Pattern Recognition Lab at FAU joins leading Bavarian partners in building sovereign AI foundation models for healthcare
Erlangen, April 23, 2026 — The Pattern Recognition Lab at Friedrich-Alexander-Universität Erlangen-Nürnberg is pleased to announce that the collaborative project AI-BAY-HEALTH – A Sovereign Multi-Modal Healthcare Foundation Model for Bavaria has officially been approved. The initiative brings together leading researchers across Bavaria to develop next-generation foundation models for biomedical and clinical data, with the goal of advancing trustworthy, high-impact artificial intelligence for healthcare and life sciences.
AI-BAY-HEALTH unites expertise from FAU, TUM, LMU, and university hospitals and research partners across the state. The project focuses on combining multiple modalities, including medical imaging, radiology, pathology, molecular data, omics, and clinical representations, within a shared AI framework. By connecting these complementary data sources through modern embedding and foundation model approaches, the consortium aims to create robust and interoperable models that can support a broad range of downstream research and clinical applications.
A central scientific objective of the project is to investigate whether jointly learned and integrated embeddings across modalities can improve performance on important biomedical evaluation tasks while enabling new emergent capabilities that do not arise in isolated models. In this way, AI-BAY-HEALTH will contribute to a new generation of sovereign and collaborative AI infrastructures for health research in Bavaria.
At FAU, the Pattern Recognition Lab contributes its expertise in machine learning, medical imaging, multimodal representation learning, and AI for translational healthcare. The project also establishes the basis for a shared technical infrastructure, including joint model development, collaborative code integration, and large-scale computational resources to support training and evaluation.
The approval marks the beginning of the initial implementation phase. Over the coming months, the consortium will align model development activities across sites, establish the shared computational framework, and prepare for the first major evaluation phase planned for 2027. The project is designed not only to deliver measurable scientific progress in its initial phase, but also to lay the groundwork for a sustainable long-term research platform for healthcare foundation models in Bavaria.
“AI-BAY-HEALTH is an important step toward building sovereign, multimodal AI systems for medicine in Bavaria,” said Prof. Dr. Andreas Maier, Head of the Pattern Recognition Lab at FAU. “The project creates a unique opportunity to bring together complementary expertise from across institutions and data modalities. We are excited to contribute our experience in medical AI and representation learning to this highly collaborative effort.”
The consortium is coordinated by leading researchers from Bavarian universities and research institutions and is supported within the framework of the Bavarian AI foundation model initiative. With its strong emphasis on collaboration, measurable milestones, and shared infrastructure, AI-BAY-HEALTH is positioned to become a major driver of innovation at the intersection of artificial intelligence and healthcare.
For the Pattern Recognition Lab, participation in AI-BAY-HEALTH is another important milestone in strengthening research on interpretable and clinically relevant AI methods, and in expanding collaborative structures for data-driven medicine across Bavaria.