emrQA
收藏资源简介:
emrQA是由MIT-IBM Watson AI Lab创建的一个大规模问答数据集,专门用于电子医疗记录(EMR)的问答任务。该数据集包含超过400,000个问题-答案对和100万个问题-逻辑形式对,数据来源于临床笔记和其他医疗文档。创建过程中,研究团队利用了现有的i2b2数据集中的专家注释,通过一种新颖的生成框架来构建数据集。emrQA的应用领域主要集中在医疗领域,旨在帮助医生从EMR中快速准确地获取信息,以支持临床决策和提高医疗服务质量。
emrQA is a large-scale question answering (QA) dataset created by the MIT-IBM Watson AI Lab, specifically designed for question answering tasks on electronic medical records (EMR). This dataset contains over 400,000 question-answer pairs and 1 million question-logical form pairs, with data sourced from clinical notes and other medical documents. During its development, the research team utilized expert annotations from the existing i2b2 dataset and constructed the dataset via a novel generative framework. The application scenarios of emrQA are mainly focused on the healthcare domain, aiming to assist clinicians in quickly and accurately retrieving information from EMRs to support clinical decision-making and improve the quality of medical services.

- 1emrQA: A Large Corpus for Question Answering on Electronic Medical RecordsMIT-IBM Watson AI Lab · 2018年



