Extending error bounds for radial basis function interpolation to measuring the error in higher order Sobolev norms

Hangelbroek, T.; Rieger, C.

doi:10.1090/mcom/3960

Extending error bounds for radial basis function interpolation to measuring the error in higher order Sobolev norms
HTML articles powered by AMS MathViewer

by T. Hangelbroek and C. Rieger;

Math. Comp. 94 (2025), 381-407

DOI: https://doi.org/10.1090/mcom/3960

Published electronically: March 20, 2024

HTML | PDF | Request permission

Abstract:

Radial basis functions (RBFs) are prominent examples for reproducing kernels with associated reproducing kernel Hilbert spaces (RKHSs). The convergence theory for the kernel-based interpolation in that space is well understood and optimal rates for the whole RKHS are often known. Schaback added the doubling trick [Math. Comp. 68 (1999), pp. 201–216], which shows that functions having double the smoothness required by the RKHS (along with specific, albeit complicated boundary behavior) can be approximated with higher convergence rates than the optimal rates for the whole space. Other advances allowed interpolation of target functions which are less smooth, and different norms which measure interpolation error. The current state of the art of error analysis for RBF interpolation treats target functions having smoothness up to twice that of the native space, but error measured in norms which are weaker than that required for membership in the RKHS.

Motivated by the fact that the kernels and the approximants they generate are smoother than required by the native space, this article extends the doubling trick to error which measures higher smoothness. This extension holds for a family of kernels satisfying easily checked hypotheses which we describe in this article, and includes many prominent RBFs. In the course of the proof, new convergence rates are obtained for the abstract operator considered by Devore and Ron in [Trans. Amer. Math. Soc. 362 (2010), pp. 6205–6229], and new Bernstein estimates are obtained relating high order smoothness norms to the native space norm.

References

Helmut Abels, Pseudodifferential and singular integral operators, De Gruyter Graduate Lectures, De Gruyter, Berlin, 2012. An introduction with applications. MR 2884718
Milton Abramowitz and Irene A. Stegun, Handbook of mathematical functions with formulas, graphs, and mathematical tables, National Bureau of Standards Applied Mathematics Series, No. 55, U. S. Government Printing Office, Washington, DC, 1964. For sale by the Superintendent of Documents. MR 167642
Jean-Pierre Aubin, Behavior of the error of the approximate solutions of boundary value problems for linear elliptic operators by Gelerkin’s and finite difference methods, Ann. Scuola Norm. Sup. Pisa Cl. Sci. (3) 21 (1967), 599–637. MR 233068
M. D. Buhmann, Radial functions on compact support, Proc. Edinburgh Math. Soc. (2) 41 (1998), no. 1, 33–46. MR 1604361, DOI 10.1017/S0013091500019416
C. Canuto and A. Quarteroni, Approximation results for orthogonal polynomials in Sobolev spaces, Math. Comp. 38 (1982), no. 157, 67–86. MR 637287, DOI 10.1090/S0025-5718-1982-0637287-3
Maria Charina, Costanza Conti, and Nira Dyn, Multivariate compactly supported $C^\infty$ functions by subdivision, Appl. Comput. Harmon. Anal. 70 (2024), Paper No. 101630, 13. MR 4695556, DOI 10.1016/j.acha.2024.101630
Yifan Chen, Bamdad Hosseini, Houman Owhadi, and Andrew M. Stuart, Solving and learning nonlinear PDEs with Gaussian processes, J. Comput. Phys. 447 (2021), Paper No. 110668, 29. MR 4311012, DOI 10.1016/j.jcp.2021.110668
Andrew Chernih and Simon Hubbert, Closed form representations and properties of the generalised Wendland functions, J. Approx. Theory 177 (2014), 17–33. MR 3128401, DOI 10.1016/j.jat.2013.09.005
W. M. Czarnecki, S. Osindero, M. Jaderberg, G. Swirszcz, and R. Pascanu, Sobolev training for neural networks, Adv. Neural Inf. Process Syst. 30 (2017).
T. De Ryck, S. Lanthaler, and S. Mishra, On the approximation of functions by tanh neural networks, Neural Netw. 143 (2021), 732–750.
Ronald Devore and Amos Ron, Approximation using scattered shifts of a multivariate function, Trans. Amer. Math. Soc. 362 (2010), no. 12, 6205–6229. MR 2678971, DOI 10.1090/S0002-9947-2010-05070-6
Stephen Hilbert, A mollifier useful for approximations in Sobolev spaces and some applications to approximating solutions of differential equations, Math. Comp. 27 (1973), 81–89. MR 331715, DOI 10.1090/S0025-5718-1973-0331715-3
Eleonora Di Nezza, Giampiero Palatucci, and Enrico Valdinoci, Hitchhiker’s guide to the fractional Sobolev spaces, Bull. Sci. Math. 136 (2012), no. 5, 521–573. MR 2944369, DOI 10.1016/j.bulsci.2011.12.004
F. W. J. Olver, A. B. Olde Daalhuis, D. W. Lozier, B. I. Schneider, R. F. Boisvert, C. W. Clark, B. R. Miller, and B. V. Saunders (eds.), NIST digital library of mathematical functions, http://dlmf.nist.gov/, Release 1.0.18 of 2018-03-27.
Jean Duchon, Sur l’erreur d’interpolation des fonctions de plusieurs variables par les $D^{m}$-splines, RAIRO Anal. Numér. 12 (1978), no. 4, 325–334, vi (French, with English summary). MR 519016, DOI 10.1051/m2an/1978120403251
G. E. Fasshauer, RBF collocation methods as pseudospectral methods, Boundary elements XXVII, WIT Trans. Model. Simul., vol. 39, WIT Press, Southampton, 2005, pp. 47–56. MR 2234627
I. M. Gel’fand and N. Ya. Vilenkin, Generalized functions. Vol. 4: Applications of harmonic analysis, Academic Press, New York-London, 1964. Translated by Amiel Feinstein. MR 173945
Ingo Gühring, Gitta Kutyniok, and Philipp Petersen, Error bounds for approximations with deep ReLU neural networks in $W^{s,p}$ norms, Anal. Appl. (Singap.) 18 (2020), no. 5, 803–859. MR 4131039, DOI 10.1142/S0219530519410021
Tim Gutzmer and Jens Markus Melenk, Approximation orders for natural splines in arbitrary dimensions, Math. Comp. 70 (2001), no. 234, 699–703. MR 1813144, DOI 10.1090/S0025-5718-00-01299-0
T. Hangelbroek, F. J. Narcowich, C. Rieger, and J. D. Ward, An inverse theorem for compact Lipschitz regions in $\Bbb {R}^d$ using localized kernel bases, Math. Comp. 87 (2018), no. 312, 1949–1989. MR 3787398, DOI 10.1090/mcom/3256
T. Hangelbroek, F. J. Narcowich, and J. D. Ward, Polyharmonic and related kernels on manifolds: interpolation and approximation, Found. Comput. Math. 12 (2012), no. 5, 625–670. MR 2970852, DOI 10.1007/s10208-011-9113-5
Thomas Hangelbroek and Aaron Lauve, The polyharmonic Dirichlet problem and path counting, J. Math. Pures Appl. (9) 102 (2014), no. 3, 449–481 (English, with English and French summaries). MR 3249719, DOI 10.1016/j.matpur.2013.12.001
Thomas C. Hangelbroek, On a polyharmonic Dirichlet problem and boundary effects in surface spline approximation, SIAM J. Math. Anal. 50 (2018), no. 4, 4616–4654. MR 3849117, DOI 10.1137/18M1167188
Douglas Samuel Jones, The theory of generalised functions, 2nd ed., Cambridge University Press, Cambridge-New York, 1982. MR 665103, DOI 10.1017/CBO9780511569210
Francis J. Narcowich, Joseph D. Ward, and Holger Wendland, Sobolev bounds on functions with scattered zeros, with applications to radial basis function surface fitting, Math. Comp. 74 (2005), no. 250, 743–763. MR 2114646, DOI 10.1090/S0025-5718-04-01708-9
Francis J. Narcowich and Joseph D. Ward, Norms of inverses and condition numbers for matrices associated with scattered data, J. Approx. Theory 64 (1991), no. 1, 69–94. MR 1086096, DOI 10.1016/0021-9045(91)90087-Q
Francis J. Narcowich, Joseph D. Ward, and Holger Wendland, Sobolev error estimates and a Bernstein inequality for scattered data interpolation via radial basis functions, Constr. Approx. 24 (2006), no. 2, 175–186. MR 2239119, DOI 10.1007/s00365-005-0624-7
J. Nitsche, Ein Kriterium für die Quasi-Optimalität des Ritzschen Verfahrens, Numer. Math. 11 (1968), 346–348 (German). MR 233502, DOI 10.1007/BF02166687
R. Schaback, Improved error bounds for scattered data interpolation by radial basis functions, Math. Comp. 68 (1999), no. 225, 201–216. MR 1604379, DOI 10.1090/S0025-5718-99-01009-1
Robert Schaback, Native Hilbert spaces for radial basis functions. I, New developments in approximation theory (Dortmund, 1998) Internat. Ser. Numer. Math., vol. 132, Birkhäuser, Basel, 1999, pp. 255–282. MR 1724923
Robert Schaback and Holger Wendland, Kernel techniques: from machine learning to meshless methods, Acta Numer. 15 (2006), 543–639. MR 2269747, DOI 10.1017/S0962492906270016
Holger Wendland, Piecewise polynomial, positive definite and compactly supported radial functions of minimal degree, Adv. Comput. Math. 4 (1995), no. 4, 389–396. MR 1366510, DOI 10.1007/BF02123482
Holger Wendland, Scattered data approximation, Cambridge Monographs on Applied and Computational Mathematics, vol. 17, Cambridge University Press, Cambridge, 2005. MR 2131724
Zong Min Wu, Compactly supported positive definite radial functions, Adv. Comput. Math. 4 (1995), no. 3, 283–292. MR 1357720, DOI 10.1007/BF03177517
Wei Zhao, Yiu-chung Hon, and Martin Stoll, Localized radial basis functions-based pseudo-spectral method (LRBF-PSM) for nonlocal diffusion problems, Comput. Math. Appl. 75 (2018), no. 5, 1685–1704. MR 3766544, DOI 10.1016/j.camwa.2017.11.030