Understanding Gaussian Process Regression Using the Equivalent Kernel

  • Peter Sollich
  • Christopher K. I. Williams
Part of the Lecture Notes in Computer Science book series (LNCS, volume 3635)

Abstract

The equivalent kernel [1] is a way of understanding how Gaussian process regression works for large sample sizes based on a continuum limit. In this paper we show how to approximate the equivalent kernel of the widely-used squared exponential (or Gaussian) kernel and related kernels. This is easiest for uniform input densities, but we also discuss the generalization to the non-uniform case. We show further that the equivalent kernel can be used to understand the learning curves for Gaussian processes, and investigate how kernel smoothing using the equivalent kernel compares to full Gaussian process regression.

Keywords

Power Spectrum Mean Square Error Covariance Function Gaussian Process Target Function 
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.
    Silverman, B.W.: Annals of Statistics,  12, 898–916 (1984)Google Scholar
  2. 2.
    Williams, C.K.I.: In: Jordan, M.I. (ed.) Learning in Graphical Models, pp. 599–621. Kluwer Academic, Dordrecht (1998)Google Scholar
  3. 3.
    Hastie, T.J., Tibshirani, R.J.: Generalized Additive Models. Chapman and Hall, Boca Raton (1990)MATHGoogle Scholar
  4. 4.
    Sollich, P., Williams, C.K.I.: NIPS 17 (2005) (to appear)Google Scholar
  5. 5.
    Girosi, F., Jones, M., Poggio, T.: Neural Computation 7(2), 219–269 (1995)Google Scholar
  6. 6.
    Papoulis, A.: Probability, Random Variables, and Stochastic Processes, 3rd edn. McGraw-Hill, New York (1991)Google Scholar
  7. 7.
    Thomas-Agnan, C.: Numerical Algorithms.  13, 21–32 (1996)Google Scholar
  8. 8.
    Poggio, T., Voorhees, H., Yuille, A.: Tech. Report AI Memo 833, MIT AI Laboratory (1985)Google Scholar
  9. 9.
    Schölkopf, B., Smola, A.: Learning with Kernels. MIT Press, Cambridge (2002)Google Scholar
  10. 10.
    Stein, M.L.: Interpolation of Spatial Data. Springer, New York (1999)MATHGoogle Scholar
  11. 11.
    Zhu, H., Williams, C.K.I., Rohwer, R.J., Morciniec, M.: Gaussian Regeression and Optimal Finite Dimensional LinearModels. In: Bishop, C.M. (ed.) Neural Networks and Machine Learning. Springer, Heidelberg (1998)Google Scholar
  12. 12.
    Abramowitz, M., Stegun, I.A.: Handbook of Mathematical Functions, Dover, New York (1965)Google Scholar
  13. 13.
    Sollich, P., Halees, A.: Neural Computation.  14, 1393–1428 (2002)Google Scholar
  14. 14.
    Opper, M., Vivarelli, F.: NIPS 11, pp. 302–308 (1999)Google Scholar
  15. 15.
    Sollich, P.: NIPS 14, pp. 519–526 (2002)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2005

Authors and Affiliations

  • Peter Sollich
    • 1
  • Christopher K. I. Williams
    • 2
  1. 1.Dept of MathematicsKing’s College LondonLondonU.K.
  2. 2.School of InformaticsUniversity of EdinburghEdinburghU.K.

Personalised recommendations