Transferable neural networks for partial differential equations
From MaRDI portal
Numerical computation of solutions to systems of equations (65H10) Numerical optimization and variational techniques (65K10) Stability and convergence of numerical methods for initial value and initial-boundary value problems involving PDEs (65M12) Numerical methods for partial differential equations, initial value and time-dependent initial-boundary value problems (65M99) Numerical solution of discretized equations for boundary value problems involving PDEs (65N22) Artificial neural networks and deep learning (68T07)
Abstract: Transfer learning for partial differential equations (PDEs) is to develop a pre-trained neural network that can be used to solve a wide class of PDEs. Existing transfer learning approaches require much information of the target PDEs such as its formulation and/or data of its solution for pre-training. In this work, we propose to construct transferable neural feature spaces from purely function approximation perspectives without using PDE information. The construction of the feature space involves re-parameterization of the hidden neurons and uses auxiliary functions to tune the resulting feature space. Theoretical analysis shows the high quality of the produced feature space, i.e., uniformly distributed neurons. Extensive numerical experiments verify the outstanding performance of our method, including significantly improved transferability, e.g., using the same feature space for various PDEs with different domains and boundary conditions, and the superior accuracy, e.g., several orders of magnitude smaller mean squared error than the state of the art methods.
Recommendations
- Mosaic flows: a transferable deep learning framework for solving PDEs on unseen domains
- Adaptive transfer learning for PINN
- Local neural operator for solving transient partial differential equations on varied domains
- Solving parametric PDE problems with artificial neural networks
- Deep learning methods for partial differential equations and related parameter identification problems
Cites work
- Approximation by superpositions of a sigmoidal function
- DeepXDE: a deep learning library for solving differential equations
- DGM: a deep learning algorithm for solving partial differential equations
- MIONet: Learning Multiple-Input Operators via Tensor Product
- Neural operator prediction of linear instability waves in high-speed boundary layers
- Neural‐network‐based approximations for solving partial differential equations
- Nonlinear approximation and (deep) ReLU networks
- On the equivalence between kernel quadrature rules and random feature expansions
- PDE-Net 2.0: learning PDEs from data with a numeric-symbolic hybrid deep network
- Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
- Solving Allen-Cahn and Cahn-Hilliard Equations using the Adaptive Physics Informed Neural Networks
- Solving differential equations of fractional order using an optimization technique based on training artificial neural network
- The deep Ritz method: a deep learning-based numerical algorithm for solving variational problems
- The Random Feature Model for Input-Output Maps between Banach Spaces
- Transfer learning based multi-fidelity physics informed deep neural network
- Weak adversarial networks for high-dimensional partial differential equations
Cited in
(23)- Adaptive transfer learning for PINN
- Improving the expressive power of deep neural networks through integral activation transform
- Transformers as neural operators for solutions of differential equations with finite regularity
- Energy-decaying ERK-SAV finite element methods for the time-dependent Ginzburg-Landau equations under the zero electric potential gauge
- Adaptive neural network basis methods for partial differential equations with low-regular solutions
- Subspace method based on neural networks for solving the partial differential equation in weak form
- Local randomized neural networks with discontinuous Galerkin methods for KdV-type and Burgers equations
- DeepONet augmented by randomized neural networks for efficient operator learning in PDEs
- Deep collocation method: a framework for solving PDEs using neural networks with error control
- Subspace method based on neural networks for solving the partial differential equation
- High-precision physics-informed extreme learning machines for evolving interface problems
- Deep learning for model correction of dynamical systems with data scarcity
- Local ultraconvergence of quadratic rectangular element
- Overlapping Schwarz preconditioners for randomized neural networks with domain decomposition
- A multiple transferable neural network method with domain decomposition for elliptic interface problems
- Local randomized neural networks with finite difference methods for interface problems
- The neural network basis method for nonlinear partial differential equations and its Gauss-Newton optimizer
- A novel \(\left( G^{\prime}/G^2\right)\)-expansion neural networks method: application to the analytical study of nonlinear dynamical systems
- Alternately-optimized SNN method for acoustic scattering problem in unbounded domain
- Subspace method based on neural networks for solving eigenvalue problems
- A three-stage PINN for solving inverse problems of piecewise-continuous variable coefficients of PDEs
- Finite expression method with TranNet-based function learning for high-dimensional partial differential equations
- High-order accurate Taylor correction-based backward differentiation formulas with spatial transferable neural networks for moving interface problems
This page was built for publication: Transferable neural networks for partial differential equations
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6123346)
