Guardado en:
Detalles Bibliográficos
Autores principales: Liang, Siyu, Ballier, Nicolas, Levow, Gina-Anne, Wright, Richard
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:https://arxiv.org/abs/2509.25516
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866911184170516480
author Liang, Siyu
Ballier, Nicolas
Levow, Gina-Anne
Wright, Richard
author_facet Liang, Siyu
Ballier, Nicolas
Levow, Gina-Anne
Wright, Richard
contents While large multilingual automatic speech recognition (ASR) models achieve remarkable performance, the internal mechanisms of the end-to-end pipeline, particularly concerning fairness and efficacy across languages, remain underexplored. This paper introduces a fine-grained analysis of Whisper's multilingual decoder, examining its sub-token hypotheses during transcription across languages with various resource levels. Our method traces the beam search path, capturing sub-token guesses and their associated probabilities. Results reveal that higher resource languages benefit from higher likelihood of the correct token being top-ranked, greater confidence, lower predictive entropy, and more diverse alternative candidates. Lower resource languages fare worse on these metrics, but also exhibit distinct clustering patterns in sub-token usage sometimes influenced by typology in our PCA and t-SNE analysis. This sub-token probing uncovers systematic decoding disparities masked by aggregate error rates and points towards targeted interventions to ameliorate the imbalanced development of speech technology.
format Preprint
id arxiv_https___arxiv_org_abs_2509_25516
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Beyond WER: Probing Whisper's Sub-token Decoder Across Diverse Language Resource Levels
Liang, Siyu
Ballier, Nicolas
Levow, Gina-Anne
Wright, Richard
Computation and Language
While large multilingual automatic speech recognition (ASR) models achieve remarkable performance, the internal mechanisms of the end-to-end pipeline, particularly concerning fairness and efficacy across languages, remain underexplored. This paper introduces a fine-grained analysis of Whisper's multilingual decoder, examining its sub-token hypotheses during transcription across languages with various resource levels. Our method traces the beam search path, capturing sub-token guesses and their associated probabilities. Results reveal that higher resource languages benefit from higher likelihood of the correct token being top-ranked, greater confidence, lower predictive entropy, and more diverse alternative candidates. Lower resource languages fare worse on these metrics, but also exhibit distinct clustering patterns in sub-token usage sometimes influenced by typology in our PCA and t-SNE analysis. This sub-token probing uncovers systematic decoding disparities masked by aggregate error rates and points towards targeted interventions to ameliorate the imbalanced development of speech technology.
title Beyond WER: Probing Whisper's Sub-token Decoder Across Diverse Language Resource Levels
topic Computation and Language
url https://arxiv.org/abs/2509.25516