Improving Stability in Simultaneous Speech Translation: A Revision-Controllable Decoding Approach

TL;DR

This paper introduces a revision-controllable decoding method for simultaneous speech translation that significantly reduces flickering in partial results, enhancing stability without sacrificing translation quality.

Contribution

It proposes a novel revision window mechanism within beam search to control and eliminate flickering in real-time speech translation.

Findings

Substantial flickering reduction demonstrated in experiments

Translation quality remains largely unaffected

Method provides complete flickering elimination capability

Abstract

Simultaneous Speech-to-Text translation serves a critical role in real-time crosslingual communication. Despite the advancements in recent years, challenges remain in achieving stability in the translation process, a concern primarily manifested in the flickering of partial results. In this paper, we propose a novel revision-controllable method designed to address this issue. Our method introduces an allowed revision window within the beam search pruning process to screen out candidate translations likely to cause extensive revisions, leading to a substantial reduction in flickering and, crucially, providing the capability to completely eliminate flickering. The experiments demonstrate the proposed method can significantly improve the decoding stability without compromising substantially on the translation quality.