arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.11072cs.CLcs.SDeess.AS
收藏

无需对齐数据的同时语音到语音翻译

Simultaneous Speech-to-Speech Translation Without Aligned Data

LIA, University of Avignon, France(LIA,阿维尼翁大学,法国)

Tom Labiausse, Romain Fabre, Yannick Estève, Alexandre Défossez, Neil Zeghidour

AI总结:

Hibiki-Zero无需词级对齐数据,通过句子级对齐和强化学习策略实现高精度的同时语音翻译,支持多种语言并提升翻译质量与效率。

摘要:

同时语音翻译需要在实时情况下将源语音翻译成目标语言,同时处理非单调的词依赖关系。传统方法依赖于监督训练使用词级对齐数据,这种数据难以大规模收集,因此依赖于使用语言特定的启发式方法生成合成对齐,效果不佳。我们提出了Hibiki-Zero,完全消除了词级对齐的需求。这从根本上简化了训练流程,并使模型能够无缝扩展到具有不同语法结构的各种语言,从而移除了设计语言特定对齐启发式方法的瓶颈。我们首先在句子级对齐数据上进行训练,以学习高延迟的语音翻译,然后应用一种新的强化学习策略,使用GRPO来优化延迟同时保持翻译质量。Hibiki-Zero在五个X到英语任务中实现了翻译准确性、延迟、语音转换和自然度方面的最佳性能。此外,我们证明我们的模型可以适应支持新输入语言,仅需不到1000小时的语音数据。我们提供了示例、模型权重、推断代码,并发布了一个包含45小时多语言数据的基准,用于语音翻译评估。

英文摘要:

Simultaneous speech translation requires translating source speech into a target language in real-time while handling non-monotonic word dependencies. Traditional approaches rely on supervised training with word-level aligned data, which is difficult to collect at scale and thus depends on synthetic alignments using language-specific heuristics that are suboptimal. We propose Hibiki-Zero, which eliminates the need for word-level alignments entirely. This fundamentally simplifies the training pipeline and enables seamless scaling to diverse languages with varying grammatical structures, removing the bottleneck of designing language-specific alignment heuristics. We first train on sentence-level aligned data to learn speech translation at high latency, then apply a novel reinforcement learning strategy using GRPO to optimize latency while preserving translation quality. Hibiki-Zero achieves state-of-the-art performance in translation accuracy, latency, voice transfer, and naturalness across five X-to-English tasks. Moreover, we demonstrate that our model can be adapted to support a new input language with less than 1000h of speech. We provide examples, model weights, inference code and we release a benchmark containing 45h of multilingual data for speech translation evaluation.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑