Eignets for function approximation on manifolds

DOI10.1016/J.ACHA.2009.08.006zbMATH Open1201.41003arXiv0909.5000OpenAlexW2963072276MaRDI QIDQ984660FDOQ984660

H. N. Mhaskar

Publication date: 20 July 2010

Published in: Applied and Computational Harmonic Analysis (Search for Journal in Brave)

Abstract: Let $XX$ be a compact, smooth, connected, Riemannian manifold without boundary, $G:XXimesXXoRR$ be a kernel. Analogous to a radial basis function network, an eignet is an expression of the form $su{m}_{j=1}^M{a}_{j}G(circ,y_j)$, where $a_{j}inRR$, $y_{j}inXX$, $1lejleM$. We describe a deterministic, universal algorithm for constructing an eignet for approximating functions in $L^p(mu;XX)$ for a general class of measures $mu$ and kernels $G$. Our algorithm yields linear operators. Using the minimal separation amongst the centers $y_j$ as the cost of approximation, we give modulus of smoothness estimates for the degree of approximation by our eignets, and show by means of a converse theorem that these are the best possible for every emph{individual function}. We also give estimates on the coefficients $a_j$ in terms of the norm of the eignet. Finally, we demonstrate that if any sequence of eignets satisfies the optimal estimates for the degree of approximation of a smooth function, measured in terms of the minimal separation, then the derivatives of the eignets also approximate the corresponding derivatives of the target function in an optimal manner.

Full work available at URL: https://arxiv.org/abs/0909.5000

Recommendations

• Kernel Approximation on Manifolds I: Bounding the Lebesgue Constant
• scientific article; zbMATH DE number 2065335
• A direct approach for function approximation on data defined manifolds
• SPLINE APPROXIMATIONS ON MANIFOLDS
• Approximation of eigenfunctions in kernel-based spaces

zbMATH Keywords

simultaneous approximationstability estimatesRBF networksdata dependent manifoldsdirect and converse theorems of approximationkernel based approximation

Mathematics Subject Classification ID

Computational difficulty of problems (lower bounds, completeness, difficulty of approximation, etc.) (68Q17) Approximation by other special function classes (41A30)

Cites Work

• Diffusion wavelets
• Towards a theoretical foundation for Laplacian-based manifold methods
• Title not available (Why is that?)
• Derivatives of the spectral function and Sobolev norms of eigenfunctions on a closed Riemannian manifold
• The spectral function of an elliptic operator
• Universal local parametrizations via heat kernels and eigenfunctions of the Laplacian
• Title not available (Why is that?)
• From graph to manifold Laplacian: the convergence rate
• Title not available (Why is that?)
• Title not available (Why is that?)
• Some Properties of the Eigenfunctions of The Laplace-Operator on Riemannian Manifolds
• Semi-supervised learning on Riemannian manifolds
• Data analysis and representation on a general domain using eigenfunctions of Laplacian
• \(L^ p\)-theory of elliptic differential operators on manifolds of bounded geometry
• Efficient reconstruction of functions on the sphere from scattered data
• Title not available (Why is that?)
• Title not available (Why is that?)
• Image manifolds which are isometric to Euclidean space
• A quadrature formula for diffusion polynomials corresponding to a generalized heat kernel
• Localized Linear Polynomial Operators and Quadrature Formulas on the Sphere
• Polynomial approximation on the sphere using scattered data
• Diffusion polynomial frames on metric measure spaces
• When is approximation by Gaussian networks necessarily a linear process?
• A walk through energy, discrepancy, numerical integration and group invariant measures on measurable subsets of Euclidean space
• Local Paley–Wiener theorems for functions analytic on unit spheres

Cited In (33)

• Quasi Monte Carlo Integration and Kernel-Based Function Approximation on Grassmannians
• Approximation by linear combinations of translates of a single function
• Smooth function extension based on high dimensional unstructured data
• Function approximation with zonal function networks with activation functions analogous to the rectified linear unit functions
• Curve based approximation of measures on manifolds by discrepancy minimization
• Representation of functions on big data associated with directed graphs
• Super-resolution meets machine learning: approximation of measures
• A generalized diffusion frame for parsimonious representation of functions on data defined manifolds
• An analysis of training and generalization errors in shallow and deep networks
• Data Based Construction of Kernels for Classification
• Estimation of a regression function on a manifold by fully connected deep neural networks
• Regularized kernel-based reconstruction in generalized Besov spaces
• Construction and Monte Carlo estimation of wavelet frames generated by a reproducing kernel
• Applications of classical approximation theory to periodic basis function networks and computational harmonic analysis
• Function approximation by deep networks
• Tight framelets and fast framelet filter bank transforms on manifolds
• Dual framelets transform on manifolds and graphs
• Deep neural networks can stably solve high-dimensional, noisy, non-linear inverse problems
• Band-limited localized Parseval frames and Besov spaces on compact homogeneous manifolds
• Approximate Quadrature Measures on Data-Defined Spaces
• Minimum Sobolev norm interpolation of scattered derivative data
• A unified framework for harmonic analysis of functions on directed graphs and changing data
• A direct approach for function approximation on data defined manifolds
• Minimum Sobolev norm interpolation with trigonometric polynomials on the torus
• Weighted least \(\ell_p\) approximation on compact Riemannian manifolds
• \(\varepsilon\)-coverings of Hölder-Zygmund type spaces on data-defined manifolds
• Local Approximation Using Hermite Functions
• Function recovery on manifolds using scattered data
• Marcinkiewicz-Zygmund measures on manifolds
• Metric entropy, \(n\)-widths, and sampling of functions on manifolds
• Title not available (Why is that?)
• Representation of functions on big data: graphs and trees
• Applied harmonic analysis and data processing. Abstracts from the workshop held March 25--31, 2018

This page was built for publication: Eignets for function approximation on manifolds

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q984660)