Recurrent neural networks as optimal mesh refinement strategies

DOI10.1016/j.camwa.2021.05.018MaRDI QIDQ2047561

Jan Bohn, Michael Feischl

Publication date: 20 August 2021

Published in: Computers \& Mathematics with Applications (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/1909.04275

zbMATH Keywords

neural networks; adaptive mesh refinement; partial differential equations; optimal convergence

Mathematics Subject Classification ID

35J25: Boundary value problems for second-order elliptic equations

65N15: Error bounds for boundary value problems involving PDEs

65N12: Stability and convergence of numerical methods for boundary value problems involving PDEs

65N30: Finite element, Rayleigh-Ritz and Galerkin methods for boundary value problems involving PDEs

65N50: Mesh generation, refinement, and adaptive methods for boundary value problems involving PDEs

Related Items

Deep reinforcement learning for adaptive mesh refinement, Quasi-optimal \textit{hp}-finite element refinements towards singularities via deep neural network prediction, \(r\)-adaptive deep learning method for solving partial differential equations, Learning Robust Marking Policies for Adaptive Mesh Refinement, A robust unsupervised neural network framework for geometrically nonlinear analysis of inelastic truss structures, An efficient neural network method with plane wave activation functions for solving Helmholtz equation, Neural control of discrete weak formulations: Galerkin, least squares \& minimal-residual methods with quasi-optimal weights

Uses Software

• Evolino

Cites Work

• Unnamed Item
• Unnamed Item
• Axioms of adaptivity
• Counting symmetry classes of dissections of a convex regular polygon
• Stochastic recursive algorithms for optimization. Simultaneous perturbation methods
• Convergence of a standard adaptive nonconforming finite element method with optimal complexity
• Adaptive finite element methods with convergence rates
• Convergence and complexity of arbitrary order adaptive mixed element methods for the Poisson equation
• Optimal adaptive nonconforming FEM for the Stokes problem
• Quasi-optimal convergence rates for adaptive boundary element methods with data approximation. I: Weakly-singular integral equation
• On existence and uniqueness properties for solutions of stochastic fixed point equations
• Error bounds for approximations with deep ReLU networks
• Quasi-optimal convergence rates for adaptive boundary element methods with data approximation. II: Hyper-singular integral equation
• Adaptive boundary element methods with convergence rates
• Decay rates of adaptive finite elements with Dörfler marking
• A natural adaptive nonconforming FEM of quasi-optimal complexity
• Optimality of a standard adaptive finite element method
• Quasi-optimal Convergence Rate for an Adaptive Boundary Element Method
• The Adaptive Nonconforming FEM for the Pure Displacement Problem in Linear Elasticity is Optimal and Robust
• Convergence and optimality of adaptive mixed finite element methods
• A Convergent Nonconforming Adaptive Finite Element Method with Quasi-Optimal Complexity
• Quasi-Optimality of Adaptive Nonconforming Finite Element Methods for the Stokes Equations
• Quasioptimal cardinality of AFEM driven by nonresidual estimators
• Quasi-Optimal Convergence Rate for an Adaptive Finite Element Method
• Optimality of a Standard Adaptive Finite Element Method for the Stokes Problem
• Efficiency and Optimality of Some Weighted-Residual Error Estimator for Adaptive 2D Boundary Element Methods
• The completion of locally refined simplicial partitions created by bisection
• Adaptive FEM with Optimal Convergence Rates for a Certain Class of Nonsymmetric and Possibly Nonlinear Problems
• Instance optimality of the adaptive maximum strategy