Luc Steels

Components of expertise

This article discusses frameworks for studying expertise at the knowledge level and knowledge-use level. It reviews existing approaches such as inference structures, the distinction between deep and surface knowledge, problem-solving methods, and generic tasks. A new synthesis is put forward in the form of a componential framework that stresses modularity and an analysis of the pragmatic constraints on the task. The analysis of a rule from an existing expert system (the Dipmeter Advisor) is used to illustrate the framework.

The componential framework and its role in reusability

The paper describes a framework for constructing knowledge level models. The framework assumes that descriptions can be constructed from three perspectives: models, methods, and tasks. A workbench is introduced that supports the interactive specification of knowledge level models. The role of the symbol level is described. The problem of reuse is briefly discussed within this context.

Second generation expert systems

Second generation expert systems are able to combine heuristic reasoning based on rules, with deep reasoning based on a model of the problem domain. This solves a number of major problems of current expert systems, most importantly the problem of knowledge acquisition: second generation expert systems can learn new rules by examining the results of deep reasoning. The paper outlines the components of second generation expert systems and gives an example.