Can Synthetic Translations Improve Bitext Quality?

TL;DR

This paper investigates how synthetic translations can enhance the quality of mined bitext by replacing imperfect references, leading to better performance in translation tasks without extra bilingual data.

Contribution

It introduces a method to use synthetic translations for revising bitext, improving quality without additional supervision, validated through human and task-based evaluations.

Findings

Synthetic translations can improve bitext quality.

Replacing original references with synthetic ones enhances translation performance.

The approach reduces noise in mined bilingual data.

Abstract

Synthetic translations have been used for a wide range of NLP tasks primarily as a means of data augmentation. This work explores, instead, how synthetic translations can be used to revise potentially imperfect reference translations in mined bitext. We find that synthetic samples can improve bitext quality without any additional bilingual supervision when they replace the originals based on a semantic equivalence classifier that helps mitigate NMT noise. The improved quality of the revised bitext is confirmed intrinsically via human evaluation and extrinsically through bilingual induction and MT tasks.