Enhancing explanations in knowledge-based systems with hypertext

Abstract

This article investigates the use of hypertext to enhance explanations in knowledge-based systems (KBS). Three fundamental issues are addressed: (1) why is hypertext suited for enhancing explanations in KBS? (2) what kind of knowledge needs to be added to KBS to provide effective explanations? and (3) how should such knowledge be represented efficiently and made accessible naturally and conveniently to users? The central idea is to develop a hypertext deep knowledge base to supplement KBS conclusions and explanations, so that users are provided with a rich context for understanding and interpreting KBS conclusions and reasoning. Explanations in KBS can be made more intelligible and relevant to users, and more interactive by (1) representing deep knowledge needed for explanations with hypertext in a way consistent with learning and educational theories, and (2) providing access to domain knowledge from KBS output with hypertext links, thus integrating domain knowledge into the KBS output in a manner conducive to contextualized learning.