Automatic Rating of Spontaneous Speech for Low-Resource Languages

Ragheb Al-Ghezi; Yaroslav Getman; Ekaterina Voskoboinik; Mittul Singh; Mikko Kurimo

doi:10.1109/SLT54892.2023.10022381

Automatic Rating of Spontaneous Speech for Low-Resource Languages

Ragheb Al-Ghezi
, Yaroslav Getman
, Ekaterina Voskoboinik
, Mittul Singh
, Mikko Kurimo

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussa › Conference article in proceedings › Scientific › vertaisarvioitu

9 Sitaatiot (Scopus)

36 Lataukset (Pure)

Abstrakti

Automatic spontaneous speaking assessment systems bring numerous advantages to second language (L2) learning and assessment such as promoting self-learning and reducing language teachers' workload. Conventionally, these systems are developed for languages with a large number of learners due to the abundance of training data, yet languages with fewer learners such as Finnish and Swedish remain at a disadvantage due to the scarcity of required training data. Nevertheless, recent advancements in self-supervised deep learning make it possible to develop automatic speech recognition systems with a reasonable amount of training data. In turn, this advancement makes it feasible to develop systems for automatically assessing spoken proficiency of learners of underresourced languages: L2 Finnish and Finland Swedish. Our work evaluates the overall performance of the L2 ASR systems as well as the the rating systems compared to human reference ratings for both languages.

Alkuperäiskieli	Englanti
Otsikko	2022 IEEE Spoken Language Technology Workshop, SLT 2022 - Proceedings
Kustantaja	IEEE
Sivut	339-345
Sivumäärä	7
ISBN (elektroninen)	979-8-3503-9690-4
DOI - pysyväislinkit	https://doi.org/10.1109/SLT54892.2023.10022381
Tila	Julkaistu - 2023
OKM-julkaisutyyppi	A4 Artikkeli konferenssijulkaisussa
Tapahtuma	IEEE Spoken Language Technology Workshop - Doha, Qatar Kesto: 9 tammik. 2023 → 12 tammik. 2023

Julkaisusarja

Nimi	2022 IEEE Spoken Language Technology Workshop, SLT 2022 - Proceedings

Workshop

Workshop	IEEE Spoken Language Technology Workshop
Lyhennettä	SLT
Maa/Alue	Qatar
Kaupunki	Doha
Ajanjakso	09/01/2023 → 12/01/2023

Rahoitus

This work is part of Digitala project which is funded by the Academy of Finland (grant numbers 322619, 322625, 322965). The computational resources were provided by Aalto ScienceIT.

Pääsy asiakirjaan

10.1109/SLT54892.2023.10022381

SLT_2022_DigiTalaAccepted author manuscript, 173 KB

Muut tiedostot ja linkit

Linkki julkaisuun Scopuksessa

DigiTala: Aka-Digi Tala
Kurimo, M. (Vastuullinen johtaja), Getman, Y. (Projektin jäsen), Voskoboinik, E. (Projektin jäsen) & Al-Ghezi, R. (Projektin jäsen)
01/01/2020 → 31/08/2023
Projekti: RCF Academy Project targeted call

Science-IT
Hakala, M. (Manager)
Perustieteiden korkeakoulu
Laitteistot/tilat: Facility

Siteeraa tätä

@inproceedings{30ce700440fa47fca3d4b5695e3ee04c,

title = "Automatic Rating of Spontaneous Speech for Low-Resource Languages",

abstract = "Automatic spontaneous speaking assessment systems bring numerous advantages to second language (L2) learning and assessment such as promoting self-learning and reducing language teachers' workload. Conventionally, these systems are developed for languages with a large number of learners due to the abundance of training data, yet languages with fewer learners such as Finnish and Swedish remain at a disadvantage due to the scarcity of required training data. Nevertheless, recent advancements in self-supervised deep learning make it possible to develop automatic speech recognition systems with a reasonable amount of training data. In turn, this advancement makes it feasible to develop systems for automatically assessing spoken proficiency of learners of underresourced languages: L2 Finnish and Finland Swedish. Our work evaluates the overall performance of the L2 ASR systems as well as the the rating systems compared to human reference ratings for both languages.",

keywords = "automatic speaking assessment, low-resource, Wav2Vec2",

author = "Ragheb Al-Ghezi and Yaroslav Getman and Ekaterina Voskoboinik and Mittul Singh and Mikko Kurimo",

note = "Funding Information: This work is part of Digitala project which is funded by the Academy of Finland (grant numbers 322619, 322625, 322965). The computational resources were provided by Aalto ScienceIT. Publisher Copyright: {\textcopyright} 2023 IEEE.; IEEE Spoken Language Technology Workshop, SLT ; Conference date: 09-01-2023 Through 12-01-2023",

year = "2023",

doi = "10.1109/SLT54892.2023.10022381",

language = "English",

series = "2022 IEEE Spoken Language Technology Workshop, SLT 2022 - Proceedings",

publisher = "IEEE",

pages = "339--345",

booktitle = "2022 IEEE Spoken Language Technology Workshop, SLT 2022 - Proceedings",

address = "United States",

}

Al-Ghezi, R, Getman, Y, Voskoboinik, E, Singh, M & Kurimo, M 2023, Automatic Rating of Spontaneous Speech for Low-Resource Languages. julkaisussa 2022 IEEE Spoken Language Technology Workshop, SLT 2022 - Proceedings. 2022 IEEE Spoken Language Technology Workshop, SLT 2022 - Proceedings, IEEE, Sivut 339-345, IEEE Spoken Language Technology Workshop, Doha, Qatar, 09/01/2023. https://doi.org/10.1109/SLT54892.2023.10022381

Automatic Rating of Spontaneous Speech for Low-Resource Languages. / Al-Ghezi, Ragheb; Getman, Yaroslav; Voskoboinik, Ekaterina et al.
2022 IEEE Spoken Language Technology Workshop, SLT 2022 - Proceedings. IEEE, 2023. s. 339-345 (2022 IEEE Spoken Language Technology Workshop, SLT 2022 - Proceedings).

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussa › Conference article in proceedings › Scientific › vertaisarvioitu

TY - GEN

T1 - Automatic Rating of Spontaneous Speech for Low-Resource Languages

AU - Al-Ghezi, Ragheb

AU - Getman, Yaroslav

AU - Voskoboinik, Ekaterina

AU - Singh, Mittul

AU - Kurimo, Mikko

N1 - Funding Information: This work is part of Digitala project which is funded by the Academy of Finland (grant numbers 322619, 322625, 322965). The computational resources were provided by Aalto ScienceIT. Publisher Copyright: © 2023 IEEE.

PY - 2023

Y1 - 2023

N2 - Automatic spontaneous speaking assessment systems bring numerous advantages to second language (L2) learning and assessment such as promoting self-learning and reducing language teachers' workload. Conventionally, these systems are developed for languages with a large number of learners due to the abundance of training data, yet languages with fewer learners such as Finnish and Swedish remain at a disadvantage due to the scarcity of required training data. Nevertheless, recent advancements in self-supervised deep learning make it possible to develop automatic speech recognition systems with a reasonable amount of training data. In turn, this advancement makes it feasible to develop systems for automatically assessing spoken proficiency of learners of underresourced languages: L2 Finnish and Finland Swedish. Our work evaluates the overall performance of the L2 ASR systems as well as the the rating systems compared to human reference ratings for both languages.

AB - Automatic spontaneous speaking assessment systems bring numerous advantages to second language (L2) learning and assessment such as promoting self-learning and reducing language teachers' workload. Conventionally, these systems are developed for languages with a large number of learners due to the abundance of training data, yet languages with fewer learners such as Finnish and Swedish remain at a disadvantage due to the scarcity of required training data. Nevertheless, recent advancements in self-supervised deep learning make it possible to develop automatic speech recognition systems with a reasonable amount of training data. In turn, this advancement makes it feasible to develop systems for automatically assessing spoken proficiency of learners of underresourced languages: L2 Finnish and Finland Swedish. Our work evaluates the overall performance of the L2 ASR systems as well as the the rating systems compared to human reference ratings for both languages.

KW - automatic speaking assessment

KW - low-resource

KW - Wav2Vec2

UR - https://www.scopus.com/pages/publications/85147799145

U2 - 10.1109/SLT54892.2023.10022381

DO - 10.1109/SLT54892.2023.10022381

M3 - Conference article in proceedings

AN - SCOPUS:85147799145

T3 - 2022 IEEE Spoken Language Technology Workshop, SLT 2022 - Proceedings

SP - 339

EP - 345

BT - 2022 IEEE Spoken Language Technology Workshop, SLT 2022 - Proceedings

PB - IEEE

T2 - IEEE Spoken Language Technology Workshop

Y2 - 9 January 2023 through 12 January 2023

ER -

Automatic Rating of Spontaneous Speech for Low-Resource Languages

Abstrakti

Julkaisusarja

Workshop

Rahoitus

Pääsy asiakirjaan

Muut tiedostot ja linkit

Sormenjälki

Projektit

DigiTala: Aka-Digi Tala

Laitteet

Science-IT

Siteeraa tätä