BabyHuBERT: Multilingual Self-Supervised Learning for Segmenting Speakers in Child-Centered Long-Form Recordings

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Charlot, Théo, Kunze, Tarek, Poli, Maxime, Cristia, Alejandrina, Dupoux, Emmanuel, Lavechin, Marvin
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866908867635445760
author Charlot, Théo
Kunze, Tarek
Poli, Maxime
Cristia, Alejandrina
Dupoux, Emmanuel
Lavechin, Marvin
author_facet Charlot, Théo
Kunze, Tarek
Poli, Maxime
Cristia, Alejandrina
Dupoux, Emmanuel
Lavechin, Marvin
contents Child-centered daylong recordings are essential for studying early language development, but existing speech models trained on clean adult data perform poorly due to acoustic and linguistic differences. We introduce BabyHuBERT, a self-supervised speech model trained on 13,000 hours of multilingual child-centered recordings spanning 40+ languages. Evaluated on voice type classification -- distinguishing target children from female adults, male adults, and other children, a key preprocessing step for analyzing naturalistic language experiences -- BabyHuBERT-VTC achieves F1-scores from 52.1% to 74.4% across six corpora, consistently outperforming W2V2-LL4300 (English daylongs) and HuBERT (clean adult speech). Notable gains include 13.2 and 15.9 absolute F1 points over HuBERT on Vanuatu and Solomon Islands, demonstrating effectiveness on underrepresented languages. We share code and model to support researchers working with child-centered recordings across diverse linguistic contexts.
format Preprint
id arxiv_https___arxiv_org_abs_2509_15001
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle BabyHuBERT: Multilingual Self-Supervised Learning for Segmenting Speakers in Child-Centered Long-Form Recordings
Charlot, Théo
Kunze, Tarek
Poli, Maxime
Cristia, Alejandrina
Dupoux, Emmanuel
Lavechin, Marvin
Audio and Speech Processing
Machine Learning
Sound
Child-centered daylong recordings are essential for studying early language development, but existing speech models trained on clean adult data perform poorly due to acoustic and linguistic differences. We introduce BabyHuBERT, a self-supervised speech model trained on 13,000 hours of multilingual child-centered recordings spanning 40+ languages. Evaluated on voice type classification -- distinguishing target children from female adults, male adults, and other children, a key preprocessing step for analyzing naturalistic language experiences -- BabyHuBERT-VTC achieves F1-scores from 52.1% to 74.4% across six corpora, consistently outperforming W2V2-LL4300 (English daylongs) and HuBERT (clean adult speech). Notable gains include 13.2 and 15.9 absolute F1 points over HuBERT on Vanuatu and Solomon Islands, demonstrating effectiveness on underrepresented languages. We share code and model to support researchers working with child-centered recordings across diverse linguistic contexts.
title BabyHuBERT: Multilingual Self-Supervised Learning for Segmenting Speakers in Child-Centered Long-Form Recordings
topic Audio and Speech Processing
Machine Learning
Sound
url https://arxiv.org/abs/2509.15001