Neural parametric Fokker-Planck equation
From MaRDI portal
deep learningimplicit Euler schemeneural parametric Fokker-Planck equationnumerical analysisoptimal transporttransport information geometry
Methods of ordinary differential equations applied to PDEs (35A24) Fokker-Planck equations (35Q84) PDEs on graphs and networks (ramified or polygonal spaces) (35R02) PDEs with randomness, stochastic partial differential equations (35R60) Optimal transportation (49Q22) Applications of stochastic analysis (to PDEs, etc.) (60H30) Finite difference and finite volume methods for ordinary differential equations (65L12) Finite difference methods for initial value and initial-boundary value problems involving PDEs (65M06) Stability and convergence of numerical methods for initial value and initial-boundary value problems involving PDEs (65M12) Error bounds for initial value and initial-boundary value problems involving PDEs (65M15) Artificial neural networks and deep learning (68T07)
Abstract: In this paper, we develop and analyze numerical methods for high dimensional Fokker-Planck equations by leveraging generative models from deep learning. Our starting point is a formulation of the Fokker-Planck equation as a system of ordinary differential equations (ODEs) on finite-dimensional parameter space with the parameters inherited from generative models such as normalizing flows. We call such ODEs neural parametric Fokker-Planck equations. The fact that the Fokker-Planck equation can be viewed as the -Wasserstein gradient flow of Kullback-Leibler (KL) divergence allows us to derive the ODEs as the constrained -Wasserstein gradient flow of KL divergence on the set of probability densities generated by neural networks. For numerical computation, we design a variational semi-implicit scheme for the time discretization of the proposed ODE. Such an algorithm is sampling-based, which can readily handle the Fokker-Planck equations in higher dimensional spaces. Moreover, we also establish bounds for the asymptotic convergence analysis of the neural parametric Fokker-Planck equation as well as the error analysis for both the continuous and discrete versions. Several numerical examples are provided to illustrate the performance of the proposed algorithms and analysis.
Recommendations
- Adaptive deep density approximation for Fokker-Planck equations
- Parametric Fokker-Planck equation
- Solving Fokker-Planck equation using deep learning
- Solving time dependent Fokker-Planck equations via temporal normalizing flow
- Solving inverse stochastic problems from discrete particle observations using the Fokker-Planck equation and physics-informed neural networks
Cites work
- A blob method for diffusion
- A multiscale meshfree method for macroscopic approximations of interacting particle systems
- A practical difference scheme for Fokker-Planck equations
- A review of the mean field limits for Vlasov equations
- A spectral element method for fluid dynamics: Laminar flow in a channel expansion
- Affine natural proximal learning
- Asymptotic Dynamics of Attractive-Repulsive Swarms
- Deep learning-based numerical methods for high-dimensional parabolic partial differential equations and backward stochastic differential equations
- Discrete gradients for computational Bayesian inference
- Elliptic partial differential equations. Volume 1: Fredholm theory of elliptic problems in unbounded domains
- Error bounds for approximations with deep ReLU networks
- Exponential convergence of Langevin distributions and their discrete approximations
- Fokker-Planck particle systems for Bayesian inference: computational approaches
- Generalization of an inequality by Talagrand and links with the logarithmic Sobolev inequality
- Global-in-time weak measure solutions and finite-time aggregation for nonlocal interaction equations
- Gradient flows in metric spaces and in the space of probability measures
- scientific article; zbMATH DE number 3897897 (Why is no real title available?)
- scientific article; zbMATH DE number 50534 (Why is no real title available?)
- scientific article; zbMATH DE number 54145 (Why is no real title available?)
- scientific article; zbMATH DE number 646825 (Why is no real title available?)
- scientific article; zbMATH DE number 1909499 (Why is no real title available?)
- scientific article; zbMATH DE number 3894218 (Why is no real title available?)
- Information geometry
- Information geometry and its applications
- Logarithmic Sobolev inequalities and stochastic Ising models
- Low-dimensional reduced-order models for statistical response and uncertainty quantification: barotropic turbulence with topography
- Mean Field Analysis of Deep Neural Networks
- Molecular modeling and simulation. An interdisciplinary guide.
- Natural gradient via optimal transport
- Nonlinear approximation and (deep) ReLU networks
- Numerical solution of the Fokker-Planck equation by finite difference and finite element methods -- a comparative study
- On the Poisson equation and diffusion approximation. I
- Optimal approximation of piecewise smooth functions using deep ReLU neural networks
- Optimal Transport
- Parametric Fokker-Planck equation
- Partial differential equations and stochastic methods in molecular dynamics
- Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
- Primal dual methods for Wasserstein gradient flows
- Ricci curvature for parametric statistics via optimal transport
- Solution of Fokker-Planck equation by finite element and finite difference methods for nonlinear systems
- Solving for high-dimensional committor functions using artificial neural networks
- Solving high-dimensional Hamilton-Jacobi-Bellman PDEs using neural networks: perspectives from the theory of controlled diffusions and measures on path space
- Solving parametric PDE problems with artificial neural networks
- Some geometric calculations on Wasserstein space
- Stochastic processes and applications. Diffusion processes, the Fokker-Planck and Langevin equations
- The <scp>Data‐Driven</scp> Schrödinger Bridge
- The Brownian movement and stochastic equations
- The Density Manifold and Configuration Space Quantization
- The Fokker-Planck equation. Methods of solution and applications.
- THE GEOMETRY OF DISSIPATIVE EVOLUTION EQUATIONS: THE POROUS MEDIUM EQUATION
- The Variational Formulation of the Fokker--Planck Equation
- Theory of Maxima and the Method of Lagrange
- Vector field approximation using radial basis functions
- Wasserstein proximal of GANs
- Weak adversarial networks for high-dimensional partial differential equations
Cited in
(22)- Self-similar characteristics of neural networks based on Fokker-Planck equation
- Trend to equilibrium for the kinetic Fokker-Planck equation via the neural network approach
- Adaptive deep density approximation for Fokker-Planck equations
- Solving for high-dimensional committor functions using artificial neural networks
- Solving inverse stochastic problems from discrete particle observations using the Fokker-Planck equation and physics-informed neural networks
- Solving time dependent Fokker-Planck equations via temporal normalizing flow
- Neural network representation of the probability density function of diffusion processes
- Solving Fokker-Planck equation using deep learning
- Stationary Density Estimation of Itô Diffusions Using Deep Learning
- Adaptive deep density approximation for fractional Fokker-Planck equations
- Efficient Natural Gradient Descent Methods for Large-Scale PDE-Based Optimization Problems
- Fisher information and shape-morphing modes for solving the Fokker-Planck equation in higher dimensions
- Solving high-dimensional Fokker-Planck equation with functional hierarchical tensor
- Nonstandard finite difference schemes for linear and non-linear Fokker-Planck equations
- Deep JKO: time-implicit particle methods for general nonlinear gradient flows
- Applications of Fokker Planck equations in machine learning algorithms
- Lipschitz-regularized gradient flows and generative particle algorithms for high-dimensional scarce data
- Numerical analysis on neural network projected schemes for approximating one dimensional Wasserstein gradient flows
- Score-based physics-informed neural networks for high-dimensional Fokker-Planck equations
- Parameterized Wasserstein gradient flow
- Parameterized Wasserstein Hamiltonian flow
- A fast and accurate solver for the fractional Fokker-Planck equation with Dirac-delta initial conditions
This page was built for publication: Neural parametric Fokker-Planck equation
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5087103)
