Object's details: Accents in Speech Recognition through the Lens of a World Englishes Evaluation Set

Provider:Repozytorium Uniwersytetu Łódzkiego

Description

Title:
Accents in Speech Recognition through the Lens of a World Englishes Evaluation Set
Creator:
Del Río, Miguel;
Miller, Corey;
Profant, Ján;
Drexler-Fox, Jennifer;
Mcnamara, Quinn;
Bhandari, Nishchal;
Delworth, Natalie;
Pirkin, Ilya;
Jetté, Migüel;
Chandra, Shipra;
Ha, Peter;
Westerman, Ryan
Contributor:
Del Río, Miguel - Rev.com;
Miller, Corey - Rev.com;
Profant, Ján - Rev.com;
Drexler-Fox, Jennifer - Rev.com;
Mcnamara, Quinn - Rev.com;
Bhandari, Nishchal - Rev.com;
Delworth, Natalie - Rev.com;
Pirkin, Ilya - Rev.com;
Jetté, Migüel - Rev.com;
Chandra, Shipra - Walgreens;
Ha, Peter - Northwestern University;
Westerman, Ryan - Zoom;
Del Río, Miguel - miguel.delrio@rev.com;
Miller, Corey - corey.miller@rev.com;
Profant, Ján - ril@uni.lodz.pl;
Drexler-Fox, Jennifer - ril@uni.lodz.pl;
Mcnamara, Quinn - ril@uni.lodz.pl;
Bhandari, Nishchal - ril@uni.lodz.pl;
Delworth, Natalie - ril@uni.lodz.pl;
Pirkin, Ilya - ril@uni.lodz.pl;
Jetté, Migüel - ril@uni.lodz.pl;
Chandra, Shipra - ril@uni.lodz.pl;
Ha, Peter - ril@uni.lodz.pl;
Westerman, Ryan - ril@uni.lodz.pl
Description:
Automatic Speech Recognition (ASR) systems generalize poorly on accented speech, creating bias issues for users and providers. The phonetic and linguistic variability of accents present challenges for ASR systems in both data collection and modeling strategies. We present two promising approaches to accented speech recognition— custom vocabulary and multilingual modeling— and highlight key challenges in the space. Among these, lack of a standard benchmark makes research and comparison difficult. We address this with a novel corpus of accented speech: Earnings-22, A 125 file, 119 hour corpus of English-language earnings calls gathered from global companies. We compare commercial models showing variation in performance when taking country of origin into consideration and demonstrate targeted improvements using the methods we introduce.
Object availability:
free access
Rights: https://creativecommons.org/licenses/by-nc-nd/4.0
Date:
2024.01.03;
2023.12.28
Type:
Article
Language:
English
Format:
application/pdf
Publisher:
Wydawnictwo Uniwersytetu Łódzkiego
Relation:
Research in Language;3;
21
Subject:
accents;
dialects;
speech recognition;
bias;
multilingual
Identifier:
1731-7533;
http://hdl.handle.net/11089/49005 ;
10.18778/1731-7533.21.3.02
Data provider:
Repozytorium Uniwersytetu Łódzkiego

Similar objects

Accents in Speech Recognition through the Lens of a World Englishes Evaluation Set

Date:2024.01.03 | 2023.12.28

Type:image

Irena Maryniakowa, Dorota K. Rembiszewska, Janusz Siatkowski, Różnojęzyczne słownictwo gwarowe Podlasia, Suwalszczyzny i północno-wschodniego Mazowsza, Warszawa: Instytut Slawistyki PAN, 2014, 476 ss.

Creator:Głuszkowski, Michał

Date:2015.12.31

Type:text

Alignment in ASR and L1 Listeners’ Recognition of L2 Learner Speech: French EFL Learners & Dictation.Io

Creator:Chanethom, Vincent | Henderson, Alice

Date:2024.01.03 | 2023.12.28

Type:image

Navigating Recognition: The Symbolic Struggles in the Biographies of Young Polish Internet Content Creators

Creator:Łuczaj, Kamil

Date:2025.05.07 | 2025.04.30

Type:image

Auditive perception, acoustic and distributional properties of syllables in Polish

Creator:Śledziński, Daniel

Date:2015.12.31

Type:text

Causality and ability beliefs: An introduction to confounders and colliders

Creator:Al-Hoorie, Ali H. | Hiver, Phil

Date:2025.06.30

Type:text

Developing Unbiased Teacher Identity in Pluri-Accent Reality: Research-Based Classroom Activities

Creator:Červinková Poesová , Kristýna | Lancová , Klára

Date:2023.01.10 | 2022.12.29

Type:image

Accents of English at Czech schools: Students’ attitudes and recognition skills

Creator:Jakšič, Jan | Šturm, Pavel

Date:2018.04.03 | 2018.02.06

Type:other

Federacja Bibliotek Cyfrowych

Object's details: Accents in Speech Recognition through the Lens of a World Englishes Evaluation Set

Description

Similar objects