PROBABILISTIC AI

Research Group

The Probabilistic AI Research Group (PAIR)

Mission

Our mission is to develop AI systems that support reliable decision-making under uncertainty and can be responsibly deployed in real-world applications.

Core Research Areas

  • Probabilistic Deep Learning, Large Language Models and Agentic AI
  • Predictive Modeling for tabular, imaging, text, and multi-modal data
  • Statistical Modeling, Bayesian Inference and Uncertainty Quantification
  • Causal Modeling

The Probabilistic AI Research Group conducts applied research at the intersection of statistics, explainable artificial intelligence, and causality. While modern AI systems achieve impressive predictive performance, their practical deployment is often limited by insufficient uncertainty quantification, lack of robustness and interpretability, and missing causal understanding.

We address these limitations by combining deep learning and large language models with interpretable statistical models, valid uncertainty estimation, and causal discovery and inference methods. Our work is both methodologically grounded and application-driven, aligning academic excellence with real-world impact.
PAIR is committed to collaborative, third-party funded research in close interaction with academic and industrial partners.

Key Domains Include

  • Medical and healthcare AI
  • Energy demand management and renewable energy systems
  • Predictive modeling and decision support under uncertainty in complex systems

Selected Prior Work and Example Projects

Leadership

Prof. Dr. Beate Sick

Group Leader

Prof. Dr. Oliver Dürr

Co-Director

Members

Publications

Loading publications...