Advertisement

A Lower Bound for Learning Distributions Generated by Probabilistic Automata

  • Borja Balle
  • Jorge Castro
  • Ricard Gavaldà
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 6331)

Abstract

Known algorithms for learning PDFA can only be shown to run in time polynomial in the so-called distinguishability μ of the target machine, besides the number of states and the usual accuracy and confidence parameters. We show that the dependence on μ is necessary for every algorithm whose structure resembles existing ones. As a technical tool, a new variant of Statistical Queries termed L ∞ -queries is defined. We show how these queries can be simulated from samples and observe that known PAC algorithms for learning PDFA can be rewritten to access its target using L ∞ -queries and standard Statistical Queries. Finally, we show a lower bound: every algorithm to learn PDFA using queries with a resonable tolerance needs a number of queries larger than (1/μ) c for every c < 1.

Keywords

Hide Markov Model Target Distribution Alphabet Size Statistical Query Query Algorithm 
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.
    Abe, N., Warmuth, M.K.: On the computational complexity of approximating distributions by probabilistic automata. Mach. Learn. 9(2-3), 205–260 (1992)zbMATHCrossRefGoogle Scholar
  2. 2.
    Ben-David, S., Lindenbaum, M.: Learning distributions by their density levels: A paradigm for learning without a teacher. J. Comput. Syst. Sci. 55(1), 171–182 (1997)zbMATHCrossRefMathSciNetGoogle Scholar
  3. 3.
    Carrasco, R.C., Oncina, J.: Learning deterministic regular grammars from stochastic samples in polynomial time. RAIRO (Theoretical Informatics and Applications) 33(1), 1–20 (1999)zbMATHCrossRefMathSciNetGoogle Scholar
  4. 4.
    Castro, J., Gavaldà, R.: Towards feasible PAC-learning of probabilistic deterministic finite automata. In: Clark, A., Coste, F., Miclet, L. (eds.) ICGI 2008. LNCS (LNAI), vol. 5278, pp. 163–174. Springer, Heidelberg (2008)CrossRefGoogle Scholar
  5. 5.
    Cesa-Bianchi, N., Lugosi, G.: Prediction, Learning, and Games. Cambridge University Press, New York (2006)zbMATHCrossRefGoogle Scholar
  6. 6.
    Clark, A., Thollard, F.: PAC-learnability of probabilistic deterministic finite state automata. Journal of Machine Learning Research (2004)Google Scholar
  7. 7.
    Denis, F., Esposito, Y., Habrard, A.: Learning rational stochastic languages. In: Lugosi, G., Simon, H.U. (eds.) COLT 2006. LNCS (LNAI), vol. 4005, pp. 274–288. Springer, Heidelberg (2006)CrossRefGoogle Scholar
  8. 8.
    Dupont, P., Denis, F., Esposito, Y.: Links between probabilistic automata and hidden markov models: probability distributions, learning models and induction algorithms. Pattern Recognition 38(9), 1349–1371 (2005)zbMATHCrossRefGoogle Scholar
  9. 9.
    Gavaldà, R., Keller, P.W., Pineau, J., Precup, D.: PAC-learning of markov models with hidden state. In: Fürnkranz, J., Scheffer, T., Spiliopoulou, M. (eds.) ECML 2006. LNCS (LNAI), vol. 4212, pp. 150–161. Springer, Heidelberg (2006)CrossRefGoogle Scholar
  10. 10.
    Guttman, O., Vishwanathan, S.V.N., Williamson, R.C.: Learnability of probabilistic automata via oracles. In: Jain, S., Simon, H.U., Tomita, E. (eds.) ALT 2005. LNCS (LNAI), vol. 3734, pp. 171–182. Springer, Heidelberg (2005)CrossRefGoogle Scholar
  11. 11.
    Hsu, D., Kakade, S.M., Zhang, T.: A spectral algorithm for learning hidden markov models. CoRR abs/0811.4413 (2008)Google Scholar
  12. 12.
    Kearns, M.: Efficient noise-tolerant learning from statistical queries. J. ACM 45(6), 983–1006 (1998)zbMATHCrossRefMathSciNetGoogle Scholar
  13. 13.
    Kearns, M.J., Mansour, Y., Ron, D., Rubinfeld, R., Schapire, R.E., Sellie, L.: On the learnability of discrete distributions. In: STOC, pp. 273–282 (1994)Google Scholar
  14. 14.
    Lyngsø, R.B., Pedersen, C.N.S.: The consensus string problem and the complexity of comparing hidden markov models. J. Comput. Syst. Sci. 65(3), 545–569 (2002)CrossRefGoogle Scholar
  15. 15.
    Palmer, N., Goldberg, P.W.: PAC-learnability of probabilistic deterministic finite state automata in terms of variation distance. Theor. Comput. Sci. 387(1), 18–31 (2007)zbMATHMathSciNetGoogle Scholar
  16. 16.
    Ron, D., Singer, Y., Tishby, N.: On the learnability and usage of acyclic probabilistic finite automata. J. Comput. Syst. Sci. 56(2), 133–152 (1998)zbMATHCrossRefMathSciNetGoogle Scholar
  17. 17.
    Vidal, E., Thollard, F., de la Higuera, C., Casacuberta, F., Carrasco, R.C.: Probabilistic finite-state machines - part I. IEEE Trans. Pattern Anal. Mach. Intell. 27(7), 1013–1025 (2005)CrossRefGoogle Scholar
  18. 18.
    Vidal, E., Thollard, F., de la Higuera, C., Casacuberta, F., Carrasco, R.C.: Probabilistic finite-state machines - part II. IEEE Trans. Pattern Anal. Mach. Intell. 27(7), 1026–1039 (2005)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2010

Authors and Affiliations

  • Borja Balle
    • 1
  • Jorge Castro
    • 1
  • Ricard Gavaldà
    • 1
  1. 1.Universitat Politècnica de CatalunyaBarcelona

Personalised recommendations