Explanation-based generalization and constraint propagation with interval labels

  • Kai Zercher
Part 4: Theorem Proving And EBL
Part of the Lecture Notes in Computer Science book series (LNCS, volume 482)


Two ways of applying EBG to constraint propagation with interval labels are presented. The first method, CP-EBG-1, is described by a straightforward use of a Prolog EBG implementation. The second, CP-EBG-2, performs two phases: First, constraint propagation is done and, using EBG, a generalized final labelling is derived but no extra conditions are learned. Second, constraint propagation is again performed using the final labellings of phase 1 as the initial labelling. This time, conditions are learned which form the desired concept description.

It is shown that CP-EBG-2 learns more general concept descriptions than CP-EBG-1. A proof is outlined that CP-EBG-2 produces correct concept descriptions for the class of constraints using linear equations and interval arithmetic. Central to this proof- and to possible proofs for other constraint classes - is the notion of a moderate generalization. It guarantees that a generalization which was learned from one instance and which is now used in a new situation, does not lead to the exclusion of any solution for this new situation.


Explanation-based generalization constraint propagation interval labels moderate generalization 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. Davis, E., (1987). Constraint propagation with interval labels. Artificial Intelligence, 32,281–331.CrossRefGoogle Scholar
  2. DeJong, G., & Mooney, R. (1986). Explanation-based learning: An alternative view. Machine Learning, 1, 145–176.Google Scholar
  3. Güsgen, H.-W., & Hertzberg, J. (1988). Some fundamental properties of local constraint propagation. Artificial Intelligence, 36, 237–247.CrossRefGoogle Scholar
  4. Kedar-Cabelli, S. T., & McCarty, L. T. (1987). Explanation-based generalization as resolution theorem proving. Proceedings of the Fourth International Workshop on Machine Learning (pp. 383–389). Irvine,CA: Morgan Kaufmann.Google Scholar
  5. Keller, R. M. (1987). Defining operationality for explanation-based learning. Proceedings of the Sixth National Conference on Artificial Intelligence (pp. 482–487).Google Scholar
  6. Minton, S., (1988). Quantitative results concerning the utility of explanation-based learning. Proceedings of the Seventh National Conference on Artificial Intelligence (pp. 564–569).Google Scholar
  7. Mitchell, T. M., Keller, R., & Kedar-Cabelli, S. (1986). Explanation-based generalization: a unifying view. Machine Learning, 1, 47–80.Google Scholar
  8. Mooney, R. J., & Bennett, S. W. (1986). A domain independent explanation-based generalizer. Proceedings of the Fifth National Conference on Artificial Intelligence (pp. 551–555). Philadelphia, PA: Morgan Kaufmann.Google Scholar
  9. Papadimitriou, C. H., & Steiglitz, K. (1982). Combinatorical Optimization: Algorithms and Complexity. Englewood Cliffs, NJ: Prentice-Hall.Google Scholar
  10. Zercher, K., (1988). Model-based learning of rules for error diagnosis. In Hoeppner, W. (Ed.), Proceedings of the 12th German workshop on artificial intelligence (GWAI88) (pp. 196–205). Springer.Google Scholar
  11. Zercher, K., (1990 a). Constructing decision trees from examples and their explanation-based generalizations. In Marburger, H., (Ed.), Proceedings of the 14th German workshop on artificial intelligence (GWAI90) (pp. 267–276). Springer.Google Scholar
  12. Zercher, K., (1990 b). Learning efficient rules and decision trees for error diagnosis of robot operations. Unpblished working paper.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1991

Authors and Affiliations

  • Kai Zercher
    • 1
    • 2
  1. 1.Siemens AG, ZFE IS INF 33München 83
  2. 2.TU München, Institut für InformatikMünchen 80Germany

Personalised recommendations