Variational training of neural network approximations of solution maps for physical models
From MaRDI portal
Publication:778301
DOI10.1016/J.JCP.2020.109338zbMATH Open1435.68292arXiv1905.02789OpenAlexW2944808962MaRDI QIDQ778301FDOQ778301
Authors: Yingzhou Li, Anqi Mao, Jianfeng Lu 
Publication date: 2 July 2020
Published in: Journal of Computational Physics (Search for Journal in Brave)
Abstract: A novel solve-training framework is proposed to train neural network in representing low dimensional solution maps of physical models. Solve-training framework uses the neural network as the ansatz of the solution map and train the network variationally via loss functions from the underlying physical models. Solve-training framework avoids expensive data preparation in the traditional supervised training procedure, which prepares labels for input data, and still achieves effective representation of the solution map adapted to the input data distribution. The efficiency of solve-training framework is demonstrated through obtaining solutions maps for linear and nonlinear elliptic equations, and maps from potentials to ground states of linear and nonlinear Schr"odinger equations.
Full work available at URL: https://arxiv.org/abs/1905.02789
Recommendations
- Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
- Galerkin neural networks: a framework for approximating variational equations with error control
- scientific article; zbMATH DE number 1221486
- Solving parametric PDE problems with artificial neural networks
- Model reduction and neural networks for parametric PDEs
Artificial neural networks and deep learning (68T07) NLS equations (nonlinear Schrödinger equations) (35Q55)
Cites Work
- DGM: a deep learning algorithm for solving partial differential equations
- Universal approximation bounds for superpositions of a sigmoidal function
- A sparse matrix arithmetic based on \({\mathfrak H}\)-matrices. I: Introduction to \({\mathfrak H}\)-matrices
- Data-sparse approximation by adaptive \({\mathcal H}^2\)-matrices
- Hierarchical interpolative factorization for elliptic operators: integral equations
- A fast algorithm for particle simulations
- Computing the Ground State Solution of Bose--Einstein Condensates by a Normalized Gradient Flow
- Hierarchical Matrices: Algorithms and Analysis
- A kernel-independent adaptive fast multipole algorithm in two and three dimensions
- A Multigrid Tutorial, Second Edition
- Superfast Multifrontal Method for Large Structured Linear Systems of Equations
- The black-box fast multipole method
- Fast construction of hierarchical matrix representation from matrix-vector multiplication
- Solving high-dimensional partial differential equations using deep learning
- Hierarchical Interpolative Factorization for Elliptic Operators: Differential Equations
- An \(O(N)\) direct solver for integral equations on the plane
- The Deep Ritz Method: a deep learning-based numerical algorithm for solving variational problems
- BCR-net: A neural network based on the nonstandard wavelet form
- Physics-constrained deep learning for high-dimensional surrogate modeling and uncertainty quantification without labeled data
- Solving for high-dimensional committor functions using artificial neural networks
- A multiscale neural network based on hierarchical nested bases
- Distributed-memory hierarchical interpolative factorization
- Solving the quantum many-body problem with artificial neural networks
Cited In (16)
- Strong Solutions for PDE-Based Tomography by Unsupervised Learning
- AMS-Net: Adaptive Multiscale Sparse Neural Network with Interpretable Basis Expansion for Multiphase Flow Problems
- The Random Feature Model for Input-Output Maps between Banach Spaces
- Operator learning using random features: a tool for scientific computing
- Learning to Discretize: Solving 1D Scalar Conservation Laws via Deep Reinforcement Learning
- A new efficient approximation scheme for solving high-dimensional semilinear PDEs: control variate method for deep BSDE solver
- A variational framework for computing Wannier functions using dictionary learning
- Fractional weak adversarial networks for the stationary fractional advection dispersion equations
- A Reduced Order Schwarz Method for Nonlinear Multiscale Elliptic Equations Based on Two-Layer Neural Networks
- Solving inverse problems with deep learning
- Accuracy and architecture studies of residual neural network method for ordinary differential equations
- Control variate method for deep BSDE solver using weak approximation
- Variational inference in neural functional prior using normalizing flows: application to differential equation and operator learning problems
- Variational Bayes Solution of Linear Neural Networks and Its Generalization Performance
- Predicting continuum breakdown with deep neural networks
- Using neural networks to accelerate the solution of the Boltzmann equation
Uses Software
This page was built for publication: Variational training of neural network approximations of solution maps for physical models
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q778301)
