ConvPDE-UQ: convolutional neural networks with quantified uncertainty for heterogeneous elliptic partial differential equations on varied domains
confidence intervalconvolutional encoder-decoder networksdeep learningmachine learningpartial differential equationsuncertainty quantification
Laplace operator, Helmholtz equation (reduced wave equation), Poisson equation (35J05) Probabilistic methods, particle methods, etc. for boundary value problems involving PDEs (65N75) Fundamental solutions, Green's function methods, etc. for boundary value problems involving PDEs (65N80) Learning and adaptive systems in artificial intelligence (68T05) Artificial neural networks and deep learning (68T07) Neural nets applied to problems in time-dependent statistical mechanics (82C32)
The aim is to construct new solvers for partial differential equations (PDE) based on convolutional neural networks. One considers the classical Poisson equation with Dirichlet boundary condition, $\Delta u = f$ in $\Omega$, $ u=0$ on $\partial\Omega$, $\Omega$ being the two-dimensional unit disk. The solution can be expressed by means of the Green function $G(x,y)$, $u(x)=\int_{\Omega}G(x,y)f(y)dy$, for $x\in\Omega$. A discretization procedure is applied and the purpose is to define a neural network which approximates the discretized solution mapping. The third section of the article is devoted to the presentation of the network architecture, the probabilistic training procedure, network loss functions, data generation. Numerical results, comparison of training procedures, uncertainty quantification and inference speed are shown in the fourth section and a brief summary on the developed ConvPDE-UQ framework in the fifth section. Appendix A presents results from the extension of the method to variable coefficients differential operators or Neumann boundary conditions. Appendix B deals with alternative distributions for modeling network uncertainty.
- Deep neural network approximation for high-dimensional elliptic PDEs with boundary conditions
- Solving parametric PDE problems with artificial neural networks
- Numerical solution of the parametric diffusion equation by deep neural networks
- MOD-Net: a machine learning approach via model-operator-data network for solving PDEs
- On the convergence of physics informed neural networks for linear second-order elliptic and parabolic type PDEs
- A table of integrals of the Error functions
- Algorithm 839
- Approximation by superpositions of a sigmoidal function
- Automated code generation for discontinuous Galerkin methods
- Automated solution of differential equations by the finite element method. The FEniCS book
- Automating the finite element method
- Bayesian deep convolutional encoder-decoder networks for surrogate modeling and uncertainty quantification
- Bayesian solution uncertainty quantification for differential equations
- Calibration of reduced-order model for a coupled Burgers equations based on PC-EnKF
- Deep learning-based numerical methods for high-dimensional parabolic partial differential equations and backward stochastic differential equations
- Deep UQ: learning deep neural network surrogate models for high dimensional uncertainty quantification
- DGM: a deep learning algorithm for solving partial differential equations
- DOLFIN: automated finite element computing
- Efficient Assembly of $H(\mathrm{div})$ and $H(\mathrm{curl})$ Conforming Finite Elements
- Efficient compilation of a class of variational forms
- Framework for massively parallel adaptive finite element computational fluid dynamics on tetrahedral meshes
- Geometric Optimization of the Evaluation of Finite Element Matrices
- scientific article; zbMATH DE number 193842 (Why is no real title available?)
- Multilayer feedforward networks are universal approximators
- Neural algorithm for solving differential equations
- Neural‐network‐based approximations for solving partial differential equations
- Numerical solution for high order differential equations using a hybrid neural network-optimization method
- On the efficiency of symbolic computations combined with code generation for finite element methods
- On the unimodality of the likelihood for the Cauchy distribution
- Optimal observations-based retrieval of topography in 2D shallow water equations using PC-EnKF
- Optimizations for quadrature representations of finite element tensors through automated code generation
- Optimizing the Evaluation of Finite Element Matrices
- Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
- Sparsity-promoting elastic net method with rotations for high-dimensional nonlinear inverse problem
- Topological Optimization of the Evaluation of Finite Element Matrices
- Unicorn: parallel adaptive finite element simulation of turbulent flow and fluid-structure interaction for deforming domains and complex geometry
- Unified form language: a domain-specific language for weak formulations of partial differential equations
- ConvPDE-UQ
- PPINN: parareal physics-informed neural network for time-dependent PDEs
- PhyCRNet: physics-informed convolutional-recurrent network for solving spatiotemporal PDEs
- A modified batch intrinsic plasticity method for pre-training the random coefficients of extreme learning machines
- Learning time-dependent PDEs with a linear and nonlinear separate convolutional neural network
- Output-based adaptive aerodynamic simulations using convolutional neural networks
- Local extreme learning machines and domain decomposition for solving linear and nonlinear partial differential equations
- On computing the hyperparameter of extreme learning machines: algorithm and application to computational PDEs, and comparison with classical and high-order finite elements
- A survey of unsupervised learning methods for high-dimensional uncertainty quantification in black-box-type problems
- The Random Feature Model for Input-Output Maps between Banach Spaces
- DeepXDE: a deep learning library for solving differential equations
- Deep neural networks for waves assisted by the Wiener-Hopf method
- Fully probabilistic deep models for forward and inverse problems in parametric PDEs
- Accuracy and architecture studies of residual neural network method for ordinary differential equations
- Addressing discontinuous root-finding for subsequent differentiability in machine learning, inverse problems, and control
- Label-free learning of elliptic partial differential equation solvers with generalizability across boundary value problems
- A deep neural network-based method for solving obstacle problems
- Numerical computation of partial differential equations by hidden-layer concatenated extreme learning machine
- Bi-fidelity modeling of uncertain and partially unknown systems using DeepONets
- QuadConv: quadrature-based convolutions with applications to non-uniform PDE data compression
- Energy-dissipative evolutionary deep operator neural networks
- Physics-informed ConvNet: learning physical field from a shallow neural network
- Learning domain-independent Green's function for elliptic partial differential equations
- On the locality of local neural operator in learning fluid dynamics
- A causality-DeepONet for causal responses of linear dynamical systems
- Operator learning using random features: a tool for scientific computing
- Machine learning approaches for the solution of the Riemann problem in fluid dynamics: a case study
- f-PICNN: a physics-informed convolutional neural network for partial differential equations with space-time domain
- Uncertainty quantification for noisy inputs-outputs in physics-informed neural networks and neural operators
- Convolutional autoencoders, clustering, and POD for low-dimensional parametrization of flow equations
- From obstacle problems to neural insights: feedforward neural network modeling of ice thickness
- The deep Ritz methods for the elastic contact problems with the Signorini, Tresca-friction, Coulomb-friction, and non-convex non-smooth subgradient-type boundary conditions
- MFPC-Net: multi-fidelity physics-constrained neural process
- Active operator learning with predictive uncertainty quantification for partial differential equations
- An inverse scattering inspired Fourier neural operator for time-dependent PDE learning
This page was built for publication: ConvPDE-UQ: convolutional neural networks with quantified uncertainty for heterogeneous elliptic partial differential equations on varied domains
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2222287)
