TransLaw：模拟香港判例法专业翻译的大规模数据集与多智能体基准
TransLaw: A Large-Scale Dataset and Multi-Agent Benchmark Simulating Professional Translation of Hong Kong Case Law
AI总结：
针对香港判例法英译中资源匮乏、法律术语和格式要求严格的问题，构建了首个大规模句对齐平行语料库HKCFA Judgment 97-22，并提出多智能体框架TransLaw，通过分解翻译任务、集成法律词汇库和检索增强生成，显著提升翻译质量，但仍未达到人类专家的风格自然度。
摘要：
根据《基本法》第8-9条，香港法院判决书需从英文翻译成繁体中文，但由于平行资源短缺以及对法律术语、引用格式和司法风格的严格要求，这一任务仍受到限制。我们引入了HKCFA Judgment 97-22，这是首个用于香港判例法的大规模句对齐平行语料库，包含344份专业翻译的判决书（11,099个句对；210万词元），涵盖1997年至2022年。基于这一资源，我们提出了TransLaw，一个多智能体框架，将翻译分解为词级表达、句级翻译和多维审查，集成了专门的香港法律词汇数据库、检索增强生成和迭代反馈，并包括涵盖语义对齐、术语、引用和风格的四维专家审查。通过对13个开源和商业大语言模型进行基准测试，我们证明TransLaw在所有评估模型上均显著优于单智能体基线，并在3次迭代内收敛。由10名持证法律翻译人员使用我们提出的Legal ACS指标进行的人工评估证实了法律语义准确性的提升，同时表明TransLaw在风格自然度上仍落后于人类专家。数据集和基准代码可在以下网址获取：https://xxx。
英文摘要：
Translating Hong Kong Court Judgments from English to Traditional Chinese is mandated by Articles 8-9 of the Basic Law, yet remains constrained by a shortage of parallel resources and rigorous demands on legal terminology, citation format, and judicial style. We introduce HKCFA Judgment 97-22, the first large-scale sentence-aligned parallel corpus for HK case law, comprising 344 professionally translated judgments (11,099 sentence pairs; 2.1M tokens) spanning 1997-2022. Building on this resource, we propose TransLaw, a multi-agent framework that decomposes translation into word-level expression, sentence-level translation, and multidimensional review, integrating a specialized Hong Kong legal glossary database, Retrieval-Augmented Generation, and iterative feedback, with four-dimensional expert review covering semantic alignment, terminology, citation, and style. Benchmarking 13 open-source and commercial LLMs, we demonstrate that TransLaw significantly outperforms single-agent baselines across all evaluated models, with convergence within 3 iterations. Human evaluation by 10 certified legal translators using our proposed Legal ACS metric confirms gains in legal-semantic accuracy, while showing that TransLaw still trails human experts in stylistic naturalness. The dataset and benchmark code are available at https://github.com/xuanxixi/TransLaw.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
