术语
计算机科学
一致性(知识库)
翻译(生物学)
自然语言处理
机器翻译
人工智能
过程(计算)
质量(理念)
计算机辅助翻译
情报检索
数据科学
干预(咨询)
语言翻译
作者
Yan Huang,Wei Liu,Jiayi Wang,Huidong Zhu
标识
DOI:10.1038/s41598-026-77426-4
摘要
This paper proposes an Automated Terminology Enhancement Translation (ATET) Framework that integrates Large Language Models (LLMs) with a terminology enhancement mechanism to improve the accuracy and consistency of domain-specific translations. LLMs often struggle with errors in translating specialized terminology. The ATET Framework addresses this issue by introducing an automated process for identifying and correcting terms during translation, thereby reducing reliance on manual intervention and minimizing translation errors. Experimental results demonstrate that the ATET Framework significantly improves translation quality across multiple language pairs (zh-en, de-en, en-cs, en-zh, en-de, cs-en), particularly in handling domain-specific terms. Fine-tuning further enhances the accuracy of terminology translation, validating the Framework’s applicability and potential in multilingual translation tasks. This research offers an effective solution for advancing machine translation technology and highlights its practical value in specialized translation scenarios.
科研通智能强力驱动
Strongly Powered by AbleSci AI