Speech-to-Text Translation with Phoneme-Augmented CoT: Enhancing Cross-Lingual Transfer in Low-Resource Scenarios

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Gállego, Gerard I., Pareras, Oriol, Garcia, Martí Cortada, Takanori, Lucas, Hernando, Javier
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914063553921024
author Gállego, Gerard I.
Pareras, Oriol
Garcia, Martí Cortada
Takanori, Lucas
Hernando, Javier
author_facet Gállego, Gerard I.
Pareras, Oriol
Garcia, Martí Cortada
Takanori, Lucas
Hernando, Javier
contents We propose a Speech-to-Text Translation (S2TT) approach that integrates phoneme representations into a Chain-of-Thought (CoT) framework to improve translation in low-resource and zero-resource settings. By introducing phoneme recognition as an intermediate step, we enhance cross-lingual transfer, enabling translation even for languages with no labeled speech data. Our system builds on a multilingual LLM, which we extend to process speech and phonemes. Training follows a curriculum learning strategy that progressively introduces more complex tasks. Experiments on multilingual S2TT benchmarks show that phoneme-augmented CoT improves translation quality in low-resource conditions and enables zero-resource translation, while slightly impacting high-resource performance. Despite this trade-off, our findings demonstrate that phoneme-based CoT is a promising step toward making S2TT more accessible across diverse languages.
format Preprint
id arxiv_https___arxiv_org_abs_2505_24691
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Speech-to-Text Translation with Phoneme-Augmented CoT: Enhancing Cross-Lingual Transfer in Low-Resource Scenarios
Gállego, Gerard I.
Pareras, Oriol
Garcia, Martí Cortada
Takanori, Lucas
Hernando, Javier
Computation and Language
Sound
Audio and Speech Processing
We propose a Speech-to-Text Translation (S2TT) approach that integrates phoneme representations into a Chain-of-Thought (CoT) framework to improve translation in low-resource and zero-resource settings. By introducing phoneme recognition as an intermediate step, we enhance cross-lingual transfer, enabling translation even for languages with no labeled speech data. Our system builds on a multilingual LLM, which we extend to process speech and phonemes. Training follows a curriculum learning strategy that progressively introduces more complex tasks. Experiments on multilingual S2TT benchmarks show that phoneme-augmented CoT improves translation quality in low-resource conditions and enables zero-resource translation, while slightly impacting high-resource performance. Despite this trade-off, our findings demonstrate that phoneme-based CoT is a promising step toward making S2TT more accessible across diverse languages.
title Speech-to-Text Translation with Phoneme-Augmented CoT: Enhancing Cross-Lingual Transfer in Low-Resource Scenarios
topic Computation and Language
Sound
Audio and Speech Processing
url https://arxiv.org/abs/2505.24691