| Texto completo | |
| Autor(es): Mostrar menos - |
Candido Junior, Arnaldo
;
Casanova, Edresson
;
Soares, Anderson
;
de Oliveira, Frederico Santos
;
Oliveira, Lucas
;
Fernandes Junior, Ricardo Corso
;
Pinto da Silva, Daniel Peixoto
;
Fayet, Fernando Gorgulho
;
Carlotto, Bruno Baldissera
;
Stefanel Gris, Lucas Rafael
;
Aluisio, Sandra Maria
Número total de Autores: 11
|
| Tipo de documento: | Artigo Científico |
| Fonte: | Language Resources and Evaluation; v. N/A, p. 33-pg., 2022-11-21. |
| Resumo | |
Automatic Speech recognition (ASR) is a complex and challenging task. In recent years, there have been significant advances in the area. In particular, for the Brazilian Portuguese (BP) language, there were around 376 h publicly available for the ASR task until the second half of 2020. With the release of new datasets in early 2021, this number increased to 574 h. The existing resources, however, are composed of audios containing only read and prepared speech. There is a lack of datasets including spontaneous speech, which are essential in several ASR applications. This paper presents CORAA (Corpus of Annotated Audios) ASR with 290 h, a publicly available dataset for ASR in BP containing validated pairs of audio-transcription. CORAA ASR also contains European Portuguese audios (4.6 h). We also present a public ASR model based on Wav2Vec 2.0 XLSR-53, fine-tuned over CORAA ASR. Our model achieved a Word Error Rate (WER) of 24.18% on CORAA ASR test set and 20.08% on Common Voice test set. When measuring the Character Error Rate (CER), we obtained 11.02% and 6.34% for CORAA ASR and Common Voice, respectively. CORAA ASR corpora were assembled to both improve ASR models in BP with phenomena from spontaneous speech and motivate young researchers to start their studies on ASR for Portuguese. All the corpora are publicly available at under the CC BY-NC-ND 4.0 license. (AU) | |
| Processo FAPESP: | 19/07665-4 - Centro de Inteligência Artificial |
| Beneficiário: | Fabio Gagliardi Cozman |
| Modalidade de apoio: | Auxílio à Pesquisa - Programa eScience e Data Science - Centros de Pesquisa Aplicada |