Replacing points by compacta in neural network approximation. (English) Zbl 1075.54013

Kainen, Paul C.

J. Franklin Inst. 341, No. 4, 391-399 (2004).

Dugundji remarks that in Hausdorff spaces, “the compact subsets behave as points do and have the same separation properties”. That is, compacta can replace points in some mathematical assertions. In the paper under review this paradigm of Dugundji is applied to some problems which arise in nonlinear approximation. The author prove statements about realization of distances by pairs of points in metric spaces, about approximative compactness in Cartesian products of metric spaces, about various type of compactness of sums of two sets in linear metric spaces. A typical result is Theorem 5.1: Let \(S\) and \(P\) be nonempty subsets of metric spaces \(X\) and \(Y\), respectively. Suppose that \(P\) is compact. If \(S\) is boundedly compact or approximatively compact, then so is \({S\times P}\).
Applications are given to approximation of \({L_p}\)-functions on the \(d\)-dimensional cube, \({1\leq p<\infty}\), by linear combinations of half-space characteristic functions; i.e., by Heaviside perceptron networks.

Reviewer: Victor Milman (Minsk)

MSC:

54E35	Metric spaces, metrizability
54B10	Product spaces in general topology
41A65	Abstract approximation theory (approximation in normed linear spaces and other abstract spaces)
68T05	Learning and adaptive systems in artificial intelligence

Keywords:

approximative compactness; proximinal set; Cartesian product; F-space; nonlinear approximation; Heaviside neural network

References:

[1]	Dugundji, J., Topology (1966), Allyn and Bacon: Allyn and Bacon Boston
[2]	Kainen, P. C.; Kůrková, V.; Sanguineti, M., Minimization of error functionals over variable basis functions, SIAM J. Optim., 14, 732-742 (2003) · Zbl 1061.49020
[4]	Dunford, N.; Schwartz, J. T., Linear Operators, Part I (1967), Interscience Publication: Interscience Publication New York
[5]	Gaal, S. A., Point Set Topology (1964), Academic Press: Academic Press New York · Zbl 0124.15401
[6]	Day, M. M., Normed Linear Spaces (1962), Academic Press: Academic Press New York · Zbl 0100.10802
[7]	Dontchev, A. L.; Zolezzi, T., Well-Posed Optimization Problems, Lecture Notes in Mathematics, Vol. 1543 (1993), Springer: Springer Berlin, Heidelberg · Zbl 0797.49001
[8]	Efimov, N. V.; Stečkin, S. B., Approximative compactness and Čebyshev sets, Dokl. Akad. Nauk SSSR, 140, 522-524 (1961), (in Russian) (English transl. in Soviet Math. Dokl. 2, 1961)
[9]	Vlasov, L. P., The concept of approximative compactness and its variants, Mat. Zametki, 16, 337-348 (1974), (English transl. Math Notes 16 (1974) 786-792)
[10]	Deutsch, F., Existence of best approximations, J. Approximation Theory, 28, 132-154 (1980) · Zbl 0464.41016
[11]	Akhieser, N. I., Theory of Approximation (1992), Dover: Dover New York, (orig. 1956) · Zbl 1159.01320
[12]	Singer, I., Best Approximation in Normed Linear Spaces by Elements of Linear Subspaces (1970), Springer: Springer Berlin · Zbl 0197.38601
[13]	Nicolescu, M., Sur la meillieure approximation d’une fonction donnée par les fonctions d’une famille donnée, Bul. Fac. Sti. Cernauti, 12, 120-128 (1938)
[14]	Holmes, R. B., A Course on Optimization and Best Approximation, Lecture Notes in Mathematics, Vol. 257 (1972), Springer: Springer Berlin · Zbl 0234.46016
[15]	Kainen, P. C.; Kůrková, V.; Vogt, A., Best approximation by Heaviside perceptron networks, Neural Networks, 13, 695-697 (2000)
[16]	Kainen, P. C.; Kůrková, V.; Vogt, A., Best approximation by linear combinations of characteristic functions of half-spaces, J. Approximation Theory, 122, 151-159 (2003) · Zbl 1038.41011
[17]	Poggio, T.; Smale, S., The mathematics of learningdealing with data, Notices Amer. Math. Soc., 50, 536-544 (2003)
[19]	de Figuierdo, R. J.P.; Chen, G., Optimal disturbance rejection for nonlinear control systems, IEEE Trans. Automat. Control, 34, 1242-1248 (1989)