Expert systems: perils and promise

ABSTRACT

Based on a review of some actual expert-system projects, guidelines are proposed for choosing appropriate applications and managing the development process.

REFERENCES

• 1: Maryam Alavi, An assessment of the prototyping approach to information systems development, Communications of the ACM, v.27 n.6, p.556-563, June 1984 [doi>10.1145/358080.358095]
• 2: Bachant, J.. and McDermott. J. Rl revisited: Four years in the trenches. AZ Msg. 5, 3 (Fall 1984), 21-32. A cogent summary of the building of one of the best-known expert systems.
• 3: Daniel G. Bobrow, If PROLOG is the answer, what is the question? or what it takes to support AI programming paradigms, IEEE Transactions on Software Engineering, v.11 n.11, p.1401-1408, Nov. 1985 [doi>10.1109/TSE.1985.231888]
• 4: Bobrow, DC., and Stefik, M. The Loops Manual. AI Systems, Xerox Corporation, Palo Alto. Calif., 1984. A description of a multiparadigm system used in expert-system development.
• 5: Bobrow, D.G., and Stefik, M. Perspectives on artificial intelligence programming. Science V231,4741 (Feb. 28, 1986), 951. An account of the programming styles and environments that are used to develop systems in the AI community.
• 6: Brown, J.S. The low road, the middle road, and the high road. In The AZ Business, P.H. Winston and K. Prendergast, Eds. MIT Press, Cambridge, Mass., 1984. Alternative approaches to the development of intelligent systems: their advantages and pitfalls.
• 7: Buchanan. B., and Shortliffe, E.H. Rule Based Expert Programs: The MYCIN Experiments of the Stanford Heuristic Programming Project. Addison-Wesley, Reading, Mass., 1984. An account of 10 years' experience developing rule-based expert systems, using the MYCIN medical diagnosis system as the primary example.
• 8: Buchanan, B., Barstow, D.. Bechtel, R., Bennett, J., Clancey, W., Kulikowski, C.. Mitchell, T., and Waterman, D.A. Constructing an expert system. In Building Expert Systems, F. Hayes-Roth, D.A. Waterman, and D.B. Lenat, Eds. Addison-Wesley, Reading, Mass., 1983. An account of the developmental stages of an expert system.
• 9: Chandrasekaran. B.. and Mittal, S Conceptual representation of medical knowledge for diagnosis by computer: MDX and related systems. In Advances in Computers, Vol. 22, M.C. Yovits. Ed. Academic Press, New York, 1983, pp. 218-295. An approach to partitioning medical knowledge, and how this approach was used in building a diagnostic system.
• 10: Davis, R. Knowledge based systems. Science V231, 4741 (Feb. 28, 1986). 957. A general review of the concepts and current status of building knowledge-based systems.
• 11: de Kleer, J. How circuits work. In Qualitative Reasoning about Physical Systems, D.G. Bobrow. Ed. MIT Press, Cambridge, Mass., 1985, pp. 205-280. A description of a qualitative reasoner that uses constraints to build models of how circuits do and do not work.
• 12: Peter J. Denning, The science of computing—expert systems, American Scientist, 74, p.18-20, Jan.-Feb. 1986
• 13: Fagin, R.. and Halpern, J.Y. Belief, awareness, and limited reasoning: Preliminary report. In Proceedings of the 9th International Joint Conference on Artificial Intelligence. 1985, pp. 491-501. The theoretical foundations underlying the formalization of commonsense understanding of differences such as belief and knowledge.
• 14: Forgy, CL. The OPS5 user's manual. Tech. Rep., Dept. of Computer Science, Carnegie-Mellon Univ., Pittsburgh, Pa., 1980. The manual describing one of the primary languages used in expert-system development (and used in the Rl system).
• 15: Peter Friedland, Special section on architectures for knowledge-based systems, Communications of the ACM, v.28 n.9, p.902-903, Sept. 1985 [doi>10.1145/4284.214937]
• 16: Adele J. Goldberg, About This Issue…, ACM Computing Surveys (CSUR), v.14 n.3, p.319-320, Sept. 1982 [doi>10.1145/356887.356888]
• 17: Kunz. J., Kehler. T., and Williams. M. Applications development using a hybrid AI development system. AZ Msg. 5, 3 (Fall 1984), 41-54. Describes a commercially available multiparadigm tool useful for expert-system development.
• 18: Lane. CD., Walton, J.D., and Shortliffe, E.H. Graphical access to a medical expert system: 11. Design of an interface for physicians (Memo KSL-85-15). Methods Znf. Med. 25 (1986). Describes the development of the user interface for the Oncocin system.
• 19: Lindsay, R.K.. Buchanan, B., Feigenbaum, E., and Lederberg, J. Applications of Artificial Zntelligence for Organic Chemistry. Kaufman, Los Altos, Calif., 1985. The history of the Dendral project, one of the earliest AI expert systems.
• 20: McDermott, J. RI: The formative years. AZ Mng. V2, 2 (Spring 1981), 21-29. An easy-to-read description of the trials and tribulations of moving an expert system from a university to business environment.
• 21: McDermott, J. Rl: A rule-based configurer of computer systems. Artif. Zntelt. 19, 1 (Jan. 1982), 39-88. Technical description of the first commercially successful expert system for building configuration descriptions of computer systems.
• 22: Marvin Lee Minsky, Computation: Finite and Infinite Machines, Prentice-Hall, Inc., Upper Saddle River, NJ, 1967
• 23: Mittal, S., and Dym, C. Knowledge acquisition from multiple experts. AZ Mag. 7, 2 (Summer 1985), 32-36. Describes the approach used in the Pride project for integrating expertise from multiple experts.
• 24: Mittal. S., Bobrow. D.G., and de Kleer, J. DARN: A community memory for a diagnosis and repair task. Xerox ISL Lab. Rep., Xerox Corporation, Palo Alto, Calif., 1985. An approach to building a knowledge base that will aid people doing hardware repair.
• 25: Mittal, S., Dym, C.. and Morjaria, M. Pride: An expert system for the design of paper handling systems. In Applications of Knowledge-Based Systems to Engineering Analysis and Design, C.L. Dym. Ed. American Society of Mechanical Engineers, New York, 1985. A general description of Pride, an expert system that aids in the design of paper handling systems.
• 26: Pople, H.E. CADUCEUS: An experimental expert system for medical diagnosis. In The AZ Business, P. Winston and K. Prendergast, Eds. MIT Press, Cambridge, Mass., 1984. Describes the genesis of a medical diagnostic system that is based on the knowledge of one of the best internal medicine specialists in the United States.
• 27: Prerau, D.S. Selection of an appropriate domain for an expert system. AZ Msg. 7, 2 (Summer 1985), 26-30. An account of the criteria used to select expert-system tasks in GTE laboratories.
• 28: Sheil, B. Power tools for programmers. In Znteractive Programming Environments, Barstow et al., Eds. McGraw-Hill, New York, 1984, pp. 19-30. A description of the need for tools that support the exploratory program process.
• 29: Shore, J. The Sachertorte Algorithm. Viking Press, New York, 1985. A nontechnical book describing programming and computers.
• 30: Shortliffe, E.H., Scott, AC., Bischoff, M.B., Campbell, A.B., van Melle, W., and Jacobs, C.D. ONCOCIN: An expert system for oncology protocol management. In Proceedings of the 7th International Joint Conference on Artificial Intelligence. Menlo Park, 1981, pp. 876-881. A good example of how a well-designed user interface facilitates expert-system acceptance.
• 31: Smith, R. On the development of commercial expert systems. AZ Mag. V5, 3 (Fall 1984), 21-34. A description of the development of the dipmeter advisor, an expert system for geologic analysis.
• 32: Mark Stefik, The next knowledge medium, AI Magazine, v.7 n.1, p.34-46, March 1986
• 33: Mark Stefik , Daniel Bobrow, Object-oriented programming: Themes and variations, AI Magazine, v.6 n.4, p.40-62, Winter 1986
• 34: Winograd, T. Extended inference modes in reasoning by computer systems. Artif. Intell. 13 (Winter 1980), 5-26. A very readable account of the need to extend conventional approaches to reasoning to expand the capabilities of AI systems.

INDEX TERMS

Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE

General Terms:
Design, Languages, Management