Accents in Speech Recognition through the Lens of a World Englishes Evaluation Set
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...
Saved in:
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Lodz University Press
2023-12-01
|
| Series: | Research in Language |
| Subjects: | |
| Online Access: | https://czasopisma.uni.lodz.pl/research/article/view/21579 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841560824746868736 |
|---|---|
| author | Miguel Del Río Corey Miller Ján Profant Jennifer Drexler-Fox Quinn Mcnamara Nishchal Bhandari Natalie Delworth Ilya Pirkin Migüel Jetté Shipra Chandra Peter Ha Ryan Westerman |
| author_facet | Miguel Del Río Corey Miller Ján Profant Jennifer Drexler-Fox Quinn Mcnamara Nishchal Bhandari Natalie Delworth Ilya Pirkin Migüel Jetté Shipra Chandra Peter Ha Ryan Westerman |
| author_sort | Miguel Del Río |
| collection | DOAJ |
| 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. |
| format | Article |
| id | doaj-art-ffe1012f0388429c8dcac424bf0fd8dd |
| institution | Kabale University |
| issn | 1731-7533 |
| language | English |
| publishDate | 2023-12-01 |
| publisher | Lodz University Press |
| record_format | Article |
| series | Research in Language |
| spelling | doaj-art-ffe1012f0388429c8dcac424bf0fd8dd2025-01-03T14:52:21ZengLodz University PressResearch in Language1731-75332023-12-0121322524410.18778/1731-7533.21.3.0221571Accents in Speech Recognition through the Lens of a World Englishes Evaluation SetMiguel Del Río0Corey Miller1Ján Profant2Jennifer Drexler-Fox3Quinn Mcnamara4Nishchal Bhandari5Natalie Delworth6Ilya Pirkin7Migüel Jetté8Shipra Chandra9Peter Ha10Ryan Westerman11Rev.comRev.comRev.comRev.comRev.comRev.comRev.comRev.comRev.comWalgreensNorthwestern University ZoomAutomatic 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.https://czasopisma.uni.lodz.pl/research/article/view/21579accentsdialectsspeech recognitionbiasmultilingual |
| spellingShingle | Miguel Del Río Corey Miller Ján Profant Jennifer Drexler-Fox Quinn Mcnamara Nishchal Bhandari Natalie Delworth Ilya Pirkin Migüel Jetté Shipra Chandra Peter Ha Ryan Westerman Accents in Speech Recognition through the Lens of a World Englishes Evaluation Set Research in Language accents dialects speech recognition bias multilingual |
| title | Accents in Speech Recognition through the Lens of a World Englishes Evaluation Set |
| title_full | Accents in Speech Recognition through the Lens of a World Englishes Evaluation Set |
| title_fullStr | Accents in Speech Recognition through the Lens of a World Englishes Evaluation Set |
| title_full_unstemmed | Accents in Speech Recognition through the Lens of a World Englishes Evaluation Set |
| title_short | Accents in Speech Recognition through the Lens of a World Englishes Evaluation Set |
| title_sort | accents in speech recognition through the lens of a world englishes evaluation set |
| topic | accents dialects speech recognition bias multilingual |
| url | https://czasopisma.uni.lodz.pl/research/article/view/21579 |
| work_keys_str_mv | AT migueldelrio accentsinspeechrecognitionthroughthelensofaworldenglishesevaluationset AT coreymiller accentsinspeechrecognitionthroughthelensofaworldenglishesevaluationset AT janprofant accentsinspeechrecognitionthroughthelensofaworldenglishesevaluationset AT jenniferdrexlerfox accentsinspeechrecognitionthroughthelensofaworldenglishesevaluationset AT quinnmcnamara accentsinspeechrecognitionthroughthelensofaworldenglishesevaluationset AT nishchalbhandari accentsinspeechrecognitionthroughthelensofaworldenglishesevaluationset AT nataliedelworth accentsinspeechrecognitionthroughthelensofaworldenglishesevaluationset AT ilyapirkin accentsinspeechrecognitionthroughthelensofaworldenglishesevaluationset AT migueljette accentsinspeechrecognitionthroughthelensofaworldenglishesevaluationset AT shiprachandra accentsinspeechrecognitionthroughthelensofaworldenglishesevaluationset AT peterha accentsinspeechrecognitionthroughthelensofaworldenglishesevaluationset AT ryanwesterman accentsinspeechrecognitionthroughthelensofaworldenglishesevaluationset |