Advertisement

Efficient proof encoding

  • Uroš Pompe
Theory
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1314)

Abstract

This paper proposes a method of storing the proofs of the learning examples in an efficient manner. FOIL-like top down learners usually store the computed answers of a partially induced clause as a set of ground substitutions. The need for the re-computation of the root part of the SLDNF-tree is reduced that way, but the approach is spaceinefficient when the literals in the clause are nondeterminate. We introduce a weak syntactic language bias that does not practically restrict the hypothesis space. Further more, we present a proof encoding scheme, using a mesh-like data structure, that exploits the properties of this bias to store the computed answers efficiently. We show that such encoding grows at most linearly with respect to the clause length. The result is not influenced by the presence of nondeterminism in the background knowledge.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    W. W. Cohen. Learnability of restricted logic programs. In S. Muggleton, editor, Proceedings of The Third International Workshop on Inductive Logic Programming ILP'93, pages 41–71, Bled, Slovenia, Apr. 1993.Google Scholar
  2. 2.
    W. W. Cohen.Pac-learning a restricted class of recursive logic programs.In S. Muggleton, editor, Proceedings of The Third International Workshop on Inductive Logic Programming ILP'93, pages 73–86, Bled, Slovenia, Apr. 1993.Google Scholar
  3. 3.
    L. De Raedt and S. Dieroski. First order jk-clausal theories are pac-learnable. In S. Wrobel, editor, Proceedings of the Fourth International Workshop on Inductive Logic Programming, Bad Honnef/Bonn, Germany, Sept. 1994.Google Scholar
  4. 4.
    B. Dolšak and S. Muggleton. The application of inductive logic programming to finite elements mesh design. In S. Muggleton, editor, Inductive Logic Programming. Academic Press, 1992.Google Scholar
  5. 5.
    S. Džeroski. Handling noise in inductive logic programming. Master's thesis, University of Ljubljana, Faculty of electrical engineering and computer science, Ljubljana, Slovenia, 1991.Google Scholar
  6. 6.
    M. J. Kearns and U. V. Vazirani. An Introduction to Computational Learning Theory. MIT Press, Cambridge, Massachusetts, 1994.Google Scholar
  7. 7.
    J.-U. Kietz. Some lower bounds for the computational complexity of inductive logic programming. In Proceedings of Sixth European Conference on Machine Learning, pages 115–123, Berlin, Germany, 1993. Springer Verlag.Google Scholar
  8. 8.
    M. Kovačič. Stochastic inductive logic programming. PhD thesis, University of Ljubljana, Faculty of electrical engineering and computer science, Ljubljana, Slovenia, 1995.Google Scholar
  9. 9.
    J. W. Lloyd. Foundations of Logic Programming. Springer Verlag, Berlin, Germany, second edition, 1987.Google Scholar
  10. 10.
    S. Muggleton. Inductive Logic Programming. Academic Press, London, England, 1992.Google Scholar
  11. 11.
    U. Pompe. Restricting the hypothesis space, guiding the search, and handling of redundant information in inductive logic programming. Master's thesis, University of Ljubljana, Faculty of Computer and Information Science, Ljubljana, Slovenia, May 1996.Google Scholar
  12. 12.
    U. Pompe and I. Kononenko. Linear space induction in first order logic with relief. In R. Kruse, R. Viertl, and G. Della Riccia, editors, CISM Lecture notes, (to appear). Springer Verlag, Udine, Italy, Sept. 1994. Presented at: International School for the Synthesis of Expert Knowledge.Google Scholar
  13. 13.
    U. Pompe, M. Kovačič, and I. Kononenko. Sfoil: Stochastic approach to inductive logic programming. In Proceedings of the Second Electrotechnical and Computer Science Conference ERK'93, pages 189–192, Portoroi, Slovenia, Sept. 1993.Google Scholar
  14. 14.
    J. R. Quinlan. Learning logical definitions from relations.Machine Learning, 5:239–266, 1990.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1997

Authors and Affiliations

  • Uroš Pompe
    • 1
  1. 1.Faculty of Computer and Information ScienceUniversity of LjubljanaLjubljanaSlovenia

Personalised recommendations