Skip to main content

Tools and a methodology for acquiring high level knowledge from domain experts in the context of medical problem solving


Research objectives and content
This project aims at investigating the potential for the dynamic acquisition of background knowledge in the context of semi automated systems which build or refine knowledge-bases. The principal idea to be explored is that the necessary expert knowledge is best acquired dynamically, at the time it is useful. This need appears during the system's operation, when a disagreement appears between the system's hypotheses and the user. It is through the justification by the expert of its previous judgments that we can incrementally narrow the gap between the system's knowledge and the expert's knowledge. This approach can be adopted to enhance an interactive machine learning algorithm designed as an aid to scientific discovery like PASTEUR (developed by the applicant), or to enhance a system for the refinement of existing knowledge bases like REFINER (developed at the host laboratory). The resulting systems will be evaluated on medical problems in collaboration with Aberdeen's Medical School.
Training content (objective, benefit and expected impact)
Benefit from internationally recognized excellence of Aberdeen Computer Science laboratory on both Knowledge Base Systems and computational research on scientific discovery. Apply and develop further the method developed during doctoral studies in the framework of the host's collaboration with the medical school.
Links with industry / industrial relevance (22)
The project will hopefully lead to the production of an operational system (such as a knowledge-base patient monitoring system) that could be used within the hospital in Aberdeen or elsewhere.

Funding Scheme

RGI - Research grants (individual fellowships)


University of Aberdeen
King's College
AB9 2UB Aberdeen
United Kingdom

Participants (1)

Not available