ICONKAT: an integrated constructivist knowledge acquisition tool

Abstract

In this paper, we report on a continuing research effort aimed at the development of an integrated knowledge acquisition system, ICONKAT. We describe the components of the tool and discuss how they may be used to facilitate the design, construction, testing, maintenance and explanation of knowledge bases. ICONKAT's knowledge elicitation subsystem, based on both personal construct theory and assimilation theory, interactively assists the domain expert in the task of building a model of his or her expertise. ICONKAT employs a collection of modeling primitives (i.e. the glue) as the material basis for the construction of a conceptual domain model. The maintenance subsystem provides support tools for use by the knowledge engineering team, as well as the domain expert, when testing the system's performance, refining the knowledge base, and maintaining the overall system. The components of the maintenance subsystem employ a variety of mediating representations (e.g. concept maps, repertory grids) to furnish various perspectives of the evolving domain model as embodied in the modeling primitives. Moreover, the domain model that emerges from the knowledge acquisition process is subsequently exported from the development environment to the delivery environment where it serves as the foundation of the explanation capability for the deployed system. ICONKAT is currently employed in the design and construction of an expert system for the diagnosis of first pass functional cardiac images.