Improving Accented Speech Recognition with Multi-Domain Training - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

Improving Accented Speech Recognition with Multi-Domain Training

Résumé

Thanks to the rise of self-supervised learning, automatic speech recognition (ASR) systems now achieve near-human performance on a wide variety of datasets. However, they still lack generalization capability and are not robust to domain shifts like accent variations. In this work, we use speech audio representing four different French accents to create fine-tuning datasets that improve the robustness of pre-trained ASR models. By incorporating various accents in the training set, we obtain both in-domain and out-of-domain improvements. Our numerical experiments show that we can reduce error rates by up to 25% (relative) on African and Belgian accents compared to single-domain training while keeping a good performance on standard French.
Fichier principal
Vignette du fichier
2023056083.pdf (212.49 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Licence : CC BY NC ND - Paternité - Pas d'utilisation commerciale - Pas de modification

Dates et versions

hal-04163554 , version 1 (17-07-2023)

Licence

Paternité - Pas d'utilisation commerciale - Pas de modification

Identifiants

Citer

Lucas Maison, Yannick Estève. Improving Accented Speech Recognition with Multi-Domain Training. ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), IEEE, Jun 2023, Rhodes Island, Greece. pp.1-5, ⟨10.1109/ICASSP49357.2023.10096268⟩. ⟨hal-04163554⟩

Collections

UNIV-AVIGNON LIA
4 Consultations
56 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More