Simultaneous Translation Policies: From Fixed to Adaptive
This work addresses the trade-off between quality and latency in simultaneous translation, offering a practical solution for real-time applications, though it is incremental as it builds on existing fixed policies.
The paper tackled the problem of simultaneous translation by proposing adaptive policies that balance translation quality and latency, achieving up to 4 BLEU points improvement over fixed policies and surpassing full-sentence translation in greedy mode with lower latency.
Adaptive policies are better than fixed policies for simultaneous translation, since they can flexibly balance the tradeoff between translation quality and latency based on the current context information. But previous methods on obtaining adaptive policies either rely on complicated training process, or underperform simple fixed policies. We design an algorithm to achieve adaptive policies via a simple heuristic composition of a set of fixed policies. Experiments on Chinese -> English and German -> English show that our adaptive policies can outperform fixed ones by up to 4 BLEU points for the same latency, and more surprisingly, it even surpasses the BLEU score of full-sentence translation in the greedy mode (and very close to beam mode), but with much lower latency.