The Approximate and Qualitative Reasoning (AQR)
Group is part of the
Institute for
Representation and Reasoning within
the Division
of Informatics at the University
of Edinburgh. The Division has earned top ratings for both research
(5A) and teaching (Excellent) in the most recent UK national quality assessment,
and is the premier research centre for AI in Europe.
The members of the
AQR group are particularly active in the areas of fuzzy and qualitative
modelling, model-based reasoning and knowledge reuse. The following are
the strands of our current research:
- Integration of fuzzy, qualitative and graphical models to increase the
effectiveness and efficiency of practical reasoning systems.
-
Exploitation of computational learning mechanisms for domain modelling, with a
focus on knowledge formulation and refinement.
-
Investigation of causal and compositional modelling methods for explanation
generation, within both physical and non-physical domains.
- Use of multiple models within a model family to allow fast instantiation
and adaptation of representations of knowledge.
-
Development of novel constraint satisfaction techniques for problems which
involve dynamic, uncertain and/or flexible constraints.
On the whole, these strands of research concern the development of techniques
for representing the common-sense knowledge of the average person and the
tacit knowledge used by engineers and scientists. This allows the subjective
element of common-sense knowledge to be incorporated within formal and
domain-independent algorithms for reasoning about domain-specific problems.
Most of our research projects are/were funded by EPSRC and industrial
partners.
[Introduction] [People]
[Projects] [Publications]
[Demos] [Links]
[Contact]