Learning Optimal Policy for Simultaneous Machine Translation via Binary Search
This addresses the challenge of real-time translation for users needing low-latency outputs, but it is incremental as it builds on existing SiMT methods with a novel policy learning approach.
The paper tackles the problem of learning a precise translation policy for simultaneous machine translation to balance latency and quality, and the result is a method that exceeds strong baselines across all latency scenarios in experiments on four translation tasks.
Simultaneous machine translation (SiMT) starts to output translation while reading the source sentence and needs a precise policy to decide when to output the generated translation. Therefore, the policy determines the number of source tokens read during the translation of each target token. However, it is difficult to learn a precise translation policy to achieve good latency-quality trade-offs, because there is no golden policy corresponding to parallel sentences as explicit supervision. In this paper, we present a new method for constructing the optimal policy online via binary search. By employing explicit supervision, our approach enables the SiMT model to learn the optimal policy, which can guide the model in completing the translation during inference. Experiments on four translation tasks show that our method can exceed strong baselines across all latency scenarios.