Learning of Erasing Primitive Formal Systems from Positive Examples

  • Jin Uemura
  • Masako Sato
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 2842)

Abstract

An elementary formal system, EFS for short, introduced by Smullyan is a kind of logic program over strings, and is regarded as a grammar to generate a language. Arikawa and his colleagues introduced some subclasses of EFSs which correspond to Chomsky hierarchy, and showed that they constitute a useful framework for language learning. This paper considers a subclass of EFSs, called primitive EFSs, in view of inductive inference in Gold framework from positive examples. Shinohara showed that the class of languages generated by primitive EFSs is inferable from positive examples, where ε substitutions, i.e., substitutions that may substitute the empty string for variables is not allowed. In the present paper, we consider allowing ε substitutions, and call such EFSs erasing EFSs. It is unknown whether or not the class of erasing pattern languages is learnable from positive examples. An erasing pattern language is a language generated by an erasing EFS with just one axiom.

We first show that the class \(\mathcal{PFSL}\) of languages generated by erasing primitive EFSs does not have finite elasticity, but has M-finite thickness. The notions of finite elasticity and M-finite thickness were introduced by Wright, and Moriyama and Sato, respectively, to present sufficient conditions for learnability from positive examples. Moriyama and Sato showed that a language class with M-finite thickness is learnable from positive examples if and only if for each language in the class, there is a finite tell-tale set of the language. Then we show the class \(\mathcal{PFSL}\) is learnable from positive examples by presenting a finite tell-tale set for each language in the class.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Angluin, D.: Inductive inference of formal languages from positive data. Information and Control 45, 117–135 (1980)MATHCrossRefMathSciNetGoogle Scholar
  2. 2.
    Arikawa, S., Shinohara, T., Yamamoto, A.: Learning elementary formal system. Theoretical Computer Science 95, 97–113 (1992)MATHCrossRefMathSciNetGoogle Scholar
  3. 3.
    Arimura, H., Shinohara, T., Otsuki, S.: Finding minimal generalizations for unions of pattern languages and its application to inductive inference from positive data. In: Enjalbert, P., Mayr, E.W., Wagner, K.W. (eds.) STACS 1994. LNCS, vol. 775, pp. 646–660. Springer, Heidelberg (1994)Google Scholar
  4. 4.
    Gold, E.M.: Language identification in the limit. Information and Computation 10, 447–474 (1967)MATHGoogle Scholar
  5. 5.
    Moriyama, T., Sato, M.: Properties of language classes with finite elasticity. IEICE Transactions on Information and Systems E78-D(5), 532–538 (1995)Google Scholar
  6. 6.
    Motoki, T., Shinohara, T., Wright, K.: The correct definition of finite elasticity: Corrigendum to identification of unions. In: Proceedings of the 4th Annual Workshop on Computational Learning Theory, pp. 375–375 (1991)Google Scholar
  7. 7.
    Mukouchi, Y., Arikawa, S.: Towards a mathematical theory of machine discovery from facts. Theoretical Computer Science 137, 53–84 (1995)MATHCrossRefMathSciNetGoogle Scholar
  8. 8.
    Mukouchi, Y.: Note on Learnability of Subclasses of Erasing Elementary Formal Systems from Positive Examples (in preparation)Google Scholar
  9. 9.
    Reidenbach, D.: Result on Inductive Inference of Extended Pattern Languages. In: Cesa-Bianchi, N., Numao, M., Reischuk, R. (eds.) ALT 2002. LNCS (LNAI), vol. 2533, pp. 308–320. Springer, Heidelberg (2002)CrossRefGoogle Scholar
  10. 10.
    Sato, M.: Inductive Inference of Formal Languages. Bulletin of Informatics and Cybernetics 27(1), 85–106 (1995)MATHMathSciNetGoogle Scholar
  11. 11.
    Shinohara, T.: Polynomial time inference of extended regular pattern languages. In: Goto, E., Nakajima, R., Yonezawa, A., Nakata, I., Furukawa, K. (eds.) RIMS 1982. LNCS, vol. 147, pp. 115–127. Springer, Heidelberg (1983)Google Scholar
  12. 12.
    Shinohara, T.: Inductive inference of formal systems from positive data. Bulletin of Information and Cybernetics 22, 9–18 (1986)MATHMathSciNetGoogle Scholar
  13. 13.
    Shinohara, T.: Rich classes inferable from positive data. Information and Control 108, 175–186 (1994)MATHMathSciNetGoogle Scholar
  14. 14.
    Smullyan, R.M.: Theory of Formal Systems. Princeton University Press, Princeton (1961)MATHGoogle Scholar
  15. 15.
    Uemura, J., Sato, M.: Compactness and Learning of Unions of Erasing Regular Pattern Languages. In: Cesa-Bianchi, N., Numao, M., Reischuk, R. (eds.) ALT 2002. LNCS (LNAI), vol. 2533, pp. 293–307. Springer, Heidelberg (2002)CrossRefGoogle Scholar
  16. 16.
    Wright, K.: Identification of unions of languages drawn from positive data. In: Proceedings of the 2nd Annual Workshop on Computational Learning Theory, pp. 328–333 (1989)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2003

Authors and Affiliations

  • Jin Uemura
    • 1
  • Masako Sato
    • 1
  1. 1.Department of Mathematics and Information SciencesOsaka Prefecture UniversitySakai, OsakaJapan

Personalised recommendations