Prediction with Expert Advice by Following the Perturbed Leader for General Weights

  • Marcus Hutter
  • Jan Poland
Part of the Lecture Notes in Computer Science book series (LNCS, volume 3244)

Abstract

When applying aggregating strategies to Prediction with Expert Advice, the learning rate must be adaptively tuned. The natural choice of \(\sqrt{\mbox{complexity/current loss}}\) renders the analysis of Weighted Majority derivatives quite complicated. In particular, for arbitrary weights there have been no results proven so far. The analysis of the alternative “Follow the Perturbed Leader” (FPL) algorithm from [KV03] (based on Hannan’s algorithm) is easier. We derive loss bounds for adaptive learning rate and both finite expert classes with uniform weights and countable expert classes with arbitrary weights. For the former setup, our loss bounds match the best known results so far, while for the latter our results are new.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [ACBG02]
    Auer, P., Cesa-Bianchi, N., Gentile, C.: Adaptive and self-confident on-line learning algorithms. Journal of Computer and System Sciences 64(1), 48–75 (2002)MATHCrossRefMathSciNetGoogle Scholar
  2. [AG00]
    Auer, P., Gentile, C.: Adaptive and self-confident on-line learning algorithms. In: Proceedings of the 13th Conference on Computational Learning Theory, pp. 107–117. Morgan Kaufmann, San Francisco (2000)Google Scholar
  3. [CB97]
    Cesa-Bianchi, N., et al.: How to use expert advice. Journal of the ACM 44(3), 427–485 (1997)MATHCrossRefMathSciNetGoogle Scholar
  4. [FS97]
    Freund, Y., Schapire, R.E.: A decision-theoretic generalization of on-line learning and an application to boosting. Journal of Computer and System Sciences 55(1), 119–139 (1997)MATHCrossRefMathSciNetGoogle Scholar
  5. [Gen03]
    Gentile, C.: The robustness of the p-norm algorithm. Machine Learning 53(3), 265–299 (2003)MATHCrossRefMathSciNetGoogle Scholar
  6. [Han57]
    Hannan, J.: Approximation to Bayes risk in repeated plays. In: Dresher, M., Tucker, A.W., Wolfe, P. (eds.) Contributions to the Theory of Games, vol. 3, pp. 97–139. Princeton University Press, Princeton (1957)Google Scholar
  7. [Hut03a]
    Hutter, M.: Convergence and loss bounds for Bayesian sequence prediction. IEEE Transactions on Information Theory 49(8), 2061–2067 (2003)CrossRefMathSciNetGoogle Scholar
  8. [Hut03b]
    Hutter, M.: Optimality of universal Bayesian prediction for general loss and alphabet. Journal of Machine Learning Research 4, 971–1000 (2003)CrossRefMathSciNetGoogle Scholar
  9. [KV03]
    Kalai, A., Vempala, S.: Efficient algorithms for online decision. In: Proceedings of the 16th Annual Conference on Learning Theory (COLT 2003), Berlin. LNCS (LNAI), pp. 506–521. Springer, Heidelberg (2003)Google Scholar
  10. [LW89]
    Littlestone, N., Warmuth, M.K.: The weighted majority algorithm. In: 30th Annual Symposium on Foundations of Computer Science, Research Triangle Park, North Carolina, pp. 256–261. IEEE, Los Alamitos (1989)CrossRefGoogle Scholar
  11. [LW94]
    Littlestone, N., Warmuth, M.K.: The weighted majority algorithm. Information and Computation 108(2), 212–261 (1994)MATHCrossRefMathSciNetGoogle Scholar
  12. [McD89]
    McDiarmid, C.: On the method of bounded differences. In: Kastens, U., Zimmermann, E., Hutt, B. (eds.) GAG: A Practical Compiler Generator. LNCS, vol. 141, pp. 148–188. Springer, Heidelberg (1989)Google Scholar
  13. [Vov90]
    Vovk, V.G.: Aggregating strategies. In: Proceedings of the Third Annual Workshop on Computational Learning Theory, pp. 371–383. ACM Press, New York (1990)Google Scholar
  14. [Vov95]
    Vovk, V.G.: A game of prediction with expert advice. In: Proceedings of the 8th Annual Conference on Computational Learning Theory, pp. 51–60. ACM Press, New York (1995)CrossRefGoogle Scholar
  15. [YEYS04]
    Yaroshinsky, R., El-Yaniv, R., Seiden, S.: How to better use expert advice. Machine Learning (2004)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2004

Authors and Affiliations

  • Marcus Hutter
    • 1
  • Jan Poland
    • 1
  1. 1.IDSIAManno-LuganoSwitzerland

Personalised recommendations