Knowledge Engineering Issues In Biomedicine

This paper examines knowledge engineering issues relevant to the design of advanced Computer Assisted Problem Solving environment. It is stated that problem solving not only reduces to apply a set of operators but also imply careful formulation and modelling steps aimed at finding an adequate statement of the problem. Learning efforts are developped in parallel and determine the definite capacity of systems to evolve and autonomously adapt to new situations. Three main cognitive activities are considered (stating, solving and learning) regarding both the associated human skills and their possible modelling by machines. These notions are illustrated by their application to cytopathology, a field combining diagnosis and visual skills.