Transformer-Based Multilingual G2P Converter for E-Learning System

Jueting Liu; Chang Ren; Yaoxuan Luan; Sicheng Li; Tianshi Xie; Cheryl Seals; Marisha Speights Atkins

doi:10.1007/978-3-031-05643-7_35

Transformer-Based Multilingual G2P Converter for E-Learning System

Jueting Liu, Chang Ren, Yaoxuan Luan, Sicheng Li, Tianshi Xie, Cheryl Seals^*, Marisha Speights Atkins

^*Corresponding author for this work

Communication Sciences and Disorders

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

1 Scopus citations

Abstract

Phonetic transcription is an approach to represent speech sounds to specific symbols. The most common alphabet we used is the International Phonetic Alphabet (IPA), and the characters in the IPA are phonetic symbols. To support the phonetic transcription process in the phonetic exams of our linguistic E-learning system, we designed a machine translation tool that aims to translate English words to their phonetic formats. This progress can also be expressed as grapheme to phoneme (G2P). The Transformer model has been utilized to develop this G2P module. Also, to improve the functionality of the E-learning system, we trained multiple language models and generated a multilingual G2P translator. Moreover, we evaluated our G2P system by word error rate (WER) and phoneme error rate (PER) with edit distance.

Original language	English (US)
Title of host publication	Artificial Intelligence in HCI - 3rd International Conference, AI-HCI 2022, Held as Part of the 24th HCI International Conference, HCII 2022, Proceedings
Editors	Helmut Degen, Stavroula Ntoa
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	546-556
Number of pages	11
ISBN (Print)	9783031056420
DOIs	https://doi.org/10.1007/978-3-031-05643-7_35
State	Published - 2022
Event	3rd International Conference on Artificial Intelligence in HCI, AI-HCI 2022 Held as Part of the 24th HCI International Conference, HCII 2022 - Virtual, Online Duration: Jun 26 2022 → Jul 1 2022

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	13336 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	3rd International Conference on Artificial Intelligence in HCI, AI-HCI 2022 Held as Part of the 24th HCI International Conference, HCII 2022
City	Virtual, Online
Period	6/26/22 → 7/1/22

Keywords

E-learning
Grapheme-to-phoneme
International phonetic alphabet
Transformer

ASJC Scopus subject areas

Theoretical Computer Science
General Computer Science

Access to Document

10.1007/978-3-031-05643-7_35

Cite this

Liu, J., Ren, C., Luan, Y., Li, S., Xie, T., Seals, C., & Speights Atkins, M. (2022). Transformer-Based Multilingual G2P Converter for E-Learning System. In H. Degen, & S. Ntoa (Eds.), Artificial Intelligence in HCI - 3rd International Conference, AI-HCI 2022, Held as Part of the 24th HCI International Conference, HCII 2022, Proceedings (pp. 546-556). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13336 LNAI). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-05643-7_35

Liu, Jueting ; Ren, Chang ; Luan, Yaoxuan et al. / Transformer-Based Multilingual G2P Converter for E-Learning System. Artificial Intelligence in HCI - 3rd International Conference, AI-HCI 2022, Held as Part of the 24th HCI International Conference, HCII 2022, Proceedings. editor / Helmut Degen ; Stavroula Ntoa. Springer Science and Business Media Deutschland GmbH, 2022. pp. 546-556 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{97570c9c79cf4f65a7f7078e91c984fc,

title = "Transformer-Based Multilingual G2P Converter for E-Learning System",

abstract = "Phonetic transcription is an approach to represent speech sounds to specific symbols. The most common alphabet we used is the International Phonetic Alphabet (IPA), and the characters in the IPA are phonetic symbols. To support the phonetic transcription process in the phonetic exams of our linguistic E-learning system, we designed a machine translation tool that aims to translate English words to their phonetic formats. This progress can also be expressed as grapheme to phoneme (G2P). The Transformer model has been utilized to develop this G2P module. Also, to improve the functionality of the E-learning system, we trained multiple language models and generated a multilingual G2P translator. Moreover, we evaluated our G2P system by word error rate (WER) and phoneme error rate (PER) with edit distance.",

keywords = "E-learning, Grapheme-to-phoneme, International phonetic alphabet, Transformer",

author = "Jueting Liu and Chang Ren and Yaoxuan Luan and Sicheng Li and Tianshi Xie and Cheryl Seals and {Speights Atkins}, Marisha",

note = "Publisher Copyright: {\textcopyright} 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.; 3rd International Conference on Artificial Intelligence in HCI, AI-HCI 2022 Held as Part of the 24th HCI International Conference, HCII 2022 ; Conference date: 26-06-2022 Through 01-07-2022",

year = "2022",

doi = "10.1007/978-3-031-05643-7_35",

language = "English (US)",

isbn = "9783031056420",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "546--556",

editor = "Helmut Degen and Stavroula Ntoa",

booktitle = "Artificial Intelligence in HCI - 3rd International Conference, AI-HCI 2022, Held as Part of the 24th HCI International Conference, HCII 2022, Proceedings",

address = "Germany",

}

Liu, J, Ren, C, Luan, Y, Li, S, Xie, T, Seals, C & Speights Atkins, M 2022, Transformer-Based Multilingual G2P Converter for E-Learning System. in H Degen & S Ntoa (eds), Artificial Intelligence in HCI - 3rd International Conference, AI-HCI 2022, Held as Part of the 24th HCI International Conference, HCII 2022, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 13336 LNAI, Springer Science and Business Media Deutschland GmbH, pp. 546-556, 3rd International Conference on Artificial Intelligence in HCI, AI-HCI 2022 Held as Part of the 24th HCI International Conference, HCII 2022, Virtual, Online, 6/26/22. https://doi.org/10.1007/978-3-031-05643-7_35

Transformer-Based Multilingual G2P Converter for E-Learning System. / Liu, Jueting; Ren, Chang; Luan, Yaoxuan et al.
Artificial Intelligence in HCI - 3rd International Conference, AI-HCI 2022, Held as Part of the 24th HCI International Conference, HCII 2022, Proceedings. ed. / Helmut Degen; Stavroula Ntoa. Springer Science and Business Media Deutschland GmbH, 2022. p. 546-556 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13336 LNAI).

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

TY - GEN

T1 - Transformer-Based Multilingual G2P Converter for E-Learning System

AU - Liu, Jueting

AU - Ren, Chang

AU - Luan, Yaoxuan

AU - Li, Sicheng

AU - Xie, Tianshi

AU - Seals, Cheryl

AU - Speights Atkins, Marisha

PY - 2022

Y1 - 2022

N2 - Phonetic transcription is an approach to represent speech sounds to specific symbols. The most common alphabet we used is the International Phonetic Alphabet (IPA), and the characters in the IPA are phonetic symbols. To support the phonetic transcription process in the phonetic exams of our linguistic E-learning system, we designed a machine translation tool that aims to translate English words to their phonetic formats. This progress can also be expressed as grapheme to phoneme (G2P). The Transformer model has been utilized to develop this G2P module. Also, to improve the functionality of the E-learning system, we trained multiple language models and generated a multilingual G2P translator. Moreover, we evaluated our G2P system by word error rate (WER) and phoneme error rate (PER) with edit distance.

AB - Phonetic transcription is an approach to represent speech sounds to specific symbols. The most common alphabet we used is the International Phonetic Alphabet (IPA), and the characters in the IPA are phonetic symbols. To support the phonetic transcription process in the phonetic exams of our linguistic E-learning system, we designed a machine translation tool that aims to translate English words to their phonetic formats. This progress can also be expressed as grapheme to phoneme (G2P). The Transformer model has been utilized to develop this G2P module. Also, to improve the functionality of the E-learning system, we trained multiple language models and generated a multilingual G2P translator. Moreover, we evaluated our G2P system by word error rate (WER) and phoneme error rate (PER) with edit distance.

KW - E-learning

KW - Grapheme-to-phoneme

KW - International phonetic alphabet

KW - Transformer

UR - http://www.scopus.com/inward/record.url?scp=85131127062&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85131127062&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-05643-7_35

DO - 10.1007/978-3-031-05643-7_35

M3 - Conference contribution

AN - SCOPUS:85131127062

SN - 9783031056420

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 546

EP - 556

BT - Artificial Intelligence in HCI - 3rd International Conference, AI-HCI 2022, Held as Part of the 24th HCI International Conference, HCII 2022, Proceedings

A2 - Degen, Helmut

A2 - Ntoa, Stavroula

PB - Springer Science and Business Media Deutschland GmbH

T2 - 3rd International Conference on Artificial Intelligence in HCI, AI-HCI 2022 Held as Part of the 24th HCI International Conference, HCII 2022

Y2 - 26 June 2022 through 1 July 2022

ER -

Liu J, Ren C, Luan Y, Li S, Xie T, Seals C et al. Transformer-Based Multilingual G2P Converter for E-Learning System. In Degen H, Ntoa S, editors, Artificial Intelligence in HCI - 3rd International Conference, AI-HCI 2022, Held as Part of the 24th HCI International Conference, HCII 2022, Proceedings. Springer Science and Business Media Deutschland GmbH. 2022. p. 546-556. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-05643-7_35