Adaptive expert systems and analogical problem solving
Conventional expert systems are "brittle" in the sense that they require substantial human intervention to compensate for even slight variations in descriptions, and break easily when they reach the edge of their knowledge. In response to this problem, this paper describes a prototype of a new generation of expert systems, called an adaptive expert system (AES), which is capable of adapting its knowledge dynamically and analogically. AES combines the focussed power of expert systems with the analogical problem solving abilities of case-based reasoning systems, and demonstrates much higher "IQs" than the expert systems currently available on the market.
Key WordsAnalogical problem solving Case-based reasoning Expert systems
Unable to display preview. Download preview PDF.
- J.G. Carbonell, "Learning by Analogy: Formulating and Generalizing Plans from Past Experience", Machine Learning (I): pp 137–162, 1983.Google Scholar
- F. Hayes-Roth, Waterman, D.A. & Lenat, D.B., "Principles of Pattern-Directed Inference Systems", Pattern-Directed Inference Systems, Academic Press, 1978.Google Scholar
- J. H. Holland, "Escaping Brittleness: The possibilities of general purpose learning algorithms applied to parallel rule-based systems", Machine Learning (II): pp. 593–623, Morgan Kaufmann, 1986.Google Scholar
- J. Kolodner, "Retrieval and Organizational Strategies in Conceptual Memory", Ph.D. Thesis, Yale University, 1980.Google Scholar
- R. L. Simpson, Jr. "A computer model of case-based reasoning in problem solving: an investigation in the domain of dispute mediation", Ph.D. thesis, Georgia Tech., 1985.Google Scholar
- H. H. Zhou, "A Computational Model of Cumulative Learning", to appear in Machine Learning.Google Scholar
- H. H. Zhou, "CSM: A Genetic Classifier System with Memory for Learning by Analogy", Ph.D. thesis, Computer Science Department, Vanderbilt University, Dec. 1987.Google Scholar