ESKAPE/CF: a knowledge acquisition tool for expert systems using cognitive feedback

Abstract

The major bottleneck in the construction of expert systems is the time-consuming process of acquiring knowledge from experts. Automated knowledge acquisition tools have demonstrated the ability to reduce the time required to construct expert system knowledge bases and are supported by both knowledge engineers and experts. However, due to limitations in their underlying psychological paradigms, existing tools may not be well-suited to extracting semantic or procedural knowledge from an expert. This thesis designs and implements an Expert System Knowledge Acquisition and Policy Evaluation tool using Cognitive Feedback (ESKAPE/CF) , based on Lens model techniques which have demonstrated effectiveness in capturing policy knowledge. The system is designed to be used interactively by an expert to reduce the historically lengthy interactions with a knowledge engineer. Additionally, the use of cognitive feedback techniques should enable the system to capture expertise that has heretofore been unobtainable by existing knowledge acquisition tools.