Deep learning architectures for nonlinear operator functions and nonlinear inverse problems
From MaRDI portal
Publication:2113263
DOI10.4171/MSL/28zbMATH Open1485.35420arXiv1912.11090MaRDI QIDQ2113263FDOQ2113263
Maarten V. De Hoop, Christopher A. Wong, M. Lassas
Publication date: 11 March 2022
Published in: Mathematical Statistics and Learning (Search for Journal in Brave)
Abstract: We develop a theoretical analysis for special neural network architectures, termed operator recurrent neural networks, for approximating nonlinear functions whose inputs are linear operators. Such functions commonly arise in solution algorithms for inverse boundary value problems. Traditional neural networks treat input data as vectors, and thus they do not effectively capture the multiplicative structure associated with the linear operators that correspond to the data in such inverse problems. We therefore introduce a new family that resembles a standard neural network architecture, but where the input data acts multiplicatively on vectors. Motivated by compact operators appearing in boundary control and the analysis of inverse boundary value problems for the wave equation, we promote structure and sparsity in selected weight matrices in the network. After describing this architecture, we study its representation properties as well as its approximation properties. We furthermore show that an explicit regularization can be introduced that can be derived from the mathematical analysis of the mentioned inverse problems, and which leads to certain guarantees on the generalization properties. We observe that the sparsity of the weight matrices improves the generalization estimates. Lastly, we discuss how operator recurrent networks can be viewed as a deep learning analogue to deterministic algorithms such as boundary control for reconstructing the unknown wavespeed in the acoustic wave equation from boundary measurements.
Full work available at URL: https://arxiv.org/abs/1912.11090
Recommendations
- Deep learning for inverse problems with unknown operator
- Deep neural networks can stably solve high-dimensional, noisy, non-linear inverse problems
- Convolutional Neural Networks in Phase Space and Inverse Problems
- NETT: solving inverse problems with deep neural networks
- Big in Japan: regularizing networks for solving inverse problems
Learning and adaptive systems in artificial intelligence (68T05) Neural nets and related approaches to inference from stochastic processes (62M45) Inverse problems for PDEs (35R30)
Cites Work
- Title not available (Why is that?)
- Title not available (Why is that?)
- Title not available (Why is that?)
- Probability Inequalities for Sums of Bounded Random Variables
- Stability estimates for the hyperbolic Dirichlet to Neumann map in anisotropic media
- Scattering control for the wave equation with unknown wave speed
- Two dimensional compact simple Riemannian manifolds are boundary distance rigid
- Boundary rigidity with partial data
- A Lipschitz stable reconstruction formula for the inverse problem for the wave equation
- An iterative thresholding algorithm for linear inverse problems with a sparsity constraint
- Reconstruction of piecewise smooth wave speeds using multiple scattering
- A global uniqueness theorem for an inverse boundary value problem
- Stability estimates for the anisotropic wave equation from the Dirichlet-to-Neumann map
- Boundary regularity for the Ricci equation, geometric convergence, and Gel'fand's inverse boundary problem
- Title not available (Why is that?)
- Semiglobal boundary rigidity for Riemannian metrics
- Inverse problems for Lorentzian manifolds and non-linear hyperbolic equations
- Inverse problem for the Riemannian wave equation with Dirichlet data and Neumann data on disjoint sets
- To the reconstruction of a riemannian manifold via its spectral data (Bc–Method)
- Boundary rigidity and filling volume minimality of metrics close to a flat one
- Equivalence of time-domain inverse problems and boundary spectral problems
- Iterative time-reversal control for inverse problems
- Inverse boundary value problems for partial differential equations
- Theory and applications of time reversal and interferometric imaging
- Time Reversal and Refocusing in Random Media
- On statistical Calderón problems
- Stability For a Multidimensional Inverse Spectral Theorem
- Consistency of Bayesian inference with Gaussian process priors in an elliptic inverse problem
- Error bounds for approximations with deep ReLU networks
- Multiscale sparse microcanonical models
- SwitchNet: A Neural Network Model for Forward and Inverse Scattering Problems
- Solving ill-posed inverse problems using iterative deep neural networks
- Deep Convolutional Neural Network for Inverse Problems in Imaging
- Lipschitz Certificates for Layered Network Structures Driven by Averaged Activation Operators
- NETT: solving inverse problems with deep neural networks
- Solving inverse problems using data-driven models
- Recovery of a Smooth Metric via Wave Field and Coordinate Transformation Reconstruction
- Regularization strategy for an inverse problem for a 1 + 1 dimensional wave equation
- Solving an inverse obstacle problem for the wave equation by using the boundary control method
- Splines Are Universal Solutions of Linear Inverse Problems with Generalized TV Regularization
- Title not available (Why is that?)
- Networks for nonlinear diffusion problems in imaging
- Deep learning for photoacoustic tomography from sparse data
- Learning the invisible: a hybrid deep learning-shearlet framework for limited angle computed tomography
- Discrete regularization and convergence of the inverse problem for 1+1 dimensional wave equation
- Quantifying the generalization error in deep learning in terms of data distribution and neural network smoothness
- INVERSE PROBLEMS FOR LINEAR AND NON-LINEAR HYPERBOLIC EQUATIONS
- Statistical guarantees for Bayesian uncertainty quantification in nonlinear inverse problems with Gaussian process priors
Cited In (7)
- A Data-Assisted Two-Stage Method for the Inverse Random Source Problem
- Out-of-distributional risk bounds for neural operators with applications to the Helmholtz equation
- The Calderón's problem via DeepONets
- Improved architectures and training algorithms for deep operator networks
- Convergence Rates for Learning Linear Operators from Noisy Data
- Data-consistent neural networks for solving nonlinear inverse problems
- Inverse problems on low-dimensional manifolds
Uses Software
This page was built for publication: Deep learning architectures for nonlinear operator functions and nonlinear inverse problems
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2113263)
