank" >https://ieeexplore.ieee.org/document/8683422</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1109/ICASSP.2019.8683422" target="_blank" >10.1109/ICASSP.2019.8683422</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Improving Noise Robustness of Automatic Speech Recognition via Parallel Data and Teacher-student Learning

Popis výsledku v původním jazyce

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.

Název v anglickém jazyce

Improving Noise Robustness of Automatic Speech Recognition via Parallel Data and Teacher-student Learning

Popis výsledku anglicky

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.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

<a href="/cs/project/LQ1602" target="_blank" >LQ1602: IT4Innovations excellence in science</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2019

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Proceedings of ICASSP

ISBN

978-1-5386-4658-8

ISSN

—

e-ISSN

—

Počet stran výsledku

5

Strana od-do

6475-6479

Název nakladatele

IEEE Signal Processing Society

Místo vydání

Brighton

Místo konání akce

Brighton

Datum konání akce

12. 5. 2019

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

000482554006141