Project Euphonia: advancing inclusive speech recognition through expanded data collection and evaluation
Speech recognition models, predominantly trained on standard speech, often exhibit lower accuracy for individuals with accents, dialects, or speech impairments. This disparity is particularly pronounced for economically or socially marginalized communities, including those with disabilities or diver...
Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2025-06-01
|
| Series: | Frontiers in Language Sciences |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/flang.2025.1569448/full |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849689365371944960 |
|---|---|
| author | Alicia Martin Robert L. MacDonald Pan-Pan Jiang Marilyn Ladewig Julie Cattiau Rus Heywood Richard Cave Jimmy Tobin Philip C. Nelson Katrin Tomanek |
| author_facet | Alicia Martin Robert L. MacDonald Pan-Pan Jiang Marilyn Ladewig Julie Cattiau Rus Heywood Richard Cave Jimmy Tobin Philip C. Nelson Katrin Tomanek |
| author_sort | Alicia Martin |
| collection | DOAJ |
| description | Speech recognition models, predominantly trained on standard speech, often exhibit lower accuracy for individuals with accents, dialects, or speech impairments. This disparity is particularly pronounced for economically or socially marginalized communities, including those with disabilities or diverse linguistic backgrounds. Project Euphonia, a Google initiative originally launched in English dedicated to improving Automatic Speech Recognition (ASR) of disordered speech, is expanding its data collection and evaluation efforts to include international languages like Spanish, Japanese, French and Hindi, in a continued effort to enhance inclusivity. This paper presents an overview of the extension of processes and methods used for English data collection to more languages and locales, progress on the collected data, and details about our model evaluation process, focusing on meaning preservation based on Generative AI. |
| format | Article |
| id | doaj-art-97199d5cd20d4962a99e4a2ff7614f0c |
| institution | DOAJ |
| issn | 2813-4605 |
| language | English |
| publishDate | 2025-06-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Language Sciences |
| spelling | doaj-art-97199d5cd20d4962a99e4a2ff7614f0c2025-08-20T03:21:39ZengFrontiers Media S.A.Frontiers in Language Sciences2813-46052025-06-01410.3389/flang.2025.15694481569448Project Euphonia: advancing inclusive speech recognition through expanded data collection and evaluationAlicia Martin0Robert L. MacDonald1Pan-Pan Jiang2Marilyn Ladewig3Julie Cattiau4Rus Heywood5Richard Cave6Jimmy Tobin7Philip C. Nelson8Katrin Tomanek9Google Research, Mountain View, CA, United StatesGoogle Research, Mountain View, CA, United StatesGoogle Research, Mountain View, CA, United StatesCP Unlimited, New York, NY, United StatesGoogle Research, Mountain View, CA, United StatesGoogle Research, Mountain View, CA, United StatesComputer Science Department, University College London (UCL), London, United KingdomGoogle Research, Mountain View, CA, United StatesGoogle Research, Mountain View, CA, United StatesGoogle Research, Mountain View, CA, United StatesSpeech recognition models, predominantly trained on standard speech, often exhibit lower accuracy for individuals with accents, dialects, or speech impairments. This disparity is particularly pronounced for economically or socially marginalized communities, including those with disabilities or diverse linguistic backgrounds. Project Euphonia, a Google initiative originally launched in English dedicated to improving Automatic Speech Recognition (ASR) of disordered speech, is expanding its data collection and evaluation efforts to include international languages like Spanish, Japanese, French and Hindi, in a continued effort to enhance inclusivity. This paper presents an overview of the extension of processes and methods used for English data collection to more languages and locales, progress on the collected data, and details about our model evaluation process, focusing on meaning preservation based on Generative AI.https://www.frontiersin.org/articles/10.3389/flang.2025.1569448/fulldisordered speechautomatic speech recognitionspeech data collectiondysarthriaartificial intelligence |
| spellingShingle | Alicia Martin Robert L. MacDonald Pan-Pan Jiang Marilyn Ladewig Julie Cattiau Rus Heywood Richard Cave Jimmy Tobin Philip C. Nelson Katrin Tomanek Project Euphonia: advancing inclusive speech recognition through expanded data collection and evaluation Frontiers in Language Sciences disordered speech automatic speech recognition speech data collection dysarthria artificial intelligence |
| title | Project Euphonia: advancing inclusive speech recognition through expanded data collection and evaluation |
| title_full | Project Euphonia: advancing inclusive speech recognition through expanded data collection and evaluation |
| title_fullStr | Project Euphonia: advancing inclusive speech recognition through expanded data collection and evaluation |
| title_full_unstemmed | Project Euphonia: advancing inclusive speech recognition through expanded data collection and evaluation |
| title_short | Project Euphonia: advancing inclusive speech recognition through expanded data collection and evaluation |
| title_sort | project euphonia advancing inclusive speech recognition through expanded data collection and evaluation |
| topic | disordered speech automatic speech recognition speech data collection dysarthria artificial intelligence |
| url | https://www.frontiersin.org/articles/10.3389/flang.2025.1569448/full |
| work_keys_str_mv | AT aliciamartin projecteuphoniaadvancinginclusivespeechrecognitionthroughexpandeddatacollectionandevaluation AT robertlmacdonald projecteuphoniaadvancinginclusivespeechrecognitionthroughexpandeddatacollectionandevaluation AT panpanjiang projecteuphoniaadvancinginclusivespeechrecognitionthroughexpandeddatacollectionandevaluation AT marilynladewig projecteuphoniaadvancinginclusivespeechrecognitionthroughexpandeddatacollectionandevaluation AT juliecattiau projecteuphoniaadvancinginclusivespeechrecognitionthroughexpandeddatacollectionandevaluation AT rusheywood projecteuphoniaadvancinginclusivespeechrecognitionthroughexpandeddatacollectionandevaluation AT richardcave projecteuphoniaadvancinginclusivespeechrecognitionthroughexpandeddatacollectionandevaluation AT jimmytobin projecteuphoniaadvancinginclusivespeechrecognitionthroughexpandeddatacollectionandevaluation AT philipcnelson projecteuphoniaadvancinginclusivespeechrecognitionthroughexpandeddatacollectionandevaluation AT katrintomanek projecteuphoniaadvancinginclusivespeechrecognitionthroughexpandeddatacollectionandevaluation |