Integrating Quantitative and Qualitative Discovery: The ABACUS System
Tóm tắt
Most research on inductive learning has been concerned with qualitative learning that induces conceptual, logic-style descriptions from the given facts. In contrast, quantitative learning deals with discovering numerical laws characterizing empirical data. This research attempts to integrate both types of learning by combining newly developed heuristics for formulating equations with the previously developed concept learning method embodied in the inductive learning program AQ11. The resulting system, ABACUS, formulates equations that bind subsets of observed data, and derives explicit, logic-style descriptions stating the applicability conditions for these equations. In addition, several new techniques for quantitative learning are introduced. Units analysis reduces the search space of equations by examining the compatibility of variables' units. Proportionality graph search addresses the problem of identifying relevant variables that should enter equations. Suspension search focusses the search space through heuristic evaluation. The capabilities of ABACUS are demonstrated by several examples from physics and chemistry.
Tài liệu tham khảo
Aho, A.V., Hopcroft, J.E., & Ullman, J.D. (1974).The design and analysis of computer algorithms. Menlo Park,CA: Addison-Wesley.
Becker, J.M. (1985).Inductive learning of decision rules with exceptions: Methodology and experimentation. Master's thesis. Department of Computer Science, University of Illinois, Champaign-Urbana, Illinois.
Chatterjee, S., & Price, B. (1977). Regression analysis by example. New York: John Wiley and Sons.
Condon, EU., Odishaw, H. (Eds.). (1967).Handbook of physics(2nd ed.).New York: McGraw-Hill.
Daniel, C., & Wood, F.S. (1971).Fitting equations to data. New York: Wiley-Interscience, John Wiley & Sons
de Kleer, J. (1975).Qualitative and quantitative knowledge in classical mechanics. Master's thesis (TR-.352), Massachusetts Institute of Technology, Cambridge, Massachusetts.
Dietterich, T., & Michalski, R.S. (1981).Inductive learning of structural descriptions: Evaluation criteria and comparative review of selected methods.Artificial Intelligence, 16(3), 257–294.
El-Shafei, N. (1986).Quantitative discovery and reasoning about failure mechanisms in pavement. Unpublished manuscript, Artificial Intelligence Laboratory, Massachusetts Institute of Technology, Cambridge, Massachusetts.
Falkenhainer, B.C. (1984).ABACUS: Adding domain constraints to quantitative scientific discovery(Technical Report UIUCDCS-F–84–927, ISG 84–7).Department of Computer Science, University of Illinois, Champaign-Urbana, Illinois.
Falkenhainer, B.C. (1985a).Proportionality graphs, units analysis, and domain constraints: Improving the power and efficiency of the scientific discovery process.Proceedings of the Ninth International Joint Conference on Artificial Intelligence(pp.552–554).Los Angeles, CA: Morgan-Kaufmann.
Falkenhainer, B.C. (1985b).Quantitative empirical learning: An analysis and methodology. Master's thesis (UIUCDCS-F–85–947, ISG 85–16), Department of Computer Science, University of Illinois, Champaign-Urbana, Illinois.
Forbus, K.D. (1984).Qualitative process theory. Doctoral dissertation (TR-798), Department of Computer Science, Massachusetts Institute of Technology, Cambridge, Massachusetts.
Hajek, P., & Havranek, T. (1978).The GUHA method: Its aims and techniques.International Journal on Man Machine Studies, 10, 3-22.
Huntley, H.E. (1952).Dimensional analysis.London: MacDonald.
Kokar, M. (1981). A procedure of identification of laws in empirical sciences.Systems Science,7, 32–41.
Kokar, M. (1986). Determining arguments of invariant functional descriptions.Machine Learning,1, 403–422.
Kowalik, J.S. (1986).Coupling symbolic and numerical computing in expert systems. Amsterdam: North-Holland.
Langhaar, H.L. (1951).Dimensional analysis and theory of models. New York: John Wiley & Sons.
Langley, P. (1979). Rediscovering physics with BACON.3. Proceedings of the Sixth International Joint Conference on Artificial Intelligence (pp.505–507). Tokyo,Japan: Morgan-Kaufmann.
Langley, P. (1981). Data-driven discovery of physical laws. Cognitive Science,5,31–54.
Langley, P., Bradshaw, G.L., & Simon, H.A. (1981). BACON.5:The discovery of conservation laws. Proceedings of the Seventh International Joint Conference on Artificial Intelligence(pp.121–126). Vancouver,BC,Canada: Morgan-Kaufmann.
Langley, P., Bradshaw, G.L., & Simon, H.A. (1983a). Rediscovering chemistry with the BACON system. In R.S. Michalski, J.G. Carbonell, & T.M. Mitchell (Eds.),Machine learning: An artificial intelligence approach. Palo Alto,CA: Tioga.
Langley, P., Bradshaw, G.H., Simon, H.A., & Zytkow, J. (1983b). Mechanisms for qualitative and quantitative discovery.Proceedings of the International Machine Learning Workshop. Monticello, Illinois.
Langley, P., Zytkow, J., Simon, H.A., & Bradshaw, G.L. (1986). The search for regularity: Four aspects of scientific discovery. In R.S. Michalski, J.G. Carbonell, & T.M. Mitchell (Eds.),Machine learning: An artificial intelligence approach(Vol.2). Los Altos,CA: Morgan Kaufmann.
Michalski, R.S. (1980). Knowledge acquisition through conceptual clustering: A theoretical framework and an algorithm for partitioning data into conjunctive concepts.Policy Analysis and Information System,4,219–244.
Michalski, R.S. (1983). A theory and methodology of inductive learning. In R.S. Michalski, J.G. Carbonell, & T.M. Mitchell (Eds.),Machine learning: An artificial intelligence approach. Palo Alto, CA: Tioga.
Michalski, R.S., & Larson, J.B. (1978). Selection of most representative training examples and incremental generation of VLI hypotheses: The underlying methodology and the description of programs ESEL and AQ11(Technical Report UIUCDCS-R–78–867).Department of Computer Science, University of Illinois,Champaign-Urbana,Illinois.
Mitchell, T.M. (1982). Generalization as search.Artificial Intelligence,18,203–226.
Roller, D.E., & Blum, R. (1981). Physics:Mechanics,waves,and thermodynamics(Vol.1). San Francisco,CA: Holden-Day.
Stepp III, R.E. (1984). Conjunctive conceptual clustering: A methodology and experimentation.Doctoral dissertation (UIUCDCS-R–84–1189),Department of Computer Science, University of Illinois,Champaign-Urbana,Illinois.
Weast, R.C.( Ed.). (1984).CRC handbook of chemistry and physics(65th ed.)CRC Press.
Zagoruiko, N.G. (1976). Empirical prediction algorithms. In J.C. Simon (Ed.),Computer oriented learning process(pp.581–595).Leyden:Noordhoff.
Zagoruiko, N.G., Elkina, V.N., & Lbov, G.S. (1985).Algorithms for revealing empirical laws.Nauka Publishing House, Siberian Division of the Soviet Union Academy of Sciences.