Accent neutralization for speech recognition of non-native speakers

Kacper Radzikowski; Mateusz Forc; Le Wang; Osamu Yoshie; Robert Nowak

doi:10.1145/3366030.3366083

Accent neutralization for speech recognition of non-native speakers

Kacper Radzikowski, Mateusz Forc, Le Wang, Osamu Yoshie, Robert Nowak

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

1 Citation (Scopus)

Abstract

These days, automatic speech recognition (ASR) systems achieve higher and higher accuracy rates. The score drops significantly, in case when the ASR system is being used with a non-native speaker of the language to be recognized. The main reason is specific pronunciation and accent features. A limited volume of labeled nonnative speech datasets makes it difficult to train new ASR systems for non-native speakers. In our research,we tried tackling the problem and its influence on the accuracy of ASR systems, using the style transfer methodology. We designed a pipeline for modifying the speech of a non-native speaker, so that it resembles the native speech to a higher extent. Our methodology can be used as a wrapper for any existing ASR system, which reduces the necessity of training new algorithms for non-native speech. The modification can be thus performed before passing the data forward to the speech recognition system itself.

Original language	English
Title of host publication	21st International Conference on Information Integration and Web-Based Applications and Services, iiWAS 2019 - Proceedings
Editors	Maria Indrawan-Santiago, Eric Pardede, Ivan Luiz Salvadori, Matthias Steinbauer, Ismail Khalil, Gabriele Anderst-Kotsis
Publisher	Association for Computing Machinery
ISBN (Electronic)	9781450371797
DOIs	https://doi.org/10.1145/3366030.3366083
Publication status	Published - 2019 Dec 2
Event	21st International Conference on Information Integration and Web-Based Applications and Services, iiWAS 2019 - Munich, Germany Duration: 2019 Dec 2 → 2019 Dec 4

Publication series

Name	ACM International Conference Proceeding Series

Conference

Conference	21st International Conference on Information Integration and Web-Based Applications and Services, iiWAS 2019
Country/Territory	Germany
City	Munich
Period	19/12/2 → 19/12/4

Keywords

Deep learning
Machine learning
Neural network
Non-native speaker
Speech recognition
Style transfer

ASJC Scopus subject areas

Software
Human-Computer Interaction
Computer Vision and Pattern Recognition
Computer Networks and Communications

Access to Document

10.1145/3366030.3366083

Cite this

Radzikowski, K., Forc, M., Wang, L., Yoshie, O., & Nowak, R. (2019). Accent neutralization for speech recognition of non-native speakers. In M. Indrawan-Santiago, E. Pardede, I. L. Salvadori, M. Steinbauer, I. Khalil, & G. Anderst-Kotsis (Eds.), 21st International Conference on Information Integration and Web-Based Applications and Services, iiWAS 2019 - Proceedings (ACM International Conference Proceeding Series). Association for Computing Machinery. https://doi.org/10.1145/3366030.3366083

Accent neutralization for speech recognition of non-native speakers. / Radzikowski, Kacper; Forc, Mateusz; Wang, Le et al.
21st International Conference on Information Integration and Web-Based Applications and Services, iiWAS 2019 - Proceedings. ed. / Maria Indrawan-Santiago; Eric Pardede; Ivan Luiz Salvadori; Matthias Steinbauer; Ismail Khalil; Gabriele Anderst-Kotsis. Association for Computing Machinery, 2019. (ACM International Conference Proceeding Series).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Radzikowski, K, Forc, M, Wang, L, Yoshie, O & Nowak, R 2019, Accent neutralization for speech recognition of non-native speakers. in M Indrawan-Santiago, E Pardede, IL Salvadori, M Steinbauer, I Khalil & G Anderst-Kotsis (eds), 21st International Conference on Information Integration and Web-Based Applications and Services, iiWAS 2019 - Proceedings. ACM International Conference Proceeding Series, Association for Computing Machinery, 21st International Conference on Information Integration and Web-Based Applications and Services, iiWAS 2019, Munich, Germany, 19/12/2. https://doi.org/10.1145/3366030.3366083

Radzikowski K, Forc M, Wang L, Yoshie O, Nowak R. Accent neutralization for speech recognition of non-native speakers. In Indrawan-Santiago M, Pardede E, Salvadori IL, Steinbauer M, Khalil I, Anderst-Kotsis G, editors, 21st International Conference on Information Integration and Web-Based Applications and Services, iiWAS 2019 - Proceedings. Association for Computing Machinery. 2019. (ACM International Conference Proceeding Series). doi: 10.1145/3366030.3366083

Radzikowski, Kacper ; Forc, Mateusz ; Wang, Le et al. / Accent neutralization for speech recognition of non-native speakers. 21st International Conference on Information Integration and Web-Based Applications and Services, iiWAS 2019 - Proceedings. editor / Maria Indrawan-Santiago ; Eric Pardede ; Ivan Luiz Salvadori ; Matthias Steinbauer ; Ismail Khalil ; Gabriele Anderst-Kotsis. Association for Computing Machinery, 2019. (ACM International Conference Proceeding Series).

@inproceedings{55cd843767774601afc3b3ad4154aa57,

title = "Accent neutralization for speech recognition of non-native speakers",

abstract = "These days, automatic speech recognition (ASR) systems achieve higher and higher accuracy rates. The score drops significantly, in case when the ASR system is being used with a non-native speaker of the language to be recognized. The main reason is specific pronunciation and accent features. A limited volume of labeled nonnative speech datasets makes it difficult to train new ASR systems for non-native speakers. In our research,we tried tackling the problem and its influence on the accuracy of ASR systems, using the style transfer methodology. We designed a pipeline for modifying the speech of a non-native speaker, so that it resembles the native speech to a higher extent. Our methodology can be used as a wrapper for any existing ASR system, which reduces the necessity of training new algorithms for non-native speech. The modification can be thus performed before passing the data forward to the speech recognition system itself.",

keywords = "Deep learning, Machine learning, Neural network, Non-native speaker, Speech recognition, Style transfer",

author = "Kacper Radzikowski and Mateusz Forc and Le Wang and Osamu Yoshie and Robert Nowak",

note = "Publisher Copyright: {\textcopyright} 2019 Association for Computing Machinery.; 21st International Conference on Information Integration and Web-Based Applications and Services, iiWAS 2019 ; Conference date: 02-12-2019 Through 04-12-2019",

year = "2019",

month = dec,

day = "2",

doi = "10.1145/3366030.3366083",

language = "English",

series = "ACM International Conference Proceeding Series",

publisher = "Association for Computing Machinery",

editor = "Maria Indrawan-Santiago and Eric Pardede and Salvadori, {Ivan Luiz} and Matthias Steinbauer and Ismail Khalil and Gabriele Anderst-Kotsis",

booktitle = "21st International Conference on Information Integration and Web-Based Applications and Services, iiWAS 2019 - Proceedings",

}

TY - GEN

T1 - Accent neutralization for speech recognition of non-native speakers

AU - Radzikowski, Kacper

AU - Forc, Mateusz

AU - Wang, Le

AU - Yoshie, Osamu

AU - Nowak, Robert

PY - 2019/12/2

Y1 - 2019/12/2

N2 - These days, automatic speech recognition (ASR) systems achieve higher and higher accuracy rates. The score drops significantly, in case when the ASR system is being used with a non-native speaker of the language to be recognized. The main reason is specific pronunciation and accent features. A limited volume of labeled nonnative speech datasets makes it difficult to train new ASR systems for non-native speakers. In our research,we tried tackling the problem and its influence on the accuracy of ASR systems, using the style transfer methodology. We designed a pipeline for modifying the speech of a non-native speaker, so that it resembles the native speech to a higher extent. Our methodology can be used as a wrapper for any existing ASR system, which reduces the necessity of training new algorithms for non-native speech. The modification can be thus performed before passing the data forward to the speech recognition system itself.

AB - These days, automatic speech recognition (ASR) systems achieve higher and higher accuracy rates. The score drops significantly, in case when the ASR system is being used with a non-native speaker of the language to be recognized. The main reason is specific pronunciation and accent features. A limited volume of labeled nonnative speech datasets makes it difficult to train new ASR systems for non-native speakers. In our research,we tried tackling the problem and its influence on the accuracy of ASR systems, using the style transfer methodology. We designed a pipeline for modifying the speech of a non-native speaker, so that it resembles the native speech to a higher extent. Our methodology can be used as a wrapper for any existing ASR system, which reduces the necessity of training new algorithms for non-native speech. The modification can be thus performed before passing the data forward to the speech recognition system itself.

KW - Deep learning

KW - Machine learning

KW - Neural network

KW - Non-native speaker

KW - Speech recognition

KW - Style transfer

UR - http://www.scopus.com/inward/record.url?scp=85123043043&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85123043043&partnerID=8YFLogxK

U2 - 10.1145/3366030.3366083

DO - 10.1145/3366030.3366083

M3 - Conference contribution

AN - SCOPUS:85123043043

T3 - ACM International Conference Proceeding Series

BT - 21st International Conference on Information Integration and Web-Based Applications and Services, iiWAS 2019 - Proceedings

A2 - Indrawan-Santiago, Maria

A2 - Pardede, Eric

A2 - Salvadori, Ivan Luiz

A2 - Steinbauer, Matthias

A2 - Khalil, Ismail

A2 - Anderst-Kotsis, Gabriele

PB - Association for Computing Machinery

T2 - 21st International Conference on Information Integration and Web-Based Applications and Services, iiWAS 2019

Y2 - 2 December 2019 through 4 December 2019

ER -

Accent neutralization for speech recognition of non-native speakers

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this