li-lab/GlobMedX

这是我们正在进行项目推进多语言医疗AI：GlobMed基准测试和低资源语言医疗问答的代理管道的样本数据。数据集包含多语言医疗问答内容，结构包括语言、原始英文数据、机器翻译数据、翻译准确性评估（由人类医学专家提供）、语言流畅性和语法质量评估、内容完整性评估、医学专家修订数据以及专家定性反馈或附加注释。数据集规模小于1,000条，涉及多种语言，专注于医疗领域。项目得到了Google Research Scholar Program的支持。

This is the sample data for our ongoing project Advancing Multilingual Medical AI: GlobMed Benchmarking and Agentic Pipelines for Medical QA in Low-Resource Languages. The dataset contains multilingual medical question-answering content, structured with columns for language, original English data, machine-translated data, evaluation of translation accuracy (provided by human medical experts), assessment of linguistic naturalness and grammatical quality, measure of content completeness, expert-revised data, and qualitative feedback or additional notes from medical experts. The dataset size is less than 1,000 entries, covers multiple languages, and focuses on the medical domain. The project is supported by the Google Research Scholar Program.