Neural parametric Fokker-Planck equation

DOI10.1137/20M1344986zbMATH Open1493.65132arXiv2002.11309MaRDI QIDQ5087103FDOQ5087103

Authors: Shu Liu, Wuchen Li, Hongyuan Zha, Haomin Zhou 

Publication date: 8 July 2022

Published in: SIAM Journal on Numerical Analysis (Search for Journal in Brave)

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 $L^2$-Wasserstein gradient flow of Kullback-Leibler (KL) divergence allows us to derive the ODEs as the constrained $L^2$-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.

Full work available at URL: https://arxiv.org/abs/2002.11309

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

zbMATH Keywords

deep learningoptimal transportimplicit Euler schemenumerical analysisneural parametric Fokker-Planck equationtransport information geometry

Mathematics Subject Classification ID

Optimal transportation (49Q22) Artificial neural networks and deep learning (68T07) Fokker-Planck equations (35Q84) PDEs with randomness, stochastic partial differential equations (35R60) 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) Methods of ordinary differential equations applied to PDEs (35A24) PDEs on graphs and networks (ramified or polygonal spaces) (35R02)

Cites Work

• Exponential convergence of Langevin distributions and their discrete approximations
• Gradient flows in metric spaces and in the space of probability measures
• Optimal Transport
• Title not available (Why is that?)
• THE GEOMETRY OF DISSIPATIVE EVOLUTION EQUATIONS: THE POROUS MEDIUM EQUATION
• The Variational Formulation of the Fokker--Planck Equation
• Title not available (Why is that?)
• Title not available (Why is that?)
• Elliptic partial differential equations. Volume 1: Fredholm theory of elliptic problems in unbounded domains
• Global-in-time weak measure solutions and finite-time aggregation for nonlocal interaction equations
• The Fokker-Planck equation. Methods of solution and applications.
• A review of the mean field limits for Vlasov equations
• Generalization of an inequality by Talagrand and links with the logarithmic Sobolev inequality
• Logarithmic Sobolev inequalities and stochastic Ising models
• A spectral element method for fluid dynamics: Laminar flow in a channel expansion
• Molecular modeling and simulation. An interdisciplinary guide.
• Title not available (Why is that?)
• On the Poisson equation and diffusion approximation. I
• Title not available (Why is that?)
• Stochastic processes and applications. Diffusion processes, the Fokker-Planck and Langevin equations
• Some geometric calculations on Wasserstein space
• Partial differential equations and stochastic methods in molecular dynamics
• A practical difference scheme for Fokker-Planck equations
• Title not available (Why is that?)
• Parametric Fokker-Planck equation
• The Density Manifold and Configuration Space Quantization
• Solving parametric PDE problems with artificial neural networks
• The Brownian movement and stochastic equations
• Asymptotic Dynamics of Attractive-Repulsive Swarms
• Solution of Fokker-Planck equation by finite element and finite difference methods for nonlinear systems
• Vector field approximation using radial basis functions
• A multiscale meshfree method for macroscopic approximations of interacting particle systems
• Deep learning-based numerical methods for high-dimensional parabolic partial differential equations and backward stochastic differential equations
• Numerical solution of the Fokker-Planck equation by finite difference and finite element methods -- a comparative study
• Information geometry and its applications
• Optimal approximation of piecewise smooth functions using deep ReLU neural networks
• Error bounds for approximations with deep ReLU networks
• Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
• Solving for high-dimensional committor functions using artificial neural networks
• Weak adversarial networks for high-dimensional partial differential equations
• Low-dimensional reduced-order models for statistical response and uncertainty quantification: barotropic turbulence with topography
• Natural gradient via optimal transport
• Information geometry
• Solving high-dimensional Hamilton-Jacobi-Bellman PDEs using neural networks: perspectives from the theory of controlled diffusions and measures on path space
• Theory of Maxima and the Method of Lagrange
• Ricci curvature for parametric statistics via optimal transport
• A blob method for diffusion
• Affine natural proximal learning
• Nonlinear approximation and (deep) ReLU networks
• Fokker-Planck particle systems for Bayesian inference: computational approaches
• Discrete gradients for computational Bayesian inference
• Primal dual methods for Wasserstein gradient flows
• The <scp>Data‐Driven</scp> Schrödinger Bridge
• Wasserstein proximal of GANs
• Mean Field Analysis of Deep Neural Networks

Cited In (16)

• Solving for high-dimensional committor functions using artificial neural networks
• Adaptive deep density approximation for fractional Fokker-Planck equations
• Solving inverse stochastic problems from discrete particle observations using the Fokker-Planck equation and physics-informed neural networks
• Solving Fokker-Planck equation using deep learning
• Solving time dependent Fokker-Planck equations via temporal normalizing flow
• Adaptive deep density approximation for Fokker-Planck equations
• Deep JKO: time-implicit particle methods for general nonlinear gradient flows
• Stationary Density Estimation of Itô Diffusions Using Deep Learning
• Applications of Fokker Planck equations in machine learning algorithms
• Neural network representation of the probability density function of diffusion processes
• Solving high-dimensional Fokker-Planck equation with functional hierarchical tensor
• Self-similar characteristics of neural networks based on Fokker-Planck equation
• Lipschitz-regularized gradient flows and generative particle algorithms for high-dimensional scarce data
• Efficient Natural Gradient Descent Methods for Large-Scale PDE-Based Optimization Problems
• Nonstandard finite difference schemes for linear and non-linear Fokker-Planck equations
• Fisher information and shape-morphing modes for solving the Fokker-Planck equation in higher dimensions

Uses Software

• Adam
• Virtual library
• Wasserstein GAN
• BigGAN

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)