A Lightweight Downscaled Approach to Automatic Speech Recognition for Small Indigenous Languages

George Vlad Stan; André Baart; Francis Dittoh; Hans Akkermans; Anna Bon

doi:10.1145/3501247.3539017

A Lightweight Downscaled Approach to Automatic Speech Recognition for Small Indigenous Languages

George Vlad Stan, André Baart, Francis Dittoh, Hans Akkermans, Anna Bon

Research output: Chapter in Book / Report / Conference proceeding › Conference contribution › Academic › peer-review

Abstract

Development of fully featured Automatic Speech Recognition (ASR) systems for a complete language vocabulary generally requires large data repositories, massive computing power, and a stable digital network infrastructure. These conditions are not met in the case of many indigenous languages. Based on our research for over a decade in West Africa, we present a lightweight and downscaled approach to AI-based ASR and describe a set of associated experiments. The aim is to produce a variety of limited-vocabulary ASRs as a basis for the development of practically useful (mobile and radio) voice-based information services that fit needs, preferences and knowledge of local rural communities.

Original language	English
Title of host publication	WebSci '22
Subtitle of host publication	[Proceedings] 14th ACM Web Science Conference 2022
Publisher	Association for Computing Machinery
Pages	451-458
Number of pages	8
ISBN (Electronic)	9781450391917
DOIs	https://doi.org/10.1145/3501247.3539017
Publication status	Published - Jun 2022
Event	14th ACM Web Science Conference, WebSci 2022 - Virtual, Online, Spain Duration: 26 Jun 2022 → 29 Jun 2022

Publication series

Name	ACM International Conference Proceeding Series

Conference

Conference	14th ACM Web Science Conference, WebSci 2022
Country/Territory	Spain
City	Virtual, Online
Period	26/06/22 → 29/06/22

Bibliographical note

Publisher Copyright:
© 2022 ACM.

Keywords

automatic speech recognition
low resource environments
machine learning
neural networks
under-resourced/indigenous languages
voice-based technologies

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1145/3501247.3539017

Persistent URL (handle)

Persistent link

Cite this

Stan, G. V., Baart, A., Dittoh, F., Akkermans, H., & Bon, A. (2022). A Lightweight Downscaled Approach to Automatic Speech Recognition for Small Indigenous Languages. In WebSci '22: [Proceedings] 14th ACM Web Science Conference 2022 (pp. 451-458). (ACM International Conference Proceeding Series). Association for Computing Machinery. https://doi.org/10.1145/3501247.3539017

@inproceedings{768edc4627014b8b91d80f53125ed150,

title = "A Lightweight Downscaled Approach to Automatic Speech Recognition for Small Indigenous Languages",

abstract = "Development of fully featured Automatic Speech Recognition (ASR) systems for a complete language vocabulary generally requires large data repositories, massive computing power, and a stable digital network infrastructure. These conditions are not met in the case of many indigenous languages. Based on our research for over a decade in West Africa, we present a lightweight and downscaled approach to AI-based ASR and describe a set of associated experiments. The aim is to produce a variety of limited-vocabulary ASRs as a basis for the development of practically useful (mobile and radio) voice-based information services that fit needs, preferences and knowledge of local rural communities. ",

keywords = "automatic speech recognition, low resource environments, machine learning, neural networks, under-resourced/indigenous languages, voice-based technologies",

author = "Stan, {George Vlad} and Andr{\'e} Baart and Francis Dittoh and Hans Akkermans and Anna Bon",

note = "Publisher Copyright: {\textcopyright} 2022 ACM.; 14th ACM Web Science Conference, WebSci 2022 ; Conference date: 26-06-2022 Through 29-06-2022",

year = "2022",

month = jun,

doi = "10.1145/3501247.3539017",

language = "English",

series = "ACM International Conference Proceeding Series",

publisher = "Association for Computing Machinery",

pages = "451--458",

booktitle = "WebSci '22",

}

Stan, GV, Baart, A, Dittoh, F, Akkermans, H & Bon, A 2022, A Lightweight Downscaled Approach to Automatic Speech Recognition for Small Indigenous Languages. in WebSci '22: [Proceedings] 14th ACM Web Science Conference 2022. ACM International Conference Proceeding Series, Association for Computing Machinery, pp. 451-458, 14th ACM Web Science Conference, WebSci 2022, Virtual, Online, Spain, 26/06/22. https://doi.org/10.1145/3501247.3539017

A Lightweight Downscaled Approach to Automatic Speech Recognition for Small Indigenous Languages. / Stan, George Vlad; Baart, André; Dittoh, Francis et al.
WebSci '22: [Proceedings] 14th ACM Web Science Conference 2022. Association for Computing Machinery, 2022. p. 451-458 (ACM International Conference Proceeding Series).

Research output: Chapter in Book / Report / Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - A Lightweight Downscaled Approach to Automatic Speech Recognition for Small Indigenous Languages

AU - Stan, George Vlad

AU - Baart, André

AU - Dittoh, Francis

AU - Akkermans, Hans

AU - Bon, Anna

PY - 2022/6

Y1 - 2022/6

N2 - Development of fully featured Automatic Speech Recognition (ASR) systems for a complete language vocabulary generally requires large data repositories, massive computing power, and a stable digital network infrastructure. These conditions are not met in the case of many indigenous languages. Based on our research for over a decade in West Africa, we present a lightweight and downscaled approach to AI-based ASR and describe a set of associated experiments. The aim is to produce a variety of limited-vocabulary ASRs as a basis for the development of practically useful (mobile and radio) voice-based information services that fit needs, preferences and knowledge of local rural communities.

AB - Development of fully featured Automatic Speech Recognition (ASR) systems for a complete language vocabulary generally requires large data repositories, massive computing power, and a stable digital network infrastructure. These conditions are not met in the case of many indigenous languages. Based on our research for over a decade in West Africa, we present a lightweight and downscaled approach to AI-based ASR and describe a set of associated experiments. The aim is to produce a variety of limited-vocabulary ASRs as a basis for the development of practically useful (mobile and radio) voice-based information services that fit needs, preferences and knowledge of local rural communities.

KW - automatic speech recognition

KW - low resource environments

KW - machine learning

KW - neural networks

KW - under-resourced/indigenous languages

KW - voice-based technologies

UR - http://www.scopus.com/inward/record.url?scp=85133667316&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85133667316&partnerID=8YFLogxK

U2 - 10.1145/3501247.3539017

DO - 10.1145/3501247.3539017

M3 - Conference contribution

AN - SCOPUS:85133667316

T3 - ACM International Conference Proceeding Series

SP - 451

EP - 458

BT - WebSci '22

PB - Association for Computing Machinery

T2 - 14th ACM Web Science Conference, WebSci 2022

Y2 - 26 June 2022 through 29 June 2022

ER -

A Lightweight Downscaled Approach to Automatic Speech Recognition for Small Indigenous Languages

Abstract

Publication series

Conference

Bibliographical note

Keywords

UN SDGs

Access to Document

Persistent URL (handle)

Other files and links

Fingerprint

Cite this