In life and social sciences, the same experimental evidence may support conflicting interpretations. Logical methods help articulate when statistical or machine-learning outputs should count as justified scientific inferences.
Core questions
- How should we reason from uncertain data?
- How should we compare contradictory outcomes?
- What counts as valid support or rejection of a hypothesis?
Methods
LUCI combines argumentation theory, non-monotonic logic, and many-valued logic to model data-driven reasoning and explainable AI.
MEPER / ReDa line
The ReDa project coordinates work on methodological foundations for constructing probabilistic evidence under uncertainty, including collaboration with clinical and biomedical partners.