Constructive approximation by Gaussian neural networks
From MaRDI portal
Let \(C_{c}(\mathbb{R}^{s})\) be the set of all continuous functions that have compact support. Then it is possible to construct Gaussian neural networks with integer weights that approximate arbitrarily well for functions in \(C_{c}(\mathbb{R}^{s})\). Further, some numerical results support the obtained theoretical results.
Recommendations
- Constructive approximation by neural networks with positive integer weights
- The rate of approximation of Gaussian radial basis neural networks in continuous function space
- The construction and approximation of neural networks operators with Gaussian activation function
- When is approximation by Gaussian networks necessarily a linear process?
- Neural networks for optimal approximation of continuous functions in \(\mathbb R^d\)
Cited in
(7)- Gaussian kernel approximation algorithm for feedforward neural network design
- When is approximation by Gaussian networks necessarily a linear process?
- Constructive deep ReLU neural network approximation
- Constructive estimation of approximating of functions with Gaussian RBF neural network
- The construction and approximation of neural networks operators with Gaussian activation function
- Constructive approximation by neural networks with positive integer weights
- Constructive approximate interpolation by neural networks
This page was built for publication: Constructive approximation by Gaussian neural networks
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2871205)
