Skip to main content
SpringerLink
Log in

Faster Hoeffding Racing: Bernstein Races via Jackknife Estimates

  • Conference paper
Algorithmic Learning Theory (ALT 2013)

Part of the book series: Lecture Notes in Computer Science ((LNAI,volume 8139))

Included in the following conference series:

Abstract

Hoeffding racing algorithms are used to achieve computational speedups in settings where the goal is to select a “best” option among a set of alternatives, but the amount of data is so massive that scoring all candidates using every data point is too costly. The key is to construct confidence intervals for scores of candidates that are used to eliminate options sequentially as more samples are processed. We propose a tighter version of Hoeffding racing based on empirical Bernstein inequalities, where a jackknife estimate is used in place of the unknown variance. We provide rigorous proofs of the accuracy of our confidence intervals in the case of U-statistics and entropy estimators, and demonstrate the efficacy of our racing algorithms with synthetic experiments.

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

Log in via an institution
Chapter
USD 29.95
Price excludes VAT (USA)
  • Available as PDF
  • Read on any device
  • Instant download
  • Own it forever
eBook
USD 39.99
Price excludes VAT (USA)
  • Available as PDF
  • Read on any device
  • Instant download
  • Own it forever
Softcover Book
USD 54.99
Price excludes VAT (USA)
  • Compact, lightweight edition
  • Dispatched in 3 to 5 business days
  • Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

Institutional subscriptions

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. Antos, A., Kontoyiannis, I.: Convergence properties of functional estimates for discrete distributions. Random Structures and Algorithms 19(3-4), 163–193 (2001)

    Article  MathSciNet  MATH  Google Scholar 

  2. Arcones, M.: A Bernstein-type inequality for U-statistics and U-processes. Statistics and Probability Letters 22(3), 239–247 (1995)

    Article  MathSciNet  MATH  Google Scholar 

  3. Audibert, J.-Y., Munos, R., Szepasvari, C.: Exploration-exploitation tradeoff using variance estimates in multi-armed bandits. Theoretical Computer Science 410(19), 1876–1902 (2009)

    Article  MathSciNet  MATH  Google Scholar 

  4. Audibert, J.-Y., Munos, R., Szepesvári, C.: Tuning bandit algorithms in stochastic environments. In: Hutter, M., Servedio, R.A., Takimoto, E. (eds.) ALT 2007. LNCS (LNAI), vol. 4754, pp. 150–165. Springer, Heidelberg (2007)

    Chapter  Google Scholar 

  5. Bernstein, S.N.: The Theory of Probabilities. Gastehizdat Publishing House, Moscow (1946)

    Google Scholar 

  6. Berry, D.A., Fristedt, B.: Bandit Problems. Chapman and Hall (1985)

    Google Scholar 

  7. Boucheron, S., Lugosi, G., Massart, P.: Concentration inequalities using the entropy method. Annals of Probability 31(3), 1583–1614 (2003)

    Article  MathSciNet  MATH  Google Scholar 

  8. DasGupta, A.: Asymptotic Theory of Statistics and Probability. Springer (2008)

    Google Scholar 

  9. Domingos, P., Hulten, G.: Mining high-speed data streams. In: KDD, pp. 71–80 (2000)

    Google Scholar 

  10. Dubhashi, D.P., Panconesi, A.: Concentration of Measure for the Analysis of Randomized Algorithms. Cambridge University Press (2009)

    Google Scholar 

  11. Efron, B., Stein, C.: The jackknife estimator of variance. Annals of Statistics 9, 586–596 (1981)

    Article  MathSciNet  MATH  Google Scholar 

  12. Even-Dar, E., Mannor, S., Mansour, Y.: Action elimination and stopping conditions for the multi-armed bandit and reinforcement learning problems. JMLR 7, 1079–1105 (2006)

    MathSciNet  MATH  Google Scholar 

  13. Hoeffding, W.: Probability inequalities for sums of bounded random variables. JASA 58(301), 13–30 (1963)

    Article  MathSciNet  MATH  Google Scholar 

  14. Ikonomovska, E., Gama, J., Zenko., B., Dzeroski, S.: Speeding up Hoeffding-based regression trees with options. In: Proceedings of ICML, pp. 537–544 (2011)

    Google Scholar 

  15. Jin, R., Agrawal, G.: Efficient decision tree construction on streaming data. In: Proceedings of the 9th ACM SIGKDD, pp. 571–576. ACM, New York (2003)

    Google Scholar 

  16. Lee, A.J.: U-statistics: Theory and Practice. CRC Press (1990)

    Google Scholar 

  17. Maron, O., Moore, A.W.: Hoeffding races: Accelerating model selection search for classification and function approximation. Advances in NIPS, 59–66 (1993)

    Google Scholar 

  18. McDiarmid, C.: On the method of bounded differences. Surveys in Combinatorics 141, 148–188 (1989)

    MathSciNet  Google Scholar 

  19. Mnih, V., Szepesvári, C., Audibert, J.-Y.: Empirical Bernstein stopping. In: Proceedings of ICML, vol. 307, pp. 672–679. ACM (2008)

    Google Scholar 

  20. Paninski, L.: Estimation of entropy and mutual information. Neural Computation 15(6), 1191–1253 (2003)

    Article  MATH  Google Scholar 

  21. Peel, T., Anthoine, S., Ralaivola, L.: Empirical Bernstein inequalities for U-statistics. Advances in NIPS 23, 1903–1911 (2010)

    Google Scholar 

  22. Pfahringer, B., Holmes, G., Kirkby, R.: New options for Hoeffding trees. In: Australian Conference on Artificial Intelligence, pp. 90–99 (2007)

    Google Scholar 

  23. Serfling, R.: Approximation theorems of mathematical statistics, Series in Probability and Mathematical Statistics, New York, NY (1980)

    Google Scholar 

  24. Steele, J.M.: An Efron-Stein inequality for nonsymmetric statistics. Annals of Statistics 14(2), 753–758 (1986)

    Article  MathSciNet  MATH  Google Scholar 

Download references

Author information

Authors and Affiliations

  1. Department of Statistics, University of California, Berkeley, CA, 97420, USA

    Po-Ling Loh

  2. Microsoft Research Cambridge, 21 Station Road, Cambridge, UK

    Sebastian Nowozin

Authors
  1. Po-Ling Loh
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Sebastian Nowozin
    View author publications

    You can also search for this author in PubMed Google Scholar

Editor information

Editors and Affiliations

  1. National University of Singapore, Republic of Singapore

    Sanjay Jain  & Frank Stephan  & 

  2. Inria Lille - Nord Europe, Villeneuve d’Ascq, France

    Rémi Munos

  3. Hokkaido University, Sapporo, Japan

    Thomas Zeugmann

Rights and permissions

Reprints and permissions

Copyright information

© 2013 Springer-Verlag Berlin Heidelberg

About this paper

Cite this paper

Loh, PL., Nowozin, S. (2013). Faster Hoeffding Racing: Bernstein Races via Jackknife Estimates. In: Jain, S., Munos, R., Stephan, F., Zeugmann, T. (eds) Algorithmic Learning Theory. ALT 2013. Lecture Notes in Computer Science(), vol 8139. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-40935-6_15

Download citation

  • DOI: https://doi.org/10.1007/978-3-642-40935-6_15

  • Publisher Name: Springer, Berlin, Heidelberg

  • Print ISBN: 978-3-642-40934-9

  • Online ISBN: 978-3-642-40935-6

  • eBook Packages: Computer ScienceComputer Science (R0)

Publish with us

Policies and ethics

Search

Navigation