Research Group:

Statistical and Symbolic Artificial Intelligence

Head of the Research Group:

Felix Weitkämper

The group for statistical and symbolic artificial intelligence investigates human-centred artificial intelligence: How can human knowledge be encoded in artificial intelligence, and how does human knowledge empower artificially intelligent systems? 
In this endeavour, we develop, analyse and deploy formalisms from two distinct subfields of AI: symbolic AI, which encodes deterministic knowledge in logical rules and explicit definitions, and probabilistic AI, which encodes uncertain knowledge and random processes in graphical models (such as Bayesian networks) and systems of probabilistic rules.
In particular, we are interested in the meeting points of these two fields. This includes statistical relational AI, which combines the expressive power of classical logic with the flexibility of probabilistic graphical models, and the study of causality, which makes use of human knowledge to go beyond the description of a given situation and make predictions about the impact of external interventions and counterfactuals.
Throughout, we are particularly interested in the role of machine learning in extracting explicit knowledge from intricate, interconnected datasets.