Eliciting and Using Experiential Knowledge and General Expertise
David S. Prerau, Mark R. Adler, and Alan S. Gunderson
Expert systems generally contain the knowledge of human domain experts in the form of heuristic rules. The heuristics encapsulate the rules of thumb that the experts have gained through extensive experience in the field. These experience-proven heuristics are used by the experts as a major part of their problemsolving approach. When confronted with rare and novel problems to which the heuristics may not apply, human experts can apply, in addition, their fundamental understanding -- or "deep knowledge" -- of the domain. In GTE's COMPASS and PROPHET expert systems, we have elicited and used expert knowledge that is at an intermediate level, lying between knowledge based on direct experience and knowledge based on fundamental domain principles. We call this intermediate level expertise knowledge. Use of this type of knowledge allowed us to extend the range of both the resulting systems beyond the heuristics gained from the experts' field experience.
This chapter discusses the elicitation and use of expertise knowledge in expert systems. We utilized these techniques in the development of both the COMPASS and PROPHET expert systems, which contain knowledge based on both the experience and the expertise of our domain experts. In this chapter, we use the COMPASS expert system as our primary example. In the first section of the chapter, we define and describe three levels of knowledge: experiential knowledge, deep knowledge, and expertise knowledge. The second section describes the COMPASS expert system and discusses its task, the analysis of GTE's No. 2 EAX switching system. The following section describes how we elicited and used the expert knowledge. Next, we discuss the differences between performing knowledge elicitation for knowledge based on experience and for knowledge based on the general expertise of our domain expert. The last major section of the chapter discusses the risks involved in basing an expert system's knowledge on the general expertise of an expert.
Most of the expert systems being produced today are based on the day-to-day experiences of domain experts. Experts have a thorough understanding of the fundamental principles that are involved, based on lengthy experience and often on extensive training. Through their familiarity with the domain, their training, and their experiences in solving problems within that domain, the experts develop heuristics, or rules of thumb. We call this knowledge based on extensive field applications experiential knowledge.
Experiential knowledge is associational, providing a direct relationship between situational data and conclusions, without the explicit use of the principles of the domain. The knowledge is usually related to a narrowly defined problem. For example, an expert automobile mechanic may hear a certain noise