Guest lecturer Prof. Dr. Anne-Christin Hauschild (Institute for Predictive Deep Learning for Medicine and Healthcare, Justus Liebig University Giessen, Hessian.AI) presented a talk on 12 February 2026 as part of our Seminar on Recent Methods at the Institute of Medical Informatics in Münster. Her lecture was titled “Navigating the Challenges of Predictive Modeling of Multi-Modal Health Data for Clinical Decision-Making.”

In her presentation, Hauschild illustrated how methods from artificial intelligence are increasingly being used to analyze large volumes of medical data. At the same time, she addressed key challenges such as limited sample sizes, potential biases in datasets, and the limited interpretability of many predictive models.

Using examples from current research projects, she demonstrated how these challenges can be addressed. The project FAIrPaCT, for instance, employs federated AI methods to analyze different types of medical data—such as clinical, molecular, and imaging data from patients with pancreatic cancer—across institutions without requiring the exchange of sensitive patient data. She also introduced new approaches to improving the explainability of AI systems, including BenchXAI and the interactive platform CLARUS.

The lecture highlighted how robust and transparent AI systems can help make better use of complex health data for clinical decision-making and ultimately contribute to advances in personalized medicine.