Paraphrase-Aligned Machine Translation

TL;DR

This paper introduces ParaAlign Translator, a fine-tuning method for LLMs that improves translation quality by aligning sentence structures with target languages, achieving better performance across resource scenarios.

Contribution

The paper presents a novel paraphrase-alignment fine-tuning approach that enhances LLM translation accuracy by structurally aligning sentences with target languages.

Findings

Improves LLaMA-3-8B translation performance in various scenarios.

Achieves parity or surpasses larger models like LLaMA-3-70B.

Enhances translation quality by structural alignment.

Abstract

Large Language Models (LLMs) have demonstrated significant capabilities in machine translation. However, their translation quality is sometimes questioned, as the generated outputs may deviate from expressions typically used by native speakers. These deviations often arise from differences in sentence structure between language systems. To address this issue, we propose ParaAlign Translator, a method that fine-tunes LLMs to paraphrase sentences, aligning their structures with those of the target language systems. This approach improves the performance of subsequent translations. Experimental results demonstrate that the proposed method enhances the LLaMA-3-8B model's performance in both resource-rich and low-resource scenarios and achieves parity with or surpassing the much larger LLaMA-3-70B model.