aisc-team-a1/Asclepius-Synthetic-Clinical-Notes
收藏资源简介:
Asclepius数据集是一个合成的临床笔记和指令数据集,用于构建临床大型语言模型。该数据集包含临床笔记、问题和答案的格式,数据来源于PMC-Patients病例报告,并使用GPT-3.5生成指令-答案对。数据集支持8种任务,包括命名实体识别、缩写扩展、关系提取、时间信息提取、共指消解、释义、摘要和问答。数据集的语言为英语,结构包括患者ID、笔记、问题、答案和任务类别。数据集的创建使用了GPT-3.5-turbo模型,并提供了多个模型变体。数据集的许可证为CC-BY-NC-SA 4.0,并提供了引用信息。
The Asclepius Dataset is a synthetic clinical note and instruction dataset developed for constructing clinical large language models (LLMs). The dataset follows the structure of clinical notes, questions, and answers, with its source data derived from PMC-Patients case reports, and its instruction-answer pairs generated via GPT-3.5. It supports eight distinct tasks, including named entity recognition, abbreviation expansion, relation extraction, temporal information extraction, coreference resolution, paraphrasing, summarization, and question answering. The dataset uses English as its language, and its schema includes patient ID, clinical note, question, answer, and task category. It was constructed using the GPT-3.5-turbo model, with multiple model variants provided. The dataset is licensed under CC-BY-NC-SA 4.0, and complete citation information is available.
Asclepius: Synthetic Clinical Notes & Instruction Dataset
数据集描述
数据集概述
- 名称: Asclepius: Synthetic Clinical Notes & Instruction Dataset
- 语言: 英语
- 标签: 医学, 合成
- 大小类别: 100K<n<1M
- 许可证: CC-BY-NC-SA 4.0
数据集组成
- 格式: 临床笔记 - 问题 - 答案
- 来源: 从 PMC-Patients 病例报告中合成的笔记,使用 GPT-3.5 生成
- 数量: 157k 合成出院总结的指令-答案对
支持的任务
- 命名实体识别
- 缩写扩展
- 关系抽取
- 时间信息抽取
- 共指消解
- 释义
- 摘要
- 问答
数据集结构
数据实例
- 文件:
synthetic.csv - 内容: 临床笔记 - 问题 - 答案对
数据字段
patient_id: PMC-Patients 中的唯一病例报告IDpatient: 病例报告文本question: GPT-3.5 从患者生成的指令answer: GPT-3.5 为给定病例报告和问题生成的答案task: 问题的对应类别
数据集创建
源数据
注释
- 使用 GPT-3.5-turbo (版本 0314)
附加信息
模型
变体
- 从 MIMIC-III 出院总结生成的指令-答案对及训练的模型可在 Physionet 获取
引用信息
@misc{kweon2023publicly, title={Publicly Shareable Clinical Large Language Model Built on Synthetic Clinical Notes}, author={Sunjun Kweon and Junu Kim and Jiyoun Kim and Sujeong Im and Eunbyeol Cho and Seongsu Bae and Jungwoo Oh and Gyubok Lee and Jong Hak Moon and Seng Chan You and Seungjin Baek and Chang Hoon Han and Yoon Bin Jung and Yohan Jo and Edward Choi}, year={2023}, eprint={2309.00237}, archivePrefix={arXiv}, primaryClass={cs.CL} }




