# Data-Driven Adaptive Simultaneous Machine Translation

arXiv: https://arxiv.org/abs/2204.12672 · HTML: https://mlsanity.com/paper/2204.12672/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 10 of its 15 references that are in the map.
- First posted on arXiv 27 Apr 2022.

## Builds on (most cited first, 10 of 10)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Neural Machine Translation by Jointly Learning to Align and Translate](https://mlsanity.com/paper/1409.0473/) (2014)
- [Neural Machine Translation of Rare Words with Subword Units](https://mlsanity.com/paper/1508.07909/) (2015)
- [Monotonic Infinite Lookback Attention for Simultaneous Machine Translation](https://mlsanity.com/paper/1906.05218/) (2019)
- [Monotonic Multihead Attention](https://mlsanity.com/paper/1909.12406/) (2019)
- [SimulMT to SimulST: Adapting Simultaneous Text Translation to End-to-End Simultaneous Speech Translation](https://mlsanity.com/paper/2011.02048/) (2020)
- [Simultaneous Translation Policies: From Fixed to Adaptive](https://mlsanity.com/paper/2004.13169/) (2020)
- [Low-Latency Neural Speech Translation](https://mlsanity.com/paper/1808.00491/) (2018)
- [Re-translation versus Streaming for Simultaneous Translation](https://mlsanity.com/paper/2004.03643/) (2020)
- [Simultaneous Translation with Flexible Policy via Restricted Imitation Learning](https://mlsanity.com/paper/1906.01135/) (2019)

## Abstract (arXiv)

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.