Approximation by ridge functions and neural networks with a bounded number of neurons

被引：12

作者：

Ismailov, Vugar E. ^{[1
]}

机构：

[1] Natl Acad Sci Azerbaijan, Inst Math & Mech, AZ-1141 Baku, Azerbaijan

来源：

APPLICABLE ANALYSIS | 2015年 / 94卷 / 11期

关键词：

neural network; MLP model; activation function; weight; ridge function; extremal element; path; orbit; 41A30; 41A63; 68T05; 92B20; MULTILAYER FEEDFORWARD NETWORKS; RECONSTRUCTION; WEIGHTS; CAPABILITIES; OPTIMIZATION; DIRECTIONS;

D O I：

10.1080/00036811.2014.979809

中图分类号：

O29 [应用数学];

学科分类号：

070104 ;

摘要：

In this paper, we consider the problem of approximation of continuous multivariate functions by neural networks with a bounded number of neurons in hidden layers. We prove the existence of single-hidden-layer networks with bounded number of neurons, which have approximation capabilities not worse than those of networks with arbitrarily many neurons. Our analysis is based on the properties of ridge functions.

引用

页码：2245 / 2260

页数：16

共 50 条

[1] On the approximation by neural networks with bounded number of neurons in hidden layers
Ismailov, Vugar E.
[J]. JOURNAL OF MATHEMATICAL ANALYSIS AND APPLICATIONS, 2014, 417 (02) : 963 - 969
[2] Approximation by ridge functions and neural networks
Petrushev, PP
[J]. SIAM JOURNAL ON MATHEMATICAL ANALYSIS, 1998, 30 (01) : 155 - 189
[3] APPROXIMATION OF FUNCTIONS BY PERCEPTRON NETWORKS WITH BOUNDED NUMBER OF HIDDEN UNITS
KURKOVA, V
[J]. NEURAL NETWORKS, 1995, 8 (05) : 745 - 750
[4] Approximation by neural networks with a bounded number of nodes at each level
Gripenberg, G
[J]. JOURNAL OF APPROXIMATION THEORY, 2003, 122 (02) : 260 - 266
[5] Function approximation by random neural networks with a bounded number of layers
Gelenbe, E
Mao, ZH
Li, YD
[J]. ESM 2003: 17TH EUROPEAN SIMULATION MULTICONFERENCE: FOUNDATIONS FOR SUCCESSFUL MODELLING & SIMULATION, 2003, : 5 - 24
[6] APPROXIMATION BY RIDGE FUNCTIONS AND NEURAL NETWORKS WITH ONE HIDDEN LAYER
CHUI, CK
LI, X
[J]. JOURNAL OF APPROXIMATION THEORY, 1992, 70 (02) : 131 - 141
[7] Upper bounds on the number of hidden neurons in feedforward networks with arbitrary bounded nonlinear activation functions
Huang, GB
Babri, HA
[J]. IEEE TRANSACTIONS ON NEURAL NETWORKS, 1998, 9 (01): : 224 - 229
[8] Approximation bounds of three-layered neural networks - A theorem on an integral transform with ridge functions
Murata, N
[J]. ELECTRONICS AND COMMUNICATIONS IN JAPAN PART III-FUNDAMENTAL ELECTRONIC SCIENCE, 1996, 79 (03): : 23 - 33
[9] Approximation by Neural Networks with Sigmoidal Functions
Dan Sheng YU
[J]. Acta Mathematica Sinica,English Series, 2013, (10) : 2013 - 2026
[10] On the approximation of functions by tanh neural networks
De Ryck, Tim
Lanthaler, Samuel
Mishra, Siddhartha
[J]. NEURAL NETWORKS, 2021, 143 : 732 - 750

← 1 2 3 4 5 →