TNet: A Model-Constrained Tikhonov Network Approach for Inverse Problems
From MaRDI portal
Publication:6154957
Abstract: Deep Learning (DL), in particular deep neural networks (DNN), by default is purely data-driven and in general does not require physics. This is the strength of DL but also one of its key limitations when applied to science and engineering problems in which underlying physical properties and desired accuracy need to be achieved. DL methods in their original forms are not capable of respecting the underlying mathematical models or achieving desired accuracy even in big-data regimes. However, many data-driven science and engineering problems, such as inverse problems, typically have limited experimental or observational data, and DL would overfit the data in this case. Leveraging information encoded in the underlying mathematical models, we argue, not only compensates missing information in low data regimes but also provides opportunities to equip DL methods with the underlying physics, hence promoting better generalization. This paper develops a model-constrained deep learning approach and its variant TNet that are capable of learning information hidden in both the training data and the underlying mathematical models to solve inverse problems governed by partial differential equations. We provide the constructions and some theoretical results for the proposed approaches. We show that data randomization can enhance the smoothness of the networks and their generalizations. Comprehensive numerical results not only confirm the theoretical findings but also show that with even as little as 20 training data samples for 1D deconvolution, 50 for inverse 2D heat conductivity problem, 100 and 50 for inverse initial conditions for time-dependent 2D Burgers' equation and 2D Navier-Stokes equations, respectively. TNet solutions can be as accurate as Tikhonov solutions while being several orders of magnitude faster. This is possible owing to the model-constrained term, replications, and randomization.
Recommendations
Cites work
- Accelerating Markov chain Monte Carlo with active subspaces
- Adversarial uncertainty quantification in physics-informed neural networks
- Approximation by superpositions of a sigmoidal function
- Computational Methods for Inverse Problems
- Deep learning
- Deep UQ: learning deep neural network surrogate models for high dimensional uncertainty quantification
- DeepXDE: a deep learning library for solving differential equations
- Foundations of machine learning
- fPINNs: Fractional Physics-Informed Neural Networks
- Hidden physics models: machine learning of nonlinear partial differential equations
- scientific article; zbMATH DE number 1667894 (Why is no real title available?)
- scientific article; zbMATH DE number 1744773 (Why is no real title available?)
- scientific article; zbMATH DE number 803211 (Why is no real title available?)
- Inverse Problem Theory and Methods for Model Parameter Estimation
- Linear and nonlinear inverse problems with practical applications
- Machine learning of linear differential equations using Gaussian processes
- Multilayer feedforward networks are universal approximators
- Multiscale topology optimization using neural network surrogate models
- NETT: solving inverse problems with deep neural networks
- Numerical Optimization
- Pattern recognition and machine learning.
- Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
- Solving ill-posed inverse problems using iterative deep neural networks
- Solving inverse-PDE problems with physics-aware neural networks
- Statistical and computational inverse problems.
- The finite element methods for elliptic problems.
- Theory and practice of finite elements.
- Understanding machine learning. From theory to algorithms
Cited in
(9)- Tubenet: A Special Trumpetnet for Explicit Solutions to Inverse Problems
- Discussion of “Experimental Design and Modeling for Forward-Inverse Maps” by R. Barton & M. Morris, appearing in Technometrics
- TDDM: a transfer learning framework for physics-guided 3D acoustic scattering inversion
- \texttt{TAEN}: a model-constrained Tikhonov autoencoder network for forward and inverse problems
- Physics-aware deep learning framework for the limited aperture inverse obstacle scattering problem
- Two new calibration techniques of lumped-parameter mathematical models for the cardiovascular system
- Improvements on uncertainty quantification with variational autoencoders
- Neural networks for Bayesian inverse problems governed by a nonlinear ODE
- Real-time optimal control of high-dimensional parametrized systems by deep learning-based reduced order models
This page was built for publication: TNet: A Model-Constrained Tikhonov Network Approach for Inverse Problems
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6154957)
