Researchers and developers from the Horizon Europe projects dAIbetes and Microb-AI-ome came together for a collaborative hackathon dedicated to federated learning and privacy-preserving artificial intelligence in healthcare. Hosted at Sapienza Università di Roma on the 5th and 6th of May, 2026, the event provided participants with a unique opportunity to gain practical experience with modern federated AI technologies while strengthening collaboration between the two sister projects.
The hackathon was built around a real-world healthcare scenario using the UCI Diabetes 130-US Hospitals dataset, comprising approximately 100,000 patient encounters from 130 hospitals. Rather than sharing sensitive patient data, participants learned how federated learning enables AI models to be trained collaboratively while keeping data securely stored at each participating institution. Throughout the event, attendees explored the complete federated learning workflow—from setting up the computing environment and developing AI applications to training distributed models and evaluating their performance on a federated platform.
Working in multidisciplinary teams, researchers, clinicians, software developers and data scientists exchanged expertise on federated infrastructures, privacy-preserving machine learning and the practical challenges of deploying AI solutions in healthcare. The event also served as a valuable networking opportunity, fostering collaboration between the dAIbetes and Microb-AI-ome consortia and encouraging the exchange of ideas across complementary research domains.
By combining practical training with cross-project collaboration, the hackathon reinforced the shared commitment of both Horizon Europe projects to advancing trustworthy, secure and clinically relevant artificial intelligence. Such initiatives contribute to building a stronger European ecosystem for federated AI, helping researchers translate innovative methods into real-world healthcare applications while ensuring the highest standards of data privacy and security.
More information can be found here.
