On the approximation of functions by tanh neural networks
arXiv:2104.08938 · doi:10.1016/j.neunet.2021.08.015
Abstract
We derive bounds on the error, in high-order Sobolev norms, incurred in the approximation of Sobolev-regular as well as analytic functions by neural networks with the hyperbolic tangent activation function. These bounds provide explicit estimates on the approximation error with respect to the size of the neural networks. We show that tanh neural networks with only two hidden layers suffice to approximate functions at comparable or better rates than much deeper ReLU neural networks.
References in corpus (11)
- Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation
- Hidden Physics Models: Machine Learning of Nonlinear Partial Differential Equations
- Fourier Neural Operator for Parametric Partial Differential Equations
- Physics Informed Neural Networks for Simulating Radiative Transfer
- In Search of the Real Inductive Bias: On the Role of Implicit Regularization in Deep Learning
- Iterative Surrogate Model Optimization (ISMO): An active learning algorithm for PDE constrained optimization with deep neural networks
- Approximation capability of two hidden layer feedforward neural networks with fixed weights
- Smooth function approximation by deep neural networks with general activation functions
- Estimates on the generalization error of Physics Informed Neural Networks (PINNs) for approximating PDEs
- Higher-order Quasi-Monte Carlo Training of Deep Neural Networks
- The Representation Power of Neural Networks: Breaking the Curse of Dimensionality