Abstract
Knowledge has traditionally been considered to have a beneficial effect on the performance of problem solvers but recent studies indicate that knowledge acquisition is not necessarily a monotonically beneficial process, because additional knowledge sometimes leads to a deterioration in system performance. This paper is concerned with the problem of harmful knowledge: that is, knowledge whose removal would improve a system's performance. In the first part of the paper a unifying framework, called the information filtering model, is developed to define the various alternative methods for eliminating such knowledge from a learning system where selection processes, called filters, may be inserted to remove potentially harmful knowledge. These filters are termed selective experience, selective attention, selective acquisition, selective retention, and selective utilization. The framework can be used by developers of learning systems as a guide for selecting an appropriate filter to reduce or eliminate harmful knowledge.
In the second part of the paper, the framework is used to identify a suitable filter for solving a problem caused by the acquisition of harmful knowledge in a learning system called LASSY. LASSY is a system that improves the performance of a PROLOG interpreter by utilizing acquired domain specific knowledge in the form of lemmas stating previously proved results. It is shown that the particular kind of problems that arise with this system are best solved using a novel utilization filter that blocks the use of lemmas in attempts to prove subgoals that have a high probability of failing.
Article PDF
Similar content being viewed by others
References
Aha, D.W., Kibler, D., amp; Albert, M.K. (1991). Instance-based learningalgorithms. Machine Learning, 6, 37–66.
Brodie, M.L., amp; Mattias, J. (1986). On integrating logic programming and databases.In L. Kerschberg (Ed.), Expert database systems. Menlo Park,CA: Benjamin/Cummings.
Buchanan, B.G., amp; Feigenbaum, E.A. (1982). Foreword. In R. Davis amp; D.B. Lenat (Eds.), Knowledge based systems in artificial intelligence. NewYork: McGraw-Hill.
Clark, K.L. amp; McCabe, F. (1979). The control facilities of ic-prolog. In D. Michie(Ed.), Expert systems in the microelectronic age. Edinburgh, Scotland: Universityof Edinburgh.
Debray, S.K., amp; Warren, D.S. (1988). Automatic mode inference for logicprograms. The Journal of Logic Programming, 5, 207–229.
Dejong, G., amp; Mooney, R. (1986). Explanation-based learning. An alternativeview. Machine Learning, 1, 145–176.
Dietterich. T.G., amp; Michalski, R.S. (1981). Inductive learning of structuraldescription: Evaluation criteria and comparative review of selected method. ArtificialIntelligence, 16, 257–294.
Etzioni, O. (1988). Hypothesis filtering: A practical approach to reliable learning.Proceedings of The Fifth International Conference on Machine Learning (pp.416–429). Ann Arbor, MI: Morgan Kaufmann.
Fikes, R.E., Hart, P.E., amp; Nilsson, N.J. (1972). Learning and executinggeneralized robot plans. Artificial Intelligence, 3, 251–288.
Gallaire, H., amp; Lasserre, C. (1982). Metalevel control for logic programs. InK.L. Clark amp; S.A. Tarnlund (Eds.), Logic programming. NewYork: Academic Press.
Gallaire, H., amp; Minker, J. (1978). Logic and data bases. NewYork: Plenum.
Gennari, J.H. (1989). Focused concept formation. Proceedings of The SixthInternational Workshop on Machine Learning (pp. 379–382). Ithaca, NY: MorganKaufmann.
Holland, J.H. (1975). Adaption in natural and artificial systems. Ann Arbor,MI: University of Michigan Press.
Holland, J.H. (1986). Escaping brittleness: The possibilities of general-purpose learningalgorithms applied to parallel rule-based systems. In R.S. Michalski, J.G. Carbonell,amp; T.M. Mitchell (Eds.), Machine learning: An artificial intelligenceapproach (vol. 2). Los Altos, CA: Morgan Kaufmann.
Iba, G.A. (1989). A heuristic approach to the discovery of macro-operators.Machine Learning, 3, 285–317.
Keller, M.R. (1987). Concept learning in context. Proceedings of FourthInternational Workshop on Machine Learning (pp. 482–487). Irvine, CA: MorganKaufmann.
Korf, R.E. (1985). Learning to solve problems by searching for macro-operators.Boston: Pitman.
Kowalski, R.A. (1979). Logic for problem solving. New York: Elseveir NorthHolland.
Kowalski, R.A. (1985). Directions for logic programming. Proceedings of IEEESymposium On Logic Programming.
Laird, J.E., Rosenbloom, P.S., amp; Newell, A. (1986). Chunking in soar: Theanatomy of a general learning mech-anism. Machine Learning, 1,11–46.
Lenat, D.B. (1983). The role of heuristics in learning by discovery: Three case studies.In R.S. Michalski, J.G. Carbonell, amp; T.M. Mitchell (Eds.), Machinelearning: An artificial intelligence appraoch (vol. 1). Palo Alto, CA: Tioga.
Lindsay, R.K., Buchanan, B.G., Feigenbaum, E.H., amp; Lederberg, J. (1980).Applications of artificial intelligence for organic chemistry: The Dendral project.New York: McGraw-Hill.
Lloyd, J.W. (1984). Foundations of logic programming. New York: Springer-Verlag.
Markovitch, S. (1989). Information filtering: Selection mechanisms in learningsystems. Doctoral dissertation, EECS Department, University of Michigan.
Markovitch, S., amp; Scott, P.D. (1988). The role of forgetting in learning.Proceedings of The Fifth International Conference on Machine Learning (pp.459–465). Ann Arbor, MI: Morgan Kaufmann.
Markovitch, S., amp; Scott, P.D. (1989a). Automatic ordering of subgoals—a machinelearning approach. Proceedings of North American Conference on LogicProgramming (pp. 224–240). Ithaca, NY: MIT Press.
Markovitch, S., amp; Scott, P.D. (1989b). Post-prolog: Structured partially orderedprolog (TR CMI-89-018) Ann Arbor, MI: Center for Machine Intelligence.
Markovitch, S., amp; Scott, P.D. (1989c). Utilization filtering: A method for reducingthe inherent harmfulness of deductively learned knowledge. Proceedings of TheEleventh International Joint Conference for Artificial Intelligence (pp. 738–743).Detroit, MI: Morgan Kaufmann.
Minton, S. (1985). Selectively generalizing plans for problem solving. Proceedingsof The Ninth International Joint Conference on Artificial Intelligence(pp. 596–599). Los Angeles, CA: Morgan Kaufmann.
Minton, S. (1988a). Learning search control knowledge: An explanation-basedapproach. Boston, MA: Kluwer.
Minton, S. (1988b). Quantitative results concerning the utility of explanation-basedlearning. Proceedings of Seventh National Conference on Artificial Intelligence(pp. 564–569). Morgan Kaufmann.
Mitchell, T.M. (1982). Generalization as search. Artificial Intelligence,,18, 203–226.
Mitchell, T.M., Keller, R.M., amp; Kedar-Cabelli, ST. (1986). Explanation-basedgeneralization: A unifying view.Machine Learning,, 1, 47–80.
Mitchell, T.M., Utgoff, P.E., amp; Banerji, R. (1983). Learning by experimentation:Acquiring and refining problem-solving heuristics. In R.S. Michalski, J.G. Carbonell,amp; T.M. Mitchell (Eds.), Machine learning: An artificial intelligenceapproach. Palo Alto, CA: Tioga.
Mooney, R. (1989). The effect of rule use on the utility of explanation-based learning.Proceedings of The Eleventh International Joint Conference for ArtificialIntelligence (pp. 725–730). Detroit, MI: Morgan Kaufmann.
Naish, L. (1985). Automatic control for logic programs. The Journal of LogicProgramming, 3, 167–183.
Parker, S.D., Carey, M., Golshani, R, Jarke, M., Sciore, E., amp; Walker, A.(1986). Logic programming anddatabases. Expert Database Systems—Proceedings from The First InternationalWorkshop (pp. 35–48). MenloPark, CA: Benjamin/Cummings.
Pearl, J. (1984). Heuristics: Intelligent search strategies for computer problemsolving. Reading, MA: Addison-Wesley.
Pereira, L.M., amp; Porto, A. (1982). Selective backtracking. In K.L. Clark amp; S.A. Tarnlund (Eds.), Logic programming. New York: Academic Press.
Porter, B.W., Bareiss, R., amp; Holte, R.C. (1990). Concept learning and heuristicclassification in weak-theorydomains. Artificial Intelligence, 45, 229–263.
Prieditis, A.E., amp; Mostow, J. (1987). Prolearn: Toward a prolog interpreter thatlearns. Proceedings of The Sixth National Conference on Artificial Intelligence(pp. 494–498). Seattle, WA: Morgan Kaufmann.
Quinlan, J.R. (1986). Induction of decision trees. Machine Learning,1, 81–106.
Ruff, R.A., amp; Dietrich, T.G. (1989). What good are experiments? Proceedingsof The Sixth International Workshop on Machine Learning (pp.109–112). Ithaca, NY: Morgan Kaufmann.
Samuel, A.L. (1959). Some studies in machine learning using the game of checkers.IBM Journal, 3, 211–229.
Scott, P.D., amp; Markovitch, S. (1989a). Learning novel domains through curiosity andconjecture. Proceedings of International Joint Conference for ArtificialIntelligence (pp. 669–674). San Mateo, CA: Morgan Kaufmann.
Scott, P.D., amp; Markovitch, S. (1989b). Uncertainty based selection of learningexperiences. Proceedings of The Sixth International Workshop on MachineLearning (pp. 368–371). Ithaca, NY: Morgan Kaufmann.
Scott, P.D., amp; Markovitch, S. (1991). Representation generation in an exploratorylearning system. In D. Fisheramp; M. Pizzani (Eds.), Concept formation: Knowledge and experience inunsupervised learning. Morgan Kaufmann.
Shannon, C.E., amp; Weaver, W. (1949). The Mathematical Theory ofCommunication. University of Illinois Press.
Tambe, M., amp; Newell, A. (1988). Some chunks are expensive. Proceedingsof The Fifth International Conference on Machine Learning (pp.451–458). Ann Arbor, MI: Morgan Kaufmann.
Utgoff, P.E. (1989). Improved training via incremental learning. Proceedings ofThe Sixth International Workshop on Machine Learning (pp. 362–365). Ithaca,NY: Morgan Kaufmann.
Valiant, L.G. (1984). A theory of the learnable. Communications of the ACM,27, 1134–1142.
Wilkins, D.C., amp; Ma, Y. (1989). Sociopathic knowledge bases (TechnicalReport UIUCDCS-R-89-1538). Urbana-Champaign,IL: University of Illinois, Department of Computer Science.
Winston, P.H. (1975). Learning structural descriptions from examples. In P. Winston(Ed.), The psychology of computer vision. New York: McGraw-Hill.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Markovitch, S., Scott, P.D. Information Filtering: Selection Mechanisms in Learning Systems. Machine Learning 10, 113–151 (1993). https://doi.org/10.1023/A:1022614725002
Issue Date:
DOI: https://doi.org/10.1023/A:1022614725002