基于开放医疗大数据的人工智能研究

医疗人工智能（artificial intelligence, AI）的发展离不开大数据的汇聚与标注，同时也引发了医疗大数据及其标注标准如何统一的问题。从AI发展对医疗大数据需求出发，指出建立医疗大数据统一标准的重要性与必要性，分析当前医疗AI技术的开发过程缺乏数据统一标准和临床共识，提出“以开放换取共识”的医疗大数据平台方案——素问系统，实现“算法工程师开发技术—临床医生编辑知识—伦理专家监控伦理”的一体化AI生产流程。最后，以素问系统中第一个开源项目——医学知识图谱，作为大数据抽取标准知识的案例，解析了如何以开放换取共识，让AI技术专家与临床专家共同打造可落地的医疗AI技术。

Artificial Intelligence Development Based on Medical Big Data Opening

The development of medical artificial intelligence (AI) is inseparable from the collection and labeling of big data, but it is also accompanied by the compatibility of medical big data labeling standards. The significance and necessity of medial AI big data unification is claimed in the paper.Starting from the problems in the development of medical big data, this paper analyzes the limitations of the lack of clinical consensus in medical data standard, and puts forward a new medical big data platform scheme called "Suwen System" aiming at "openness for consensus". The new medical big data platform allows experts from different fields to work together online, generate consensus through continuous iteration of knowledge, and realize the integrated AI production process of "algorithm engineer manufacturing-clinician editing-ethical expert monitoring".Last but not least, it takes the first opensource project, medical knowledge graph, as the example to expound how to achieve consensus by opening, so as to generate feasible medical AI technology by AI experts and clinical experts.