Fluent and Low-latency Simultaneous Speech-to-Speech Translation with Self-adaptive Training

TL;DR

This paper introduces Self-Adaptive Translation (SAT), a novel method for simultaneous speech-to-speech translation that dynamically adjusts translation length to reduce latency and improve fluency across different speech rates.

Contribution

The paper presents a new self-adaptive training approach that effectively manages translation timing, addressing the limitations of existing single-sentence focused methods.

Findings

SAT achieves lower latency than baselines at similar translation quality.

SAT produces more natural and fluent target speech.

The method is effective in both Zh <-> En translation directions.

Abstract

Simultaneous speech-to-speech translation is widely useful but extremely challenging, since it needs to generate target-language speech concurrently with the source-language speech, with only a few seconds delay. In addition, it needs to continuously translate a stream of sentences, but all recent solutions merely focus on the single-sentence scenario. As a result, current approaches accumulate latencies progressively when the speaker talks faster, and introduce unnatural pauses when the speaker talks slower. To overcome these issues, we propose Self-Adaptive Translation (SAT) which flexibly adjusts the length of translations to accommodate different source speech rates. At similar levels of translation quality (as measured by BLEU), our method generates more fluent target speech (as measured by the naturalness metric MOS) with substantially lower latency than the baseline, in both Zh <-> En directions.