Easy Guided Decoding in Providing Suggestions for Interactive Machine Translation

TL;DR

This paper introduces PSGD, a novel constrained decoding algorithm for interactive machine translation that improves translation quality and efficiency without additional training, aiding human translators in post-editing tasks.

Contribution

The paper proposes PSGD, a new decoding method for translation suggestions that outperforms existing methods in quality and speed without requiring extra training.

Findings

PSGD improves BLEU scores by 10.87 and 8.62 on two datasets.

Decoding time is reduced by 63.4% on average.

PSGD outperforms supervised systems trained with TS data.

Abstract

Machine translation technology has made great progress in recent years, but it cannot guarantee error free results. Human translators perform post editing on machine translations to correct errors in the scene of computer aided translation. In favor of expediting the post editing process, many works have investigated machine translation in interactive modes, in which machines can automatically refine the rest of translations constrained by human's edits. Translation Suggestion (TS), as an interactive mode to assist human translators, requires machines to generate alternatives for specific incorrect words or phrases selected by human translators. In this paper, we utilize the parameterized objective function of neural machine translation (NMT) and propose a novel constrained decoding algorithm, namely Prefix Suffix Guided Decoding (PSGD), to deal with the TS problem without additional training. Compared to the state of the art lexically constrained decoding method, PSGD improves translation quality by an average of $10.87$ BLEU and $8.62$ BLEU on the WeTS and the WMT 2022 Translation Suggestion datasets, respectively, and reduces decoding time overhead by an average of 63.4% tested on the WMT translation datasets. Furthermore, on both of the TS benchmark datasets, it is superior to other supervised learning systems trained with TS annotated data.