Speech recognition challenge in the wild

Arabic MGB-3

Ahmed Ali, Stephan Vogel, Steve Renals

Research output: Chapter in Book/Report/Conference proceedingConference contribution

16 Citations (Scopus)

Abstract

This paper describes the Arabic MGB-3 Challenge-Arabic Speech Recognition in the Wild. Unlike last year's Arabic MGB-2 Challenge, for which the recognition task was based on more than 1,200 hours broadcast TV news recordings from Aljazeera Arabic TV programs, MGB-3 emphasises dialectal Arabic using a multi-genre collection of Egyptian YouTube videos. Seven genres were used for the data collection: Comedy, cooking, family/kids, fashion, drama, sports, and science (TEDx). A total of 16 hours of videos, split evenly across the different genres, were divided into adaptation, development and evaluation data sets. The Arabic MGB-Challenge comprised two tasks: A) Speech transcription, evaluated on the MGB-3 test set, along with the 10 hour MGB-2 test set to report progress on the MGB-2 evaluation; B) Arabic dialect identification, introduced this year in order to distinguish between four major Arabic dialects-Egyptian, Levantine, North African, Gulf, as well as Modern Standard Arabic. Two hours of audio per dialect were released for development and a further two hours were used for evaluation. For dialect identification, both lexical features and i-vector bottleneck features were shared with participants in addition to the raw audio recordings. Overall, thirteen teams submitted ten systems to the challenge. We outline the approaches adopted in each system, and summarise the evaluation results.

Original languageEnglish
Title of host publication2017 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2017 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages316-322
Number of pages7
Volume2018-January
ISBN (Electronic)9781509047888
DOIs
Publication statusPublished - 24 Jan 2018
Event2017 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2017 - Okinawa, Japan
Duration: 16 Dec 201720 Dec 2017

Other

Other2017 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2017
CountryJapan
CityOkinawa
Period16/12/1720/12/17

Fingerprint

Audio recordings
Cooking
Transcription
Sports
Speech recognition

Keywords

  • broadcast speech
  • dialect identification
  • multi-reference WER
  • multigenre
  • Speech recognition
  • under-resource

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Human-Computer Interaction

Cite this

Ali, A., Vogel, S., & Renals, S. (2018). Speech recognition challenge in the wild: Arabic MGB-3. In 2017 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2017 - Proceedings (Vol. 2018-January, pp. 316-322). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ASRU.2017.8268952

Speech recognition challenge in the wild : Arabic MGB-3. / Ali, Ahmed; Vogel, Stephan; Renals, Steve.

2017 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2017 - Proceedings. Vol. 2018-January Institute of Electrical and Electronics Engineers Inc., 2018. p. 316-322.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Ali, A, Vogel, S & Renals, S 2018, Speech recognition challenge in the wild: Arabic MGB-3. in 2017 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2017 - Proceedings. vol. 2018-January, Institute of Electrical and Electronics Engineers Inc., pp. 316-322, 2017 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2017, Okinawa, Japan, 16/12/17. https://doi.org/10.1109/ASRU.2017.8268952
Ali A, Vogel S, Renals S. Speech recognition challenge in the wild: Arabic MGB-3. In 2017 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2017 - Proceedings. Vol. 2018-January. Institute of Electrical and Electronics Engineers Inc. 2018. p. 316-322 https://doi.org/10.1109/ASRU.2017.8268952
Ali, Ahmed ; Vogel, Stephan ; Renals, Steve. / Speech recognition challenge in the wild : Arabic MGB-3. 2017 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2017 - Proceedings. Vol. 2018-January Institute of Electrical and Electronics Engineers Inc., 2018. pp. 316-322
@inproceedings{c551d6a262f14e9782373cbf881496ab,
title = "Speech recognition challenge in the wild: Arabic MGB-3",
abstract = "This paper describes the Arabic MGB-3 Challenge-Arabic Speech Recognition in the Wild. Unlike last year's Arabic MGB-2 Challenge, for which the recognition task was based on more than 1,200 hours broadcast TV news recordings from Aljazeera Arabic TV programs, MGB-3 emphasises dialectal Arabic using a multi-genre collection of Egyptian YouTube videos. Seven genres were used for the data collection: Comedy, cooking, family/kids, fashion, drama, sports, and science (TEDx). A total of 16 hours of videos, split evenly across the different genres, were divided into adaptation, development and evaluation data sets. The Arabic MGB-Challenge comprised two tasks: A) Speech transcription, evaluated on the MGB-3 test set, along with the 10 hour MGB-2 test set to report progress on the MGB-2 evaluation; B) Arabic dialect identification, introduced this year in order to distinguish between four major Arabic dialects-Egyptian, Levantine, North African, Gulf, as well as Modern Standard Arabic. Two hours of audio per dialect were released for development and a further two hours were used for evaluation. For dialect identification, both lexical features and i-vector bottleneck features were shared with participants in addition to the raw audio recordings. Overall, thirteen teams submitted ten systems to the challenge. We outline the approaches adopted in each system, and summarise the evaluation results.",
keywords = "broadcast speech, dialect identification, multi-reference WER, multigenre, Speech recognition, under-resource",
author = "Ahmed Ali and Stephan Vogel and Steve Renals",
year = "2018",
month = "1",
day = "24",
doi = "10.1109/ASRU.2017.8268952",
language = "English",
volume = "2018-January",
pages = "316--322",
booktitle = "2017 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2017 - Proceedings",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Speech recognition challenge in the wild

T2 - Arabic MGB-3

AU - Ali, Ahmed

AU - Vogel, Stephan

AU - Renals, Steve

PY - 2018/1/24

Y1 - 2018/1/24

N2 - This paper describes the Arabic MGB-3 Challenge-Arabic Speech Recognition in the Wild. Unlike last year's Arabic MGB-2 Challenge, for which the recognition task was based on more than 1,200 hours broadcast TV news recordings from Aljazeera Arabic TV programs, MGB-3 emphasises dialectal Arabic using a multi-genre collection of Egyptian YouTube videos. Seven genres were used for the data collection: Comedy, cooking, family/kids, fashion, drama, sports, and science (TEDx). A total of 16 hours of videos, split evenly across the different genres, were divided into adaptation, development and evaluation data sets. The Arabic MGB-Challenge comprised two tasks: A) Speech transcription, evaluated on the MGB-3 test set, along with the 10 hour MGB-2 test set to report progress on the MGB-2 evaluation; B) Arabic dialect identification, introduced this year in order to distinguish between four major Arabic dialects-Egyptian, Levantine, North African, Gulf, as well as Modern Standard Arabic. Two hours of audio per dialect were released for development and a further two hours were used for evaluation. For dialect identification, both lexical features and i-vector bottleneck features were shared with participants in addition to the raw audio recordings. Overall, thirteen teams submitted ten systems to the challenge. We outline the approaches adopted in each system, and summarise the evaluation results.

AB - This paper describes the Arabic MGB-3 Challenge-Arabic Speech Recognition in the Wild. Unlike last year's Arabic MGB-2 Challenge, for which the recognition task was based on more than 1,200 hours broadcast TV news recordings from Aljazeera Arabic TV programs, MGB-3 emphasises dialectal Arabic using a multi-genre collection of Egyptian YouTube videos. Seven genres were used for the data collection: Comedy, cooking, family/kids, fashion, drama, sports, and science (TEDx). A total of 16 hours of videos, split evenly across the different genres, were divided into adaptation, development and evaluation data sets. The Arabic MGB-Challenge comprised two tasks: A) Speech transcription, evaluated on the MGB-3 test set, along with the 10 hour MGB-2 test set to report progress on the MGB-2 evaluation; B) Arabic dialect identification, introduced this year in order to distinguish between four major Arabic dialects-Egyptian, Levantine, North African, Gulf, as well as Modern Standard Arabic. Two hours of audio per dialect were released for development and a further two hours were used for evaluation. For dialect identification, both lexical features and i-vector bottleneck features were shared with participants in addition to the raw audio recordings. Overall, thirteen teams submitted ten systems to the challenge. We outline the approaches adopted in each system, and summarise the evaluation results.

KW - broadcast speech

KW - dialect identification

KW - multi-reference WER

KW - multigenre

KW - Speech recognition

KW - under-resource

UR - http://www.scopus.com/inward/record.url?scp=85050537018&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85050537018&partnerID=8YFLogxK

U2 - 10.1109/ASRU.2017.8268952

DO - 10.1109/ASRU.2017.8268952

M3 - Conference contribution

VL - 2018-January

SP - 316

EP - 322

BT - 2017 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2017 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

ER -