Back to "Padova-Verona research group on
Constructive Approximation and Applications" (CAA) Home
Page
CAA: Stable kernel-based approximation and applications
Kernel-based approximation have become particularly popular in the last decades
especially with the use of RBF (Radial Basis Functions). Our research is addressed
primarily to the analysis and construction of stable bases and their use
in image reconstruction from Radon data and meshless approximations.
People: S. De Marchi (coordinator),
A. Martínez Calomardo, F. Marchetti ,
E. Perracchione, D. Poggiali (Padova)
Collaborators: M. Buhmann (Giessen),
R. Cavoretto (Torino), A. De Rossi (Torino), B. Haasdonk (Stuttgart),
A. Iske (Hamburg), M. Rossini (Milano), G. Santin (Stuttgart), R. Schaback (Goettingen), H. Wendland (Bayreuth).
Papers on theoretical questions
- Jumping with Variably Scaled Discontinuos Kernels (VSDK)
by S. De Marchi, F. Marchetti and E. Perracchione. To appear on BIT Numer. Mathematics.
- On the Convergence of the Rescaled Localized Radial Basis
Function Method
draft (July 2019), with H. Wendland. Appl. Math. Letters (2020) online
-
Analysis of a new class of rational RBF expansions
by M. Buhmann, S. De Marchi and E. Perracchione, Accepted by IMA J. Numer. Anal. (2018)
- Recursive POD expansion for advection-diffusion-reaction equation
by M. Azaïez, T. Chácon Rebollo, E. Perracchione, J. M. Vega,
to appear on Comm. Comput. Physics, 2018, doi: 10.4208/cicp.OA-2017-0257.
- Interpolation with uncoupled separable matrix-valued kernels
by D. Wittwar, G. Santin, B. Haasdonk
Dolomites Res. Notes Approx. 11, Special Issue SPAN2018, pp. 23--39 (2018)
- Greedy kernel methods for accelerating implicit integrators for parametric ODEs
by T. Brünnette, G. Santin, and B. Haasdonk
Proceedings of ENUMATH 2017. (2018)
-
RBF-based partition of unity methods for elliptic PDEs: Adaptivity and stability issues via VSKs
by S. De Marchi, A. Martinez, E. Perracchione and M. Rossini,
J. Sci. Comput. 79(1) (2019), pp. 321-344. online version
-
Fast and stable rational RBF-based Partition of Unity interpolation
by S. De Marchi, A. Martinez, E. Perracchione. J. Comput. Appl. Math. 349 (2019), pp. 331-343
online
- Image Reconstruction from Scattered Radon Data by Weighted Positive Definite Kernel Functions
by S. De Marchi, A. Iske and G. Santin, Calcolo 55(2) (2018), https://doi.org/10.1007/s10092-018-0247-6.
- Convergence rate of the data-independent P-greedy algorithm in kernel-based approximation
by G. Santin and B. Haasdonk, Dolomites Res. Notes on Approx. 10 (2017), pp. 68-78.
-
Optimal selection of local
approximants in RBF-PU interpolation using bivariate LOOCV
arXiv preprint 1703.04282 - R. Cavoretto, A. De Rossi and
E. Perracchione
J. Sci. Comput., to appear
- A rescaled method for RBF approximation
by S. De Marchi, A. Idda and G. Santin
Springer Proceedings on Mathematics and Statistics, Vol. 201 (2017), pp.39--59.
-
Partition of unity interpolation using stable kernel-based techniques
by R. Cavoretto, S. De Marchi, A. De Rossi, E. Perracchione
and G. Santin
Appl. Numer. Math. 116 (2017), pp. 95-107
online, http://dx.doi.org/10.1016/j.apnum.2016.07.005
- Kernel-based Image Reconstruction from Scattered Radon Data
by S. De Marchi, A. Iske and A. Sironi
Dolomites Res. Notes on Approx. Vol 9 (2016), special issue for the workshop "Kernel-based methods and function approximation", pp. 19-31.
available as Hamburger Beitraege
zur Angewandten Mathematik 2016-11
-
Approximation of Eigenfunctions in Kernel-based Spaces
by G. Santin and R. Schaback
Adv. Comput. Math. 42(4) (2016), pp. 973--993.
-
RBF approximation of large datasets by partition of unity
and local stabilization
by R. Cavoretto, S. De Marchi, A. De Rossi, E. Perracchione
and G. Santin
Proceedings CMMSE (2015), Vol. I-II-III-IV, pp. 317--326.
-
Fast computation of orthonormal basis for RBF spaces through Krylov space methods
by S. De Marchi and G. Santin
BIT Numerical Mathematics 55(4) (2015), pp. 949--966.
-
A new stable basis for radial basis function interpolation
by S. De Marchi and G. Santin
J. Comp. Appl. Math., Vol. 253 (2013), pp. 1--13.
-
Stability of Kernel-Based Interpolation
by S. De Marchi and R. Schaback
Adv. Comput. Math. Vol. 32(2), 2010, p. 155-161
Examples and more:
these are examples and figures illustrating the results of the paper "Stability of Kernel-Based Interpolation".
-
Nonstandard Kernels and their Applications
by S. De Marchi and R. Schaback
Dolomites Res. Notes on Approx. (DRNA) Vol. 2, (2009), pp. 16--43.
-
Univariate Radial Basis Functions with Compact Support Cardinal Functions
by L. Bos and S. De Marchi
East J. Approx., Vol. 14(1) 2008, pp. 69--80.
-
Near-Optimal Data-independent Point Locations for Radial Basis Function Interpolation
by with R. Schaback and H. Wendland
Adv. Comput. Math. 23(3) (2005), pp. 317--330.
Papers on applications
-
Shape-Driven Interpolation with Discontinuous Kernels: Error Analysis, Edge Extraction and Applications in MPI
S.De Marchi, W. Erb, F. Marchetti, E. Perracchione and M. Rossini. Accepted by SIAM J. Sci. Comput.
- A stable meshfree PDE solver for source-type flows in porous media
R. Campagna, S. Cuomo, S. De Marchi, E. Perracchione and G. Severino, Appl. Num. Math. on line
- Learning with subsampled kernel-based methods: Environmental and financial applications
by M. Aminian Shahrokhabadi, A. Neisy, E. Perracchione, M. Polato, Dolomites Res. Notes Approx. 12(2019), pp. 17-27.
- Greedy Kernel Approximation for Sparse Surrogate Modelling
by B. Haasdonk and G. Santin
.
in Reduced-Order Modeling (ROM) for Simulation and Optimization: Powerful Algorithms as Key Enablers for Scientific Computing,
W. Keiper, A. Milde, and S. Volkwein, Eds. Cham: Springer Int. Pub., 2018, pp. 21--45.
- Comparison of
data-driven uncertainty quantification methods for a carbon dioxide storage benchmark scenario
by M. Köppel, F. Franzelin, I. Kröker, G. Santin, D. Wittwar, S. Oladyshkin, A. Barth, B. Haasdonk, W. Nowak, D. Pflüger, C. Rohde,
Accepted for publication in Comput. Geosci.
- Numerical modelling of a peripheral arterial stenosis using dimensionally reduced models and kernel
methods,
by T. Köppl, G. Santin, B. Haasdonk, R. Helmig
Int. J. Numer. Meth. Biomed. Engng., vol 34, 8 (2018), pg. e3095
- Rational RBF-based partition of unity method for efficiently
and accurately approximating 3D objects
by E. Perracchione, arXiv preprint arXiv:1802.01842, 2018. To appear in Comput. Appl. Math.
-
Approximating basins of attraction for dynamical systems
via stable radial bases
preprint - R. Cavoretto, S. De Marchi, A. De Rossi, E. Perracchione
and G. Santin
AIP Conference Proceedings, 1738, 390003 (2016); doi:10.1063/1.4952177
online
Submitted
-
Shape-Driven Interpolation with Discontinuous Kernels: Error Analysis, Edge Extraction and Applications in MPI
by S. De Marchi, W. Erb, F. Marchetti, E. Perracchione and M. Rossini (March 2109)
- Greedy regularized kernel interpolation
G. Santin, D. Wittwar, B. Haasdonk, submitted (2018).
Posters
- Smooth kernel machines for environment and finance
by M. Aminian Shahrokhabadi, E. Perracchione, M. Polato, M. Putti
poster presented at the "Dolomites Research Week on Approximation 2018", Alba di Canazei (Italy), Sept. 2018.
-
Rational stable RBF-PU interpolation via VSKs
by S. De Marchi, A. Martinez and E. Perracchione
poster presented at the "Dolomites Research Week on Approximation 2017", Alba di Canazei (Italy), Sept. 2017.
- A rescaled method for RBF approximation
by S. De Marchi, A. Idda and G. Santin
poster presented at "4th Workshop on Constructive Approximation and Applications", Alba di Canazei (TN- Italy), Sept. 2016.
- Kernel methods for Radon transform
by S. Guglielmo and G. Santin
poster presented at International CAE Conference 2013, Pacengo del Garda (Italy), Oct. 2013.
-
WSVD basis for RBF and Krylov subspaces
by S. De Marchi and G. Santin
poster presented at the "Dolomites Research Week on Approximation 2013", Alba di Canazei (Italy), Sept. 2013.
-
A New Stable Basis for RBF Approximation
by S. De Marchi and G. Santin
poster presented at the "Dolomites Research Week on Approximation 2012", Alba di Canazei (Italy), Sept. 2012.
Software
Books
Meshfree Approximation for Multi-Asset European and American Option Problems
by Stefano De Marchi, Maddalena Mandarà and Anna Viero
ISBN: 9788854851511 (2012), pp. 92.
Lecture Notes
Lectures on radial basis functions by S. De Marchi and E. Perracchione (always in progress), pp. 13?
Presentations
- S. De Marchi: New developments on rational RBF
- S. De Marchi: Kernel-based Image
Reconstruction from scattered Radon data by (anisotropic) positive definite functions
Kernel-based methods and function approximation - Department of Mathematics, University of Torino (Italy), February 5th, 2016.
- S. De Marchi: On a new orthonormal basis for RBF native spaces and its fast computation
Colloqium at the Department of Mathematics, University of Torino (Italy), on June 11th, 2014.
- G. Santin: A fast algorithm for computing a truncated orthonormal basis for RBF native spaces
Multivariate Approximation, Verona 29-30 November, 2013.
- S. De Marchi: On a new orthonormal basis for RBF native spaces
San Diego (USA), SIAM Annual meeting: July 8th, 2013.
PhD theses
Master theses
- Medical image reconstruction using kernel based methods, candidate: Amos Sironi,
University of Padua, A. Y. 2010-11.
- A new stable basis for RBF approximation, candidate: Gabriele Santin, University
of Padua, A. Y. 2011-12.
- Radial basis functions networks for ODEs: application to diabetes and insulin
therapy models, candidate: Giulia Antinori, University of Padua, A. Y. 2011-12.
- A Scilab radial basis functions toolbox, candidate: Anna Bassi, University of
Padua, A. Y. 2011-12.
- Reconstruction of medical images from Radon data in trasmission and emission
tomography, candidate: Davide Poggiali, University of Padua, A. Y. 2011-12.~demarchi/
- Kernel-based medical image reconstruction, candidate: Maria Angela Narduzzo,
University of Padua, A. Y. 2013-14.
- Kernel-based medical image reconstruction from Radon data, candidate: Silvia
Guglielmo, University of Padua, A. Y. 2013-14.
- A comparison of some RBF interpolation methods: theory and numerics, candidate:
Andrea Idda, University of Padova, A.Y. 2014-15.
- Numerical solution of PDEs on general surfaces by RBFs, candidate: Sara Carlino,
University of Padova, A.Y. 2015-16.
- A partition of unity radial basis functions method for solving PDEs, candidate: Danilo Stocchino, University of Padova A. Y. 2017-18.
Degree theses
- Radial basis functions approximation for European call option price, candidate:
Maddalena Mandarà, University of Verona, A. Y. 2007-08.
- Meshfree approximation for multi-asset American option problems, candidate: Anna
Viero, University of Verona, A. Y. 2007-08.