The paper presents the Second ASR Challenge for Non-native Children’s Speech proposed as a Special Session at Interspeech 2021, following the successful first challenge at Interspeech 2020. The goal of the challenge is to advance research on non-native children’s speech recognition technology, as speech technology still struggles when applied to both children and non-native speakers. The audio data consists of spoken responses provided by L2 students in the context of both English and German speaking proficiency examinations, the latter language added for 2021. Additional training data and a new evaluation set was released for L2 English recorded by speakers of different native languages. Participants could build systems for one or both languages. Each had a closed track where a predetermined set of audio and linguistic resources were selected, and an open track where additional data was allowed. After a description of the released corpora, the paper analyzes the results achieved by the participating systems. Some issues suggested from these results are discussed.
ETLT 2021: Shared Task on Automatic Speech Recognition for Non-Native Children’s Speech
Gretter, R.;Matassoni, Marco
;Falavigna, D.
;
2021-01-01
Abstract
The paper presents the Second ASR Challenge for Non-native Children’s Speech proposed as a Special Session at Interspeech 2021, following the successful first challenge at Interspeech 2020. The goal of the challenge is to advance research on non-native children’s speech recognition technology, as speech technology still struggles when applied to both children and non-native speakers. The audio data consists of spoken responses provided by L2 students in the context of both English and German speaking proficiency examinations, the latter language added for 2021. Additional training data and a new evaluation set was released for L2 English recorded by speakers of different native languages. Participants could build systems for one or both languages. Each had a closed track where a predetermined set of audio and linguistic resources were selected, and an open track where additional data was allowed. After a description of the released corpora, the paper analyzes the results achieved by the participating systems. Some issues suggested from these results are discussed.File | Dimensione | Formato | |
---|---|---|---|
gretter21_interspeech.pdf
solo utenti autorizzati
Tipologia:
Documento in Post-print
Licenza:
NON PUBBLICO - Accesso privato/ristretto
Dimensione
245.76 kB
Formato
Adobe PDF
|
245.76 kB | Adobe PDF | Visualizza/Apri Richiedi una copia |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.