Nonlinear approximation using Gaussian kernels

DOI10.1016/J.JFA.2010.02.001zbMATH Open1203.41015arXiv0911.2803OpenAlexW2132221321MaRDI QIDQ982495FDOQ982495

Authors: Thomas Hangelbroek, A. Ron 

Publication date: 7 July 2010

Published in: Journal of Functional Analysis (Search for Journal in Brave)

Abstract: It is well-known that non-linear approximation has an advantage over linear schemes in the sense that it provides comparable approximation rates to those of the linear schemes, but to a larger class of approximands. This was established for spline approximations and for wavelet approximations, and more recently by DeVore and Ron for homogeneous radial basis function (surface spline) approximations. However, no such results are known for the Gaussian function, the preferred kernel in machine learning and several engineering problems. We introduce and analyze in this paper a new algorithm for approximating functions using translates of Gaussian functions with varying tension parameters. At heart it employs the strategy for nonlinear approximation of DeVore and Ron, but it selects kernels by a method that is not straightforward. The crux of the difficulty lies in the necessity to vary the tension parameter in the Gaussian function spatially according to local information about the approximand: error analysis of Gaussian approximation schemes with varying tension are, by and large, an elusive target for approximators. We show that our algorithm is suitably optimal in the sense that it provides approximation rates similar to other established nonlinear methodologies like spline and wavelet approximations. As expected and desired, the approximation rates can be as high as needed and are essentially saturated only by the smoothness of the approximand.

Full work available at URL: https://arxiv.org/abs/0911.2803

Recommendations

• Regular families of kernels for nonlinear approximation
• Nonlinear approximation with general wave packets
• scientific article; zbMATH DE number 3855674
• New bases for Triebel-Lizorkin and Besov spaces
• On approximate approximations using Gaussian kernels

zbMATH Keywords

Triebel-Lizorkin spaceBesov spacenonlinear approximationshifted-dilated Gaussian

Mathematics Subject Classification ID

Multidimensional problems (41A63) Approximation by other special function classes (41A30) Approximation by arbitrary nonlinear expressions; widths and entropy (41A46)

Cites Work

• Title not available (Why is that?)
• Some Maximal Inequalities
• Title not available (Why is that?)
• Improved accuracy of multiquadric interpolation using variable shape parameters
• Approximation by radial bases and neural networks
• Fast rates for support vector machines using Gaussian kernels
• The Runge phenomenon and spatially variable shape parameters in RBF interpolation
• Fourier analysis of the approximation power of principal shift-invariant spaces
• Learnability of Gaussians with flexible variances
• Approximation using scattered shifts of a multivariate function

Cited In (20)

• An integral equation method for the numerical solution of the Burgers equation
• On linear versus nonlinear approximation in the average case setting
• Nonuniform sampling and approximation in Sobolev space from perturbation of the framelet system
• Title not available (Why is that?)
• The rate of approximation of Gaussian radial basis neural networks in continuous function space
• Cardinal interpolation with Gaussian kernels
• Highly localized RBF Lagrange functions for finite difference methods on spheres
• Gaussian sum approximation for non-linear fixed-point prediction
• \textit{hp}-VPINNs: variational physics-informed neural networks with domain decomposition
• Weak form theory-guided neural network (TgNN-wf) for deep learning of subsurface single- and two-phase flow
• Kernel methods for the approximation of some key quantities of nonlinear systems
• Regular families of kernels for nonlinear approximation
• Title not available (Why is that?)
• Title not available (Why is that?)
• Nonlinear approximation via compositions
• Deep network approximation characterized by number of neurons
• Optimal approximation by one Gaussian function to probability density functions
• Exponential tractability of \(L_2\)-approximation with function values
• Gaussianization machines for non-Gaussian function estimation models
• Green's functions: taking another look at kernel approximation, radial basis functions, and splines

This page was built for publication: Nonlinear approximation using Gaussian kernels

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q982495)