动态滞后用于同声翻译

中文摘要 AI 辅助

在级联式同声语音翻译中，机器翻译（MT）系统无法控制上游识别器的读写调度：它必须从不断增长的源语言前缀中决定要提交多少目标语言文本。我们通过在一个稳定的前缀——即任何翻译到当前部分源语言时与模型自身全源输出共享的最长前缀——与完整句子对混合的数据上微调，使一个经过句子训练的、仅解码器的LLM具备前缀感知能力，并通过一次强制解码轮次进行提示，该轮次在更多源语言到达时将已提交的目标语言向前推进，从而使系统在构造上无闪烁。我们对Qwen3-8B进行了英语到德语、日语和中文的微调，使用参考转录前缀模拟源语言流。前缀微调保持了完整句子的质量，同时提高了最差位置的分块质量，并改善了令牌级提交置信度的校准，与稳定前缀预言机相比，降低了早期源语言前缀上的期望校准误差（ECE）。在该置信度上使用一个无需训练的单一阈值是我们比较的三种延迟控制中最有效的方法：它描绘了一条连续的质-延迟前沿，优于离散的等待k和目标后缀删除的质-延迟权衡机制。该效果在FLEURS、WMT24++和CoVoST 2测试集上，在COMET和MetricX指标下均表现良好。

英文摘要

In streaming simultaneous speech translation, the speech translation system is trained to learn a read-write policy that alternates between consuming source words and generating target ones. In a cascaded setting, the output from the speech recognizer is passed to a separate machine translation component, making it more difficult to learn such a policy. Approximations such as fixed wait-k strategies or target-suffix deletion can be employed, but these approaches do not provide the model with a streaming system's flexibility to make contextual read-write decisions. This paper presents a training strategy for a cascaded machine translation system that enables it to dynamically decide how much of the growing source prefix to translate. We achieve this by fine-tuning a large language model (Qwen3-8B) on stable prefixes of the training data, which are produced by pairing every source sentence prefix in the training data with the longest translation of that prefix that is shared with the full source sentence translation. We fine-tune variants of the model on different subsets of the prefixes and compare against wait-k and target-suffix deletion. We also investigate the effect of fine-tuning the target-token generation confidence. Our experiments show that stable prefixes improve the quality-latency tradeoff when translating from English into German, Japanese, and Chinese across a range of test sets.

发表机构

• Microsoft（微软）

机构由 AI 辅助整理，请以论文原文为准。