The Advanced Reasoning Group (ARG) aims to conduct innovative research in qualitative and approximative reasoning, including methods of knowledge representation, model generation and refinement, and model-based problem solving.
The ARG has an excellent track record of developing the scientific foundations necessary for building intelligent decision support systems, especially in crime detection and prevention, engineering design analysis, and computer-based diagnosis. In particular, the group is well known for its ground-breaking work on automated diagnostic and failure analysis for circuit design in the automotive industry, and its invention of fuzzy-rough semantics-preserving techniques for explicit knowledge model formulation and simplification. The group's research also includes other advanced computational intelligence techniques, e.g. evolutionary algorithms and meta-heuristics.
- Multiple failure FMEA and sneak circuit analysis
- Model-based whole lifecycle automated system analysis
- Qualitative model-based learning
- Knowledge extraction over high dimensional data sets
- Compositional modeling and preference handling.
Applications of the above techniques are wide-reaching, ranging from laboratory demonstrations (e.g. metabolic pathway identification and simulation, and crime scenario construction and investigation) to commercial productions (e.g. automotive and aeronautical fault diagnosis, and consumer sensitive data analysis).