论文信息 - A note on a scale-sensitive dimension of linear bounded functionals in Banach spaces

A note on a scale-sensitive dimension of linear bounded functionals in Banach spaces

We show that "B is of type p > 1" is a necessary and sufficient condition for a learnability of a class of linear bounded functionals with norm < 1 restricted to the unit ball in Banach space B. On the way, we give very short probabilistic proof for Vapnik's result (Hilbert space and improved) and improve our result with Pascal Koiran for convex halls of indicator functions. The approach we use in this paper allows to connect various results about learnability and approximation.

Leonid Gurvits

[1] Vladimir Vapnik,et al. The Nature of Statistical Learning , 1995 .

[2] Noga Alon,et al. Scale-sensitive dimensions, uniform convergence, and learnability , 1993, Proceedings of 1993 IEEE 34th Annual Foundations of Computer Science.

[3] Leonid Gurvits,et al. Approximation and Learning of Convex Superpositions , 1997, J. Comput. Syst. Sci..

[4] Vladimir N. Vapnik,et al. The Nature of Statistical Learning Theory , 2000, Statistics for Engineering and Information Science.

[5] C. Darken,et al. Constructive Approximation Rates of Convex Approximation in Non-hilbert Spaces , 2022 .

[6] David Haussler,et al. What Size Net Gives Valid Generalization? , 1989, Neural Computation.

[7] Eduardo D. Sontag,et al. Rate of approximation results motivated by robust neural network learning , 1993, COLT '93.

[8] Robert E. Schapire,et al. Efficient distribution-free learning of probabilistic concepts , 1990, Proceedings [1990] 31st Annual Symposium on Foundations of Computer Science.

[9] L. Jones. A Simple Lemma on Greedy Approximation in Hilbert Space and Convergence Rates for Projection Pursuit Regression and Neural Network Training , 1992 .

[10] A. Barron. Approximation and Estimation Bounds for Artificial Neural Networks , 1991, COLT '91.

[11] J. Hoffmann-jorgensen. Probability in Banach Space , 1977 .