Re-translation versus Streaming for Simultaneous Translation

Proceedings of the 17th International Conference on Spoken Language Translation (IWSLT), pages 220–227 July 9-10, 2020. c⃝2020 Association for Computational Linguistics https://doi.org/10.18653/v1/P17 220 Re-translation versus Streaming for Simultaneous Translation Naveen Arivazhagan∗, Colin Cherry∗, Wolfgang Macherey and George Foster Google Research {navari,colincherry,wmach,fosterg}@google.com Abstract There has been great progress in improving streaming machine translation, a simultane- ous paradigm where the system appends to a growing hypothesis as more source con- tent becomes available. We study a related problem in which revisions to the hypothe- sis beyond strictly appending words are per- mitted. This is suitable for applications such as live captioning an audio feed. In this setting, we compare custom streaming ap- proaches to re-translation, a straightforward strategy where each new source token triggers a distinct translation from scratch. We find re- translation to be as good or better than state- of-the-art streaming systems, even when op- erating under constraints that allow very few revisions. We attribute much of this success to a previously proposed data-augmentation technique that adds prefix-pairs to the training data, which alongside wait-k inference forms a strong baseline for streaming translation. We also highlight re-translation’s ability to wrap arbitrarily powerful MT systems with an ex- periment showing large improvements from an upgrade to its base model. 1 Introduction In simultaneous machine translation, the goal is to translate an incoming stream of source words with as low latency as possible. A typical application is speech translation, where we often assume the eventual output modality to also be speech. In a speech-to-speech scenario, target words must be appended to existing output with no possibility for revision. The corresponding translation task, which we refer to as streaming translation, has received considerable recent attention, generating custom approaches designed to maximize quality and min- imize latency (Cho and Esipova, 2016; Gu et al., ∗Equal contributions 2017; Dalvi et al., 2018; Ma et al., 2019a). How- ever, for applications where the output modality is text, such as live captioning, the prohibition against revising output is overly stringent. The ability to revise previous partial translations makes simply re-translating each successive source prefix a viable strategy. Compared to streaming models, re-translation has the advantage of low la- tency, since it always attempts a translation of the complete source prefix, and high final-translation quality, since it is not restricted to preserving pre- vious output. It has the disadvantages of higher computational cost, and a high revision rate, vis- ible as textual instability in an online translation display. When revisions are an option, it is unclear whether one should prefer a specialized streaming model or a re-translation strategy. In light of this, we make the following con- tributions: (1) We evaluate a combination of re- translation techniques that have not previously been studied together. (2) We provide the first empirical comparison of re-translation and streaming mod- els, demonstrating that re-translation operating in a very low-revision regime can match or beat the quality-latency trade-offs of streaming models. (3) We test a 0-revision configuration of re-translation, and show that it …

Content truncated to 500 words