Schultze Joachim L
Systems Medicine, Deutsches Zentrum für Neurodegenerative Erkrankungen (DZNE), Bonn, Germany.
PRECISE Platform for Single Cell Genomics and Epigenomics, Deutsches Zentrum für Neurodegenerative Erkrankungen (DZNE) and University of Bonn, Bonn, Germany.
J CME. 2023 Jan 10;12(1):2162202. doi: 10.1080/28338073.2022.2162202. eCollection 2023.
An avalanche of medical data is starting to be build up. With the digitalisation of medicine and novel approaches such as the omics technologies, we are conquering ever bigger data spaces to be used to describe pathophysiology of diseases, define biomarkers for diagnostic purposes or identify novel drug targets. Utilising this growing lake of medical data will only be possible, if we make use of machine learning, in particular artificial intelligence (AI)-based algorithms. While the technological developments and chances of the data and information sciences are enormous, the use of AI in medicine also bears challenges and many of the current information technologies (IT) do not follow established medical traditions of mentoring, learning together, sharing insights, while preserving patient's data privacy by patient physician privilege. Other challenges to the medical sector are demands from the scientific community such as "Open Science", "Open Data", "Open Access" principles. A major question to be solved is how to guide technological developments in the IT sector to serve well-established medical traditions and processes, yet allow medicine to benefit from the many advantages of state-of-the-art IT. Here, I provide the Swarm Learning (SL) principle as a conceptual framework designed to foster medical standards, processes and traditions. A major difference to current IT solutions is the inherent property of SL to appreciate and acknowledge existing regulations in medicine that have been proven beneficial for patients and medical personal alike for centuries.
大量的医学数据开始积累起来。随着医学数字化以及诸如组学技术等新方法的出现，我们正在攻克越来越大的数据空间，用于描述疾病的病理生理学、定义用于诊断目的的生物标志物或识别新的药物靶点。只有当我们利用机器学习，特别是基于人工智能（AI）的算法时，才有可能利用这一不断增长的医学数据池。虽然数据和信息科学的技术发展和机遇巨大，但人工智能在医学中的应用也面临挑战，而且当前许多信息技术（IT）并不遵循已确立的医学传统，即指导、共同学习、分享见解，同时通过医患特权保护患者的数据隐私。医学领域面临的其他挑战来自科学界的要求，如“开放科学”“开放数据”“开放获取”原则。一个有待解决的主要问题是如何引导信息技术领域的技术发展，以服务于已确立的医学传统和流程，同时让医学从最先进的信息技术的诸多优势中受益。在此，我提出群体学习（SL）原则，作为一个旨在促进医学标准、流程和传统的概念框架。与当前的信息技术解决方案的一个主要区别在于，群体学习具有一种内在特性，即认可和承认医学中已存在数百年且已被证明对患者和医护人员都有益的现有规定。
