A first approach to learning a best basis for gravitational field modelling

DOI10.1007/S13137-020-0143-5MaRDI QIDQ2300570zbMATH OpenOpenAlexFDO

Authors V. Michel, Naomi Schneider

Publication date 27 February 2020

Published in GEM - International Journal on Geomathematics (Search for Journal in Brave)

Copyright license Creative Commons Attribution 4.0 International

Full work available at URL https://arxiv.org/abs/1901.04222

zbMATH Keywords

dictionary learningdownward continuationgreedy algorithminverse problemmatching pursuitnonlinear optimizationradial basis functionsspherical harmonics

Mathematics Subject Classification ID

Boundary value and inverse problems for harmonic functions in higher dimensions (31B20) Approximation by arbitrary linear expressions (41A45) Algorithms for approximation of functions (65D15) Numerical solutions of ill-posed problems in abstract spaces; regularization (65J20) Numerical optimization and variational techniques (65K10) Numerical methods for ill-posed problems for boundary value problems involving PDEs (65N20) Numerical methods for inverse problems for integral equations (65R32) Learning and adaptive systems in artificial intelligence (68T05) Inverse problems in geophysics (86A22)

Abstract: Gravitational field modelling is an important tool for inferring past and present dynamic processes of the Earth. Functions on the sphere such as the gravitational potential are usually expanded in terms of either spherical harmonics or radial basis functions (RBFs). The (Regularized) Functional Matching Pursuit ((R)FMP) and its variants use an overcomplete dictionary of diverse trial functions to build a best basis as a sparse subset of the dictionary and compute a model, for instance, of the gravity field, in this best basis. Thus, one advantage is that the dictionary may contain spherical harmonics and RBFs. Moreover, these methods represent a possibility to obtain an approximative and stable solution of an ill-posed inverse problem, such as the downward continuation of gravitational data from the satellite orbit to the Earth's surface, but also other inverse problems in geomathematics and medical imaging. A remaining drawback is that in practice, the dictionary has to be finite and, so far, could only be chosen by rule of thumb or trial-and-error. In this paper, we develop a strategy for automatically choosing a dictionary by a novel learning approach. We utilize a non-linear constrained optimization problem to determine best-fitting RBFs (Abel-Poisson kernels). For this, we use the Ipopt software package with an HSL subroutine. Details of the algorithm are explained and first numerical results are shown.

Recommendations

• Toward multiresolution estimation and efficient representation of gravitational fields
• The ARM-Approach Based Local Modelling of the Gravitational Field
• Physics-informed neural networks for gravity field modeling of the Earth and Moon
• Physics-informed neural networks for gravity field modeling of small bodies
• Publication:4727300
• scientific article; zbMATH DE number 1433598
• Correlation modeling of the gravity field in classical geodesy

Cites work

• A greedy algorithm for nonlinear inverse problems with an application to nonlinear inverse gravimetry
• A non-linear approximation method on the sphere
• Adaptive Barrier Update Strategies for Nonlinear Interior Methods
• Approximate decorrelation and non-isotropic smoothing of time-variable GRACE-type gravity field models
• From Sparse Solutions of Systems of Equations to Sparse Modeling of Signals and Images
• Geomathematically oriented potential theory
• scientific article; zbMATH DE number 3953478 (Why is no real title available?)
• scientific article; zbMATH DE number 41289 (Why is no real title available?)
• scientific article; zbMATH DE number 1128580 (Why is no real title available?)
• scientific article; zbMATH DE number 2057413 (Why is no real title available?)
• scientific article; zbMATH DE number 936298 (Why is no real title available?)
• Learning: Wavelet-dictionaries and continuous dictionaries.
• Lectures on constructive approximation. Fourier, spline, and wavelet methods on the real line, the sphere, and the ball
• Line Search Filter Methods for Nonlinear Programming: Local Convergence
• Line Search Filter Methods for Nonlinear Programming: Motivation and Global Convergence
• Multiscale potential theory. With applications to geoscience
• Novel algorithms of greedy-type for probability density estimation as well as linear and nonlinear inverse problems
• On the convergence theorem for the regularized functional matching pursuit (RFMP) algorithm
• On the implementation of an interior-point filter line-search algorithm for large-scale nonlinear programming
• Sparse regularization of inverse gravimetry -- case study: spatial and temporal mass variations in South America
• Special functions of mathematical (geo-)physics
• Spherical functions of mathematical geosciences. A scalar, vectorial, and tensorial setup
• Spherical harmonics
• Spline multiresolution and numerical results for joint gravitation and normal-mode inversion with an outlook on sparse regularisation
• The regularized orthogonal functional matching pursuit for ill-posed inverse problems
• The regularized weak functional matching pursuit for linear inverse problems
• Vectorial Slepian functions on the ball

Cited in

(5)

• Different radial basis functions and their applicability for regional gravity field representation on the sphere
• Regularized functional matching pursuit for the spherical multiple-shell electro-magnetoencephalography problem
• The ARM-Approach Based Local Modelling of the Gravitational Field
• High-dimensional experiments for the downward continuation using the LRFMP algorithm
• Forward modeling with forced neural networks for gravity anomaly profile

Describes a project that uses

Uses Software

• HSL
• Ipopt

This page was built for publication: A first approach to learning a best basis for gravitational field modelling

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2300570)