Leveraging Audio-Visual Data to Reduce the Multilingual Gap in Self-Supervised Speech Models

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Blandón, María Andrea Cruz, Aldeneh, Zakaria, Chi, Jie, de Seyssel, Maureen
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866909799539539968
author Blandón, María Andrea Cruz
Aldeneh, Zakaria
Chi, Jie
de Seyssel, Maureen
author_facet Blandón, María Andrea Cruz
Aldeneh, Zakaria
Chi, Jie
de Seyssel, Maureen
contents Self-supervised learning (SSL) has made significant advances in speech representation learning. Models like wav2vec 2.0 and HuBERT have achieved state-of-the-art results in tasks such as speech recognition, particularly in monolingual settings. However, multilingual SSL models tend to underperform their monolingual counterparts on each individual language, especially in multilingual scenarios with few languages such as the bilingual setting. In this work, we investigate a novel approach to reduce this performance gap by introducing limited visual grounding into bilingual speech SSL models. Our results show that visual grounding benefits both monolingual and bilingual models, with especially pronounced gains for the latter, reducing the multilingual performance gap on zero-shot phonetic discrimination from 31.5% for audio-only models to 8.04% with grounding.
format Preprint
id arxiv_https___arxiv_org_abs_2509_17523
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Leveraging Audio-Visual Data to Reduce the Multilingual Gap in Self-Supervised Speech Models
Blandón, María Andrea Cruz
Aldeneh, Zakaria
Chi, Jie
de Seyssel, Maureen
Computation and Language
Audio and Speech Processing
Self-supervised learning (SSL) has made significant advances in speech representation learning. Models like wav2vec 2.0 and HuBERT have achieved state-of-the-art results in tasks such as speech recognition, particularly in monolingual settings. However, multilingual SSL models tend to underperform their monolingual counterparts on each individual language, especially in multilingual scenarios with few languages such as the bilingual setting. In this work, we investigate a novel approach to reduce this performance gap by introducing limited visual grounding into bilingual speech SSL models. Our results show that visual grounding benefits both monolingual and bilingual models, with especially pronounced gains for the latter, reducing the multilingual performance gap on zero-shot phonetic discrimination from 31.5% for audio-only models to 8.04% with grounding.
title Leveraging Audio-Visual Data to Reduce the Multilingual Gap in Self-Supervised Speech Models
topic Computation and Language
Audio and Speech Processing
url https://arxiv.org/abs/2509.17523