Research

Our group focuses on medical AI and medical foundation models. It is our vision to develop and evaluate AI systems that are flexibly multimodal and that can be instructed to follow a large variety of tasks across different medical contexts and data modalities. For this, we consider AI agent systems that retrieve medical information, leverage tools, learn from context, and reason step-by-step across medical modalities. We focus on the following topics:

AI agents and agent systems
Multimodal large language models
Medical information retrieval

Medical reasoning
Personalized Medicine

Overview figure of an AI agent that interacts with a user. The AI agent accesses medical knowledge sources as well as patient data in a multimodal interaction with clinical users.