RosettaSpeech: Zero-Shot Speech-to-Speech Translation without Parallel Speech

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zheng, Zhisheng, Sun, Xiaohang, Dinh, Tuan, Yanamandra, Abhishek, Jain, Abhinav, Liu, Zhu, Hadap, Sunil, Bhat, Vimal, Aggarwal, Manoj, Medioni, Gerard, Harwath, David
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910022798147584
author Zheng, Zhisheng
Sun, Xiaohang
Dinh, Tuan
Yanamandra, Abhishek
Jain, Abhinav
Liu, Zhu
Hadap, Sunil
Bhat, Vimal
Aggarwal, Manoj
Medioni, Gerard
Harwath, David
author_facet Zheng, Zhisheng
Sun, Xiaohang
Dinh, Tuan
Yanamandra, Abhishek
Jain, Abhinav
Liu, Zhu
Hadap, Sunil
Bhat, Vimal
Aggarwal, Manoj
Medioni, Gerard
Harwath, David
contents End-to-end speech-to-speech translation (S2ST) systems typically struggle with a critical data bottleneck: the scarcity of parallel speech-to-speech corpora. To overcome this, we introduce RosettaSpeech, a novel zero-shot framework trained exclusively on monolingual speech-text data augmented by machine translation supervision. Unlike prior works that rely on complex cascaded pseudo-labeling, our approach strategically utilizes text as a semantic bridge during training to synthesize translation targets, thereby eliminating the need for parallel speech pairs while maintaining a direct, end-to-end inference pipeline. Empirical evaluations on the CVSS-C benchmark demonstrate that RosettaSpeech achieves state-of-the-art zero-shot performance, surpassing leading baselines by significant margins - achieving ASR-BLEU scores of 25.17 for German-to-English (+27% relative gain) and 29.86 for Spanish-to-English (+14%). Crucially, our model effectively preserves the source speaker's voice without ever seeing paired speech data. We further analyze the impact of data scaling and demonstrate the model's capability in many-to-one translation, offering a scalable solution for extending high-quality S2ST to "text-rich, speech-poor" languages.
format Preprint
id arxiv_https___arxiv_org_abs_2511_20974
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle RosettaSpeech: Zero-Shot Speech-to-Speech Translation without Parallel Speech
Zheng, Zhisheng
Sun, Xiaohang
Dinh, Tuan
Yanamandra, Abhishek
Jain, Abhinav
Liu, Zhu
Hadap, Sunil
Bhat, Vimal
Aggarwal, Manoj
Medioni, Gerard
Harwath, David
Audio and Speech Processing
Computation and Language
Machine Learning
End-to-end speech-to-speech translation (S2ST) systems typically struggle with a critical data bottleneck: the scarcity of parallel speech-to-speech corpora. To overcome this, we introduce RosettaSpeech, a novel zero-shot framework trained exclusively on monolingual speech-text data augmented by machine translation supervision. Unlike prior works that rely on complex cascaded pseudo-labeling, our approach strategically utilizes text as a semantic bridge during training to synthesize translation targets, thereby eliminating the need for parallel speech pairs while maintaining a direct, end-to-end inference pipeline. Empirical evaluations on the CVSS-C benchmark demonstrate that RosettaSpeech achieves state-of-the-art zero-shot performance, surpassing leading baselines by significant margins - achieving ASR-BLEU scores of 25.17 for German-to-English (+27% relative gain) and 29.86 for Spanish-to-English (+14%). Crucially, our model effectively preserves the source speaker's voice without ever seeing paired speech data. We further analyze the impact of data scaling and demonstrate the model's capability in many-to-one translation, offering a scalable solution for extending high-quality S2ST to "text-rich, speech-poor" languages.
title RosettaSpeech: Zero-Shot Speech-to-Speech Translation without Parallel Speech
topic Audio and Speech Processing
Computation and Language
Machine Learning
url https://arxiv.org/abs/2511.20974