Modeling Dual Read/Write Paths for Simultaneous Machine Translation

TL;DR

This paper introduces a dual-path approach for simultaneous machine translation that enforces path duality constraints, enabling joint optimization of source-to-target and target-to-source models to improve translation quality across various latency levels.

Contribution

It proposes a novel dual-path SiMT method with duality constraints that align read/write paths, enhancing model training and performance.

Findings

Outperforms strong baselines across all latency levels

Effective joint optimization of dual models via path duality

Improved translation quality on En-Vi and De-En tasks

Abstract

Simultaneous machine translation (SiMT) outputs translation while reading source sentence and hence requires a policy to decide whether to wait for the next source word (READ) or generate a target word (WRITE), the actions of which form a read/write path. Although the read/write path is essential to SiMT performance, no direct supervision is given to the path in the existing methods. In this paper, we propose a method of dual-path SiMT which introduces duality constraints to direct the read/write path. According to duality constraints, the read/write path in source-to-target and target-to-source SiMT models can be mapped to each other. As a result, the two SiMT models can be optimized jointly by forcing their read/write paths to satisfy the mapping. Experiments on En-Vi and De-En tasks show that our method can outperform strong baselines under all latency.