LexMatcher: Dictionary-centric Data Collection for LLM-based Machine Translation

TL;DR

LexMatcher is a dictionary-driven data collection method that improves LLM-based machine translation by enhancing sense coverage and handling polysemy, leading to better performance on standard benchmarks.

Contribution

This paper introduces LexMatcher, a novel data curation approach based on bilingual dictionaries, specifically designed for instruction fine-tuning of LLMs in machine translation.

Findings

Outperforms baselines on WMT2022 test sets

Improves word sense disambiguation and terminology translation

Effective in handling infrequent senses of polysemous words

Abstract

The fine-tuning of open-source large language models (LLMs) for machine translation has recently received considerable attention, marking a shift towards data-centric research from traditional neural machine translation. However, the area of data collection for instruction fine-tuning in machine translation remains relatively underexplored. In this paper, we present LexMatcher, a simple yet effective method for data curation, the design of which is driven by the coverage of senses found in bilingual dictionaries. The construction process comprises data retrieval from an existing corpus and data augmentation that supplements the infrequent senses of polysemous words. Utilizing LLaMA2 as our base model, our approach outperforms the established baselines on the WMT2022 test sets and also exhibits remarkable performance in tasks related to word sense disambiguation and specialized terminology translation. These results underscore the effectiveness of LexMatcher in enhancing LLM-based machine translation. The code, data, and models are available at https://github.com/ARIES-LM/Lexmatcher-MT.git.