Advertisement

Heuristics for Empirical Discovery

  • Pat Langley
  • Herbert A. Simon
  • Gary L. Bradshaw
Part of the Symbolic Computation book series (SYMBOLIC)

Abstract

In this paper, we review our experiences with the BACON project, which has focused on empirical methods for discovering numeric laws. The six successive versions of BACON have employed a variety of discovery methods, some very simple and others quite sophisticated. We examine methods for discovering a functional relation between two numeric terms, including techniques for detecting monotonic trends, finding constant differences, and hill-climbing through a space of parameter values. We also consider methods for discovering complex laws involving many terms, some of which build on techniques for finding two-variable relations. Finally, we introduce the notions of intrinsic properties and common divisors, and examine methods for inferring intrinsic values from symbolic data. In each case, we describe the various techniques in terms of the search required to discover useful laws.

Keywords

Intrinsic Property Dependent Term Nominal Term Oxygen Nitrogen Independent Term 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Langley P: BACON.1: A general discovery system. Proceedings of the Second National Conference of the Canadian Society for Computational Studies of Intelligence, 1978, pp 173–180Google Scholar
  2. 2.
    Langley P: Descriptive Discovery Processes: Experiments in Baconian Science. Ph. D. Th., Department of Psychology, Carnegie-Mellon University, 1979Google Scholar
  3. 3.
    Langley P: Data-driven discovery of physical laws. Cognitive Science 5, 1981, 31–54CrossRefGoogle Scholar
  4. 4.
    Bradshaw G, Langley P, Simon HA: BACON.4: The discovery of intrinsic properties. Proceedings of the Third National Conference of the Canadian Society for Computational Studies of Intelligence, 1980, pp 19–25Google Scholar
  5. 5.
    Langley P, Bradshaw G, Simon HA: Rediscovering chemistry with the BACON system. In Machine Learning: An Artificial Intelligence Approach, Michalski RS, Carbonell JG, Mitchell TM, (eds), Tioga Press, Palo Alto, CA, 1983Google Scholar
  6. 6.
    Langley P, Bradshaw G, Simon HA: BACON.5: The discovery of conservation laws. Proceedings of the Seventh International Joint Conference on Artificial Intelligence, 1981, pp 121–126Google Scholar
  7. 7.
    Langley P, Bradshaw G, Simon HA: Data-driven and expectation-driven discovery of empirical laws. Proceedings of the Fourth National Conference of the Canadian Society for Computational Studies of Intelligence, 1982, pp 137–143Google Scholar
  8. 8.
    Langley P, Bradshaw G, Zytkow J, Simon HA: Three facets of scientific discovery. Proceedings of the Eighth International Joint Conference on Artificial Intelligence, 1983, pp 465–468Google Scholar
  9. 9.
    Langley P, Zytkow J, Simon HA, Bradshaw GL: The search for regularity: Four aspects of scientific discovery. In Machine Learning, Volume 2, Michalski RS, Carbonell JG, Mitchell TM, (eds), Tioga Press, Palo Alto, CA, 1984Google Scholar
  10. 10.
    Gerwin DG: “Information processing, data inferences, and scientific generalization.” Behavioral Science 19, 1974, 314–325CrossRefGoogle Scholar
  11. 11.
    Collins JS: A regression analysis program incorporating heuristic term: selection. In Machine Intelligence 2, Dale E, Michie D, (ed), American Elsevier, New York, 1968Google Scholar
  12. 12.
    Feigenbaum EA, Buchanan BG, Lederberg J: On generality and problem solving: A case study using the DENDRAL program. In Machine Intelligence 6, Edinburgh University Press, Edinburgh, 1971Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1987

Authors and Affiliations

  • Pat Langley
  • Herbert A. Simon
  • Gary L. Bradshaw
    • 1
  1. 1.Carnegie-Mellon UniversityPittsburghUSA

Personalised recommendations