Study Published in Nature Medicine Shows Positive Impact of Leveraging Federated Learning for Healthcare AI
September 15, 2021
A study published in Nature Medicine showcases the use of federated learning to develop an AI model predicting oxygen requirements for COVID-19 patients. Involving 20 institutions across North and South America, Europe, and Asia, the research utilized federated learning to train the model without sharing patient data, thereby maintaining privacy. The model achieved an average area under the curve (AUC) of over 0.92, indicating high accuracy. This approach demonstrates the potential of federated learning in creating robust, generalizable AI models in healthcare.
For more details, read the full article on Newswire.