Scientific Modeling: A Multilevel Feedback Process
Model construction is one of the key scientific activities. In distinction to the majority of the previous machine discovery systems, model formation applies in theory-rich context. Our long-term goal is automation of model construction. This paper reports on exploratory work towards that goal. We start from the distinction between models and theories, which is critical to the presented approach. We also distinguish between modeling and two scientific activities, which are different but which support modeling: construction of operational definitions and experimentation. Then we present the basic steps of scientific model construction, outlining data structures and an algorithm which, using a number of feedback loops, incrementally develops a model of a natural phenomenon. A walk through example is used to present the algorithm: motion of a cylinder that rolls downwards on an inclined plane.
KeywordsMercury Electromagnetism Cylin
Unable to display preview. Download preview PDF.
- Bridgman, P.W., 1927, The Logic of Modern Physics, The Macmillan Company.Google Scholar
- Carnap, R., 1936–1937, Testability and meaning, Philosophy of Science 3:420–471 and 4:2–40.Google Scholar
- Gordon, A., Sleeman, D.H., and Edwards, P., 1995, Informal qualitative models: a systematic approach to their generation, in: Working Notes of AAAI Spring Symposium: Systematic Methods of Scientific Discovery, p. 18–22.Google Scholar
- Huang, K-M, and Zytkow, J.M., 1996, Robotic discovery: the dilemmas of empirical equations, in: Proceedings of the Fourth International Workshop on Rough Sets, Fuzzy Sets, and Machine Discovery, The University of Tokyo, Tokyo, pp. 217–224.Google Scholar
- Kocabas, S., 1991, Conflict resolution as discovery in particle physics, Machine Learning 6:277–309.Google Scholar
- Langley, P., Simon, H. A., Bradshaw, G. L., and Zytkow, J. M., 1987, Scientific Discovery: Computational Explorations of the Creative Processes, MIT Press, Cambridge, MA.Google Scholar
- Nordhausen, B., and Langley, P., 1993, An integrated framework for empirical discovery, Machine Learning 12:17–47.Google Scholar
- Rajamoney, S.A., 1993, The design of discrimination experiments, Machine Learning 12:185–203.Google Scholar
- Sleeman, D.H., Stacey, M.K., Edwards, P., and Gray, N.A.B., 1989, An architecture for theory-driven scientific discovery, in: Proceedings of the 4 th European Working Session on Learning (EWSL-89), K. Morik, ed., Pitman, London, 11–23.Google Scholar
- Zytkow, J.M., 1990, Deriving laws by analysis of processes and equations, in: Computational Models of Scientific Discovery and Theory Formation, P. Langley and J. Shrager, eds., Morgan Kaufmann, San Mateo (CA), pp. 129–156.Google Scholar
- Zytkow, J.M., Zhu, J., and Zembowicz, R., 1992, Operational definition refinement: a discovery process, in: Proceedings of the Tenth National Conference on Artificial Intelligence, AAAI Press, pp. 76–81.Google Scholar