DGM: a deep learning algorithm for solving partial differential equations
From MaRDI portal
Publication:2002333
deep learninghigh-dimensional partial differential equationsmachine learningpartial differential equations
Quasilinear parabolic equations (35K59) Monte Carlo methods (65C05) Stability and convergence of numerical methods for initial value and initial-boundary value problems involving PDEs (65M12) Probabilistic methods, particle methods, etc. for initial value and initial-boundary value problems involving PDEs (65M75) Learning and adaptive systems in artificial intelligence (68T05)
Abstract: High-dimensional PDEs have been a longstanding computational challenge. We propose to solve high-dimensional PDEs by approximating the solution with a deep neural network which is trained to satisfy the differential operator, initial condition, and boundary conditions. Our algorithm is meshfree, which is key since meshes become infeasible in higher dimensions. Instead of forming a mesh, the neural network is trained on batches of randomly sampled time and space points. The algorithm is tested on a class of high-dimensional free boundary PDEs, which we are able to accurately solve in up to dimensions. The algorithm is also tested on a high-dimensional Hamilton-Jacobi-Bellman PDE and Burgers' equation. The deep learning algorithm approximates the general solution to the Burgers' equation for a continuum of different boundary conditions and physical conditions (which can be viewed as a high-dimensional space). We call the algorithm a "Deep Galerkin Method (DGM)" since it is similar in spirit to Galerkin methods, with the solution approximated by a neural network instead of a linear combination of basis functions. In addition, we prove a theorem regarding the approximation power of neural networks for a class of quasilinear parabolic PDEs.
Recommendations
- Deep learning-based numerical methods for high-dimensional parabolic partial differential equations and backward stochastic differential equations
- Deep backward schemes for high-dimensional nonlinear PDEs
- Extensions of the deep Galerkin method
- Deep neural network approximations for solutions of PDEs based on Monte Carlo algorithms
- Algorithms for solving high dimensional PDEs: from nonlinear Monte Carlo to machine learning
Cites work
- A highly parallel Black--Scholes solver based on adaptive sparse grids
- Analysis of linear difference schemes in the sparse grid combination technique
- Approximation by superpositions of a sigmoidal function
- Compact sets in the space L^ p(0,T;B)
- Convergence of numerical schemes for the solution of parabolic stochastic partial differential equations
- Deep learning-based numerical methods for high-dimensional parabolic partial differential equations and backward stochastic differential equations
- Efficient Hierarchical Approximation of High‐Dimensional Option Pricing Problems
- Existence of solutions for some ``noncoercive parabolic equations
- Existence result for nonlinear parabolic equations with lower order terms
- Existence results for a Cauchy-Dirichlet parabolic problem with a repulsive gradient term
- Gradient Convergence in Gradient methods with Errors
- HJB equations in infinite dimensions under weak regularizing properties
- scientific article; zbMATH DE number 780705 (Why is no real title available?)
- scientific article; zbMATH DE number 1405266 (Why is no real title available?)
- scientific article; zbMATH DE number 3895476 (Why is no real title available?)
- scientific article; zbMATH DE number 3398554 (Why is no real title available?)
- Lattice approximations for stochastic quasi-linear parabolic partial differential equations driven by space-time white noise. I
- Monte Carlo valuation of American options
- Neural algorithm for solving differential equations
- Nonlinear parabolic equations with measure data
- Numerical solution for high order differential equations using a hybrid neural network-optimization method
- Optimal control of a stochastic heat equation with boundary-noise and boundary-control
- Option pricing with a direct adaptive sparse grid approach
- Pricing American Options: A Duality Approach
- Reynolds averaged turbulence modelling using deep neural networks with embedded invariance
- Sparse grids
- Stationary Hamilton-Jacobi equations in Hilbert spaces and applications to a stochastic optimal control problem
- Stochastic simulation: Algorithms and analysis
Cited in
(only showing first 100 items - show all)- Mini-workshop: Deep learning and inverse problems. Abstracts from the mini-workshop held March 4--10, 2018
- An immersed boundary neural network for solving elliptic equations with singular forces on arbitrary domains
- Traveling wave solutions of partial differential equations via neural networks
- Networks for nonlinear diffusion problems in imaging
- A new approach to the numerical solution of Fredholm integral equations using least squares-support vector regression
- Deep learned finite elements
- \textit{hp}-VPINNs: variational physics-informed neural networks with domain decomposition
- Neural networks-based backward scheme for fully nonlinear PDEs
- Overcoming the curse of dimensionality in the numerical approximation of Allen-Cahn partial differential equations via truncated full-history recursive multilevel Picard approximations
- Efficient approximation of solutions of parametric linear transport equations by ReLU DNNs
- A neural network-based policy iteration algorithm with global \(H^2\)-superlinear convergence for stochastic games on domains
- Hierarchical deep-learning neural networks: finite elements and beyond
- Deep autoencoder based energy method for the bending, vibration, and buckling analysis of Kirchhoff plates with transfer learning
- A deep learning improved numerical method for the simulation of rogue waves of nonlinear Schrödinger equation
- Computing Lyapunov functions using deep neural networks
- Gradient convergence of deep learning-based numerical methods for BSDEs
- A compressed lattice Boltzmann method based on ConvLSTM and resnet
- Numerical solution of the parametric diffusion equation by deep neural networks
- Image segmentation method for an illumination highlight region of interior design effects based on the partial differential equation
- Numerical simulation of Volterra-Fredholm integral equations using least squares support vector regression
- Approximation rates for neural networks with encodable weights in smoothness spaces
- Enhanced fifth order WENO shock-capturing schemes with deep learning
- Urban planning image feature enhancement and simulation based on partial differential equation method
- Optimally weighted loss functions for solving PDEs with neural networks
- Reinforcement learning and stochastic optimisation
- Exact imposition of boundary conditions with distance functions in physics-informed deep neural networks
- Multi-fidelity regression using artificial neural networks: efficient approximation of parameter-dependent output quantities
- HiDeNN-TD: reduced-order hierarchical deep learning neural networks
- Normalization effects on shallow neural networks and related asymptotic expansions
- A representative volume element network (RVE-net) for accelerating RVE analysis, microscale material identification, and defect characterization
- Data-driven peakon and periodic peakon solutions and parameter discovery of some nonlinear dispersive equations via deep learning
- Data-driven discoveries of Bäcklund transformations and soliton evolution equations via deep neural network learning schemes
- INN: interfaced neural networks as an accessible meshless approach for solving interface PDE problems
- An application of the splitting-up method for the computation of a neural network representation for the solution for the filtering equations
- On surrogate learning for linear stability assessment of Navier-Stokes equations with stochastic viscosity.
- Deep combinatorial optimisation for optimal stopping time problems: application to swing options pricing.
- A priori and a posteriori error estimates for the deep Ritz method applied to the Laplace and Stokes problem
- Efficient coupled deep neural networks for the time-dependent coupled Stokes-Darcy problems
- Deep learning schemes for parabolic nonlocal integro-differential equations
- Lookback option pricing under the double Heston model using a deep learning algorithm
- Committor functions via tensor networks
- A non-gradient method for solving elliptic partial differential equations with deep neural networks
- Solving parametric partial differential equations with deep rectified quadratic unit neural networks
- Solving partial differential equation based on extreme learning machine
- Discrete gradient flow approximations of high dimensional evolution partial differential equations via deep neural networks
- Uniform convergence guarantees for the deep Ritz method for nonlinear problems
- On stability and regularization for data-driven solution of parabolic inverse source problems
- Neural eikonal solver: improving accuracy of physics-informed neural networks for solving eikonal equation in case of caustics
- A deep domain decomposition method based on Fourier features
- Two neural-network-based methods for solving elliptic obstacle problems
- Deep learning Gauss-Manin connections
- Latent-space time evolution of non-intrusive reduced-order models using Gaussian process emulation
- DNN expression rate analysis of high-dimensional PDEs: application to option pricing
- A theoretical analysis of deep neural networks and parametric PDEs
- GINNs: graph-informed neural networks for multiscale physics
- A method for representing periodic functions and enforcing exactly periodic boundary conditions with deep neural networks
- An efficient neural network method with plane wave activation functions for solving Helmholtz equation
- Adaptive two-layer ReLU neural network. I: Best least-squares approximation
- Adaptive two-layer ReLU neural network. II: Ritz approximation to elliptic PDEs
- Using deep learning to extend the range of air pollution monitoring and forecasting
- Meta-learning pseudo-differential operators with deep neural networks
- DPM: a deep learning PDE augmentation method with application to large-eddy simulation
- On some neural network architectures that can represent viscosity solutions of certain high dimensional Hamilton-Jacobi partial differential equations
- A probabilistic generative model for semi-supervised training of coarse-grained surrogates and enforcing physical constraints through virtual observables
- Data-driven discovery of coarse-grained equations
- Structure probing neural network deflation
- Symplectic neural networks in Taylor series form for Hamiltonian systems
- Deep least-squares methods: an unsupervised learning-based numerical method for solving elliptic PDEs
- Trend to equilibrium for the kinetic Fokker-Planck equation via the neural network approach
- A derivative-free method for solving elliptic partial differential equations with deep neural networks
- Int-Deep: a deep learning initialized iterative method for nonlinear problems
- Deep learning of free boundary and Stefan problems
- PhyGeoNet: physics-informed geometry-adaptive convolutional neural networks for solving parameterized steady-state PDEs on irregular domain
- PFNN: a penalty-free neural network method for solving a class of second-order boundary-value problems on complex geometries
- Solving the linear transport equation by a deep neural network approach
- ISALT: inference-based schemes adaptive to large time-stepping for locally Lipschitz ergodic systems
- Solving inverse-PDE problems with physics-aware neural networks
- SelectNet: self-paced learning for high-dimensional partial differential equations
- DeepM\&Mnet: inferring the electroconvection multiphysics fields based on operator approximation by neural networks
- Weak form theory-guided neural network (TgNN-wf) for deep learning of subsurface single- and two-phase flow
- On obtaining sparse semantic solutions for inverse problems, control, and neural network training
- Least-squares ReLU neural network (LSNN) method for linear advection-reaction equation
- Using neural networks to accelerate the solution of the Boltzmann equation
- Mutual information for explainable deep learning of multiscale systems
- A modified batch intrinsic plasticity method for pre-training the random coefficients of extreme learning machines
- SPINN: sparse, physics-based, and partially interpretable neural networks for PDEs
- Solving and learning nonlinear PDEs with Gaussian processes
- Physics-informed neural networks for solving forward and inverse flow problems via the Boltzmann-BGK formulation
- Mesh-Conv: convolution operator with mesh resolution independence for flow field modeling
- MIM: a deep mixed residual method for solving high-order partial differential equations
- A new efficient approximation scheme for solving high-dimensional semilinear PDEs: control variate method for deep BSDE solver
- Meta-mgnet: meta multigrid networks for solving parameterized partial differential equations
- Self-adaptive deep neural network: numerical approximation to functions and PDEs
- A fast and accurate physics-informed neural network reduced order model with shallow masked autoencoder
- DeLISA: deep learning based iteration scheme approximation for solving PDEs
- Learning time-dependent PDEs with a linear and nonlinear separate convolutional neural network
- Adaptive deep density approximation for Fokker-Planck equations
- When and why PINNs fail to train: a neural tangent kernel perspective
- On quadrature rules for solving partial differential equations using neural networks
- Neural networks enforcing physical symmetries in nonlinear dynamical lattices: the case example of the Ablowitz-Ladik model
This page was built for publication: DGM: a deep learning algorithm for solving partial differential equations
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2002333)
