Reasoning visualization in expert systems—the applicability of algorithm animation techniques

  • Authors:
  • William J. Selig;James D. Johannes

  • Affiliations:
  • NASA/Marshall Space Flight Center;University of Alabama in Huntsville

  • Venue:
  • IEA/AIE '90 Proceedings of the 3rd international conference on Industrial and engineering applications of artificial intelligence and expert systems - Volume 1
  • Year:
  • 1990

Quantified Score

Hi-index 0.00

Visualization

Abstract

Visualization of the reasoning process of an expert system can be useful in many situations. Unfortunately, no current expert system development environment presents us with this visualization. The problem is three-fold. First, the information presented typically contains too much detail. It is presented at the wrong conceptual level. Second, the information desired is not always available. Third, even when the desired information is presented at a reasonable conceptual level, there may be too much of it.This paper presents the results of research into providing a means for users to flexibly and easily create visualizations of the reasoning processes of forward-chaining rule-based expert systems using algorithm animation techniques. Levels of reasoning are described in order to identify the information necessary from the expert system development environment for these visualizations.A dual process visualization environment is presented consisting of 1) a version of CLIPS modified for the identified information access requirements, and 2) VISOR, an algorithm animation based system for creating visualizations of arbitrary complexity which can be triggered by “interesting event” messages from the running expert system application. This is followed by examples from several visualizations performed during the scope of this work.