Distributed Analytics

Our Distributed Analytics (DA) team focuses on enabling privacy-preserving machine learning on sensitive medical patient data. We develop technologies that allow institutions to collaboratively analyze patient data - such as electronic health records or imaging data - without the need to centralize it. This is particularly crucial in healthcare settings, where data privacy and sovereignty are paramount. Our team has extensive experience in developing PADME (Platform for Analytics and Distributed Machine Learning for Enterprises), a modular and open platform for federated and incremental learning across distributed clinical data sources. PADME is being continuously refined and applied in diverse research and clinical use cases, including national initiatives like the German Medical Informatics Initiative (MII) through the PrivateAIM project and collaborative European projects such as the Horizon Europe project BETTER. The DA team actively contributes to the design, development, and evaluation of novel data infrastructure components that enable secondary use of health data on a national and international scale. Our work supports the integration and reusability of multimodal clinical data - ranging from structured EHRs to imaging and genomics -across institutional boundaries. Through our work, we aim to bridge the gap between cutting-edge machine learning research and real-world clinical data environments - empowering institutions to extract value from data while respecting privacy, legal, and ethical constraints.