Learning Translation Quality Evaluation on Low Resource Languages from Large Language Models

TL;DR

This paper proposes a method to enhance translation quality evaluation for low-resource languages by distilling knowledge from Large Language Models, creating synthetic datasets that improve learned metrics like BLEURT without needing human annotations.

Contribution

It introduces a novel approach to improve learned translation metrics for low-resource languages using knowledge distillation from LLMs and synthetic data generation.

Findings

Improved BLEURT-like model performance on low-resource languages

Synthetic datasets effectively enhance translation quality evaluation

Method reduces reliance on costly human annotations

Abstract

Learned metrics such as BLEURT have in recent years become widely employed to evaluate the quality of machine translation systems. Training such metrics requires data which can be expensive and difficult to acquire, particularly for lower-resource languages. We show how knowledge can be distilled from Large Language Models (LLMs) to improve upon such learned metrics without requiring human annotators, by creating synthetic datasets which can be mixed into existing datasets, requiring only a corpus of text in the target language. We show that the performance of a BLEURT-like model on lower resource languages can be improved in this way.