On the approximation by single hidden layer feedforward neural networks with fixed weights

DOI10.1016/j.neunet.2017.12.007zbMath1437.68062arXiv1708.06219OpenAlexW2747971139WikidataQ47568517 ScholiaQ47568517MaRDI QIDQ2179313

Vugar E. Ismailov, Namig J. Guliyev

Publication date: 12 May 2020

Published in: Neural Networks (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/1708.06219

zbMATH Keywords

approximationweightfeedforward neural networkactivation functionsigmoidal functionhidden layer

Mathematics Subject Classification ID

Approximation by other special function classes (41A30) Biologically inspired models of computation (DNA computing, membrane computing, etc.) (68Q07)

Related Items (26)

SelectNet: self-paced learning for high-dimensional partial differential equations ⋮ Constructive Approximation of Continuous Interval-Valued Functions ⋮ Full error analysis for the training of deep neural networks ⋮ An interval uncertainty analysis method for structural response bounds using feedforward neural network differentiation ⋮ Asymptotic expansions and Voronovskaja type theorems for the multivariate neural network operators ⋮ Robust min-max optimal control design for systems with uncertain models: a neural dynamic programming approach ⋮ On the approximation of functions by tanh neural networks ⋮ The generalized extreme learning machines: tuning hyperparameters and limiting approach for the Moore-Penrose generalized inverse ⋮ Approximation capabilities of neural networks on unbounded domains ⋮ Some aspects of approximation and interpolation of functions artificial neural networks ⋮ Some elliptic second order problems and neural network solutions: existence and error estimates ⋮ Neural network interpolation operators optimized by Lagrange polynomial ⋮ On decision regions of narrow deep neural networks ⋮ Overall error analysis for the training of deep neural networks via stochastic gradient descent with random initialisation ⋮ Neural network interpolation operators of multivariate functions ⋮ Lower bounds for artificial neural network approximations: a proof that shallow neural networks fail to overcome the curse of dimensionality ⋮ Fractional type multivariate neural network operators ⋮ Approximation error of single hidden layer neural networks with fixed weights ⋮ Computing the Approximation Error for Neural Networks with Weights Varying on Fixed Directions ⋮ Negative results for approximation using single layer and multilayer feedforward neural networks ⋮ Extreme learning machine collocation for the numerical solution of elliptic PDEs with sharp gradients ⋮ Asymptotic expansion for neural network operators of the Kantorovich type and high order of approximation ⋮ Two-hidden-layer feed-forward networks are universal approximators: a constructive approach ⋮ On the representation by bivariate ridge functions ⋮ Rates of approximation by neural network interpolation operators ⋮ The construction and approximation of ReLU neural network operators

Cites Work

• Unnamed Item
• Unnamed Item
• Approximation by neural networks with scattered data
• Approximation by max-product neural network operators of Kantorovich type
• Approximation by neural networks with weights varying on a finite set of directions
• The approximation operators with sigmoidal functions
• Approximation by Ridge functions and neural networks with one hidden layer
• Approximation by superposition of sigmoidal and radial basis functions
• Lower bounds for approximation by MLP neural networks
• Fundamentality of ridge functions
• Neural network operators: constructive interpolation of multivariate functions
• An approximation by neural networks with a fixed weight
• A neural network model with bounded-weights for pattern classification
• Limitations of the approximation capabilities of neural networks with one hidden layer
• On the approximation of the step function by some sigmoid functions
• Limitations of shallow nets approximation
• Pointwise and uniform approximation by multivariate neural network operators of the max-product type
• Max-product neural network and quasi-interpolation operators activated by sigmoidal functions
• Recounting the Rationals
• Deep vs. shallow networks: An approximation theory perspective
• Convergence for a family of neural network operators in Orlicz spaces
• Approximation by ridge functions and neural networks with a bounded number of neurons
• Ridge Functions
• Measure Theoretic Results for Approximation by Neural Networks with Limited Weights
• Approximation by sums of ridge functions with fixed directions
• A Single Hidden Layer Feedforward Network with Only One Neuron in the Hidden Layer Can Approximate Any Univariate Function
• Constructive Approximation by Superposition of Sigmoidal Functions
• Approximation by superpositions of a sigmoidal function

This page was built for publication: On the approximation by single hidden layer feedforward neural networks with fixed weights