摘要: Results concerning the uniform approximation of nonlinear functionals on compact sets are discussed. The results interest in connection with neural network-like classifiers for continuous-time signals and approximations input-output maps dynamic systems. It is shown that function-space feedforward networks one input layer bounded linear hidden each universal approximators real continuous subsets a normed space. >
sci-hub.se 参考文章(5)
Ken-Ichi Funahashi, On the approximate realization of continuous mappings by neural networks Neural Networks. ,vol. 2, pp. 183- 192 ,(1989) , 10.1016/0893-6080(89)90003-8
Irwin W. Sandberg, Structure theorems for nonlinear systems Multidimensional Systems and Signal Processing. ,vol. 2, pp. 267- 286 ,(1991) , 10.1007/BF01952236
Eric J. Hartman, James D. Keeler, Jacek M. Kowalski, Layered neural networks with Gaussian hidden units as universal approximations Neural Computation. ,vol. 2, pp. 210- 215 ,(1990) , 10.1162/NECO.1990.2.2.210
I.W. Sandberg, Approximation theorems for discrete-time systems IEEE Transactions on Circuits and Systems. ,vol. 38, pp. 564- 566 ,(1991) , 10.1109/31.76498
J. Park, I. W. Sandberg, Universal approximation using radial-basis-function networks Neural Computation. ,vol. 3, pp. 246- 257 ,(1991) , 10.1162/NECO.1991.3.2.246