Data-Driven Adaptive Simultaneous Machine Translation
This work improves simultaneous translation for real-time applications, but it is incremental as it builds on existing wait-k methods.
The paper tackled the problem of simultaneous machine translation by addressing the limitations of the fixed wait-k policy, proposing an efficient training scheme that achieves better translation quality and latency compared to strong baselines on two language pairs.
In simultaneous translation (SimulMT), the most widely used strategy is the wait-k policy thanks to its simplicity and effectiveness in balancing translation quality and latency. However, wait-k suffers from two major limitations: (a) it is a fixed policy that can not adaptively adjust latency given context, and (b) its training is much slower than full-sentence translation. To alleviate these issues, we propose a novel and efficient training scheme for adaptive SimulMT by augmenting the training corpus with adaptive prefix-to-prefix pairs, while the training complexity remains the same as that of training full-sentence translation models. Experiments on two language pairs show that our method outperforms all strong baselines in terms of translation quality and latency.