Saved in:
Bibliographic Details
Main Authors: Chang, Ke-Ching, Chen, Chung-Chi, Yen, An-Zi
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2412.05916
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917861501435904
author Chang, Ke-Ching
Chen, Chung-Chi
Yen, An-Zi
author_facet Chang, Ke-Ching
Chen, Chung-Chi
Yen, An-Zi
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2412_05916
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Paraphrase-Aligned Machine Translation
Chang, Ke-Ching
Chen, Chung-Chi
Yen, An-Zi
Computation and Language
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.
title Paraphrase-Aligned Machine Translation
topic Computation and Language
url https://arxiv.org/abs/2412.05916