Clinical data and the federated learning concept for an AI/ML device
£40.00
Understand how federated learning could be used to train AI/ML medical devices while protecting patients’ health data
English
With many organisations wanting to develop artificial intelligence (AI)/machine learning (ML) powered applications, it is essential that high quality data are available in large numbers. AI powered inferencing in healthcare applications is gaining popularity, especially in diagnostics and therapy, but this poses several critical concerns over the data flow. For many patients, privacy of their healthcare records is a concern, and healthcare providers too want to ensure that there is ethical and transparent methodology concerning data sharing for research purposes. This article by Abhineet Johri proposes the use of federated learning in a general data protection compliant manner to ensure that healthcare providers remain the data controllers. However, the ability to train on unseen data brings several challenges like ensuring quality of data and identification of outliers, and the regulatory concerns relating to the quality of algorithm results (inferences). This article summarises the various challenges associated with using federated learning and offers solutions to these problems.
Specification: 8 pages plus covers, in English, supplied by email as a PDF.
This article has been published in the February 2025 issue of the Journal of Medical Device Regulation.
