Abstract
We consider the problem of on-line prediction competitive with a benchmark class of continuous but highly irregular prediction rules. It is known that if the benchmark class is a reproducing kernel Hilbert space, there exists a prediction algorithm whose average loss over the first N examples does not exceed the average loss of any prediction rule in the class plus a “regret term” of O(N − 1/2). The elements of some natural benchmark classes, however, are so irregular that these classes are not Hilbert spaces. In this paper we develop Banach-space methods to construct a prediction algorithm with a regret term of O(N \(^{\rm -1/{\it p}}\)), where p∈(2,∞) and p–2 reflects the degree to which the benchmark class fails to be a Hilbert space.
Preview
Unable to display preview. Download preview PDF.
Similar content being viewed by others
References
Adams, R.A., Fournier, J.J.F.: Sobolev Spaces, 2nd edn. Pure and Applied Mathematics, vol. 140. Academic Press, Amsterdam (2003)
Aronszajn, N.: Theory of reproducing kernels. Transactions of the American Mathematical Society 68, 337–404 (1950)
Auer, P., Cesa-Bianchi, N., Gentile, C.: Adaptive and self-confident on-line learning algorithms. Journal of Computer and System Sciences 64, 48–75 (2002)
Cesa-Bianchi, N., Conconi, A., Gentile, C.: On the generalization ability of on-line learning algorithms. IEEE Transactions on Information Theory 50, 2050–2057 (2004)
Cesa-Bianchi, N., Long, P.M., Warmuth, M.K.: Worst-case quadratic loss bounds for on-line prediction of linear functions by gradient descent. IEEE Transactions on Neural Networks 7, 604–619 (1996)
Clarkson, J.A.: Uniformly convex spaces. Transactions of the American Mathematical Society 40, 396–414 (1936)
Figiel, T.: On the moduli of convexity and smoothness. Studia Mathematica 56, 121–155 (1976), available free of charge at: http://matwbn.icm.edu.pl
Hanner, O.: On the uniform convexity of L p and l p. Arkiv för Matematik 3, 239–244 (1956)
Karatzas, I., Shreve, S.E.: Brownian Motion and Stochastic Calculus, 2nd edn. Springer, New York (1991)
Kivinen, J., Warmuth, M.K.: Exponential Gradient versus Gradient Descent for linear predictors. Information and Computation 132, 1–63 (1997)
Kolmogorov, A.N.: Grundbegriffe der Wahrscheinlichkeitsrechnung. Springer, Berlin (1933); English translation: Foundations of the theory of probability. Chelsea, New York (1950)
Lindenstrauss, J.: On the modulus of smoothness and divergent series in Banach spaces. Michigan Mathematical Journal 10, 241–252 (1963)
Liokumovich, V.I.: The existence of B-spaces with non-convex modulus of convexity (in Russian). ИзвестиЯ вьıсших УЧебньІх заведений. Математика 12, 43–50 (1973)
Nikolsky, S.M.: On imbedding, continuation and approximation theorems for differentiable functions of several variables. Russian Mathematical Surveys 16(5), 55–104 (1961), Russian original in: Успехи математических наук 16(5), 63–114 (1961)
Tyrrell Rockafellar, R.: Convex Analysis. Princeton University Press, Princeton (1970)
Rudin, W.: Functional Analysis, 2nd edn. McGraw-Hill, Boston (1991)
Shafer, G., Vovk, V.: Probability and Finance: It’s Only a Game! Wiley, New York (2001)
Sorenson, H.W.: Least-squares estimation: from Gauss to Kalman. IEEE Spectrum 7, 63–68 (1970)
Steinwart, I.: On the influence of the kernel on the consistency of support vector machines. Journal of Machine Learning Research 2, 67–93 (2001)
Vovk, V.: Competing with wild prediction rules. Technical Report arXiv:cs.LG/0512059, arXiv.org., e-Print archive (January 2006)
Vovk, V.: On-line regression competitive with reproducing kernel Hilbert spaces. In: Cai, J.-Y., Cooper, S.B., Li, A. (eds.) TAMC 2006. LNCS, vol. 3959, pp. 452–463. Springer, Heidelberg (2006); Full version: Technical Report arXiv:cs.LG/0511058, arXiv.org e-Print archive (January 2006)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2006 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Vovk, V. (2006). Competing with Wild Prediction Rules. In: Lugosi, G., Simon, H.U. (eds) Learning Theory. COLT 2006. Lecture Notes in Computer Science(), vol 4005. Springer, Berlin, Heidelberg. https://doi.org/10.1007/11776420_41
Download citation
DOI: https://doi.org/10.1007/11776420_41
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-35294-5
Online ISBN: 978-3-540-35296-9
eBook Packages: Computer ScienceComputer Science (R0)