Learning the temporal evolution of multivariate densities via normalizing flows

Abstract

In this work, we propose a method to learn multivariate probability distributions using sample path data from stochastic differential equations. Specifically, we consider temporally evolving probability distributions (e.g., those produced by integrating local or nonlocal Fokker-Planck equations). We analyze this evolution through machine learning assisted construction of a time-dependent mapping that takes a reference distribution (say, a Gaussian) to each and every instance of our evolving distribution. If the reference distribution is the initial condition of a Fokker-Planck equation, what we learn is the time-T map of the corresponding solution. Specifically, the learned map is a multivariate normalizing flow that deforms the support of the reference density to the support of each and every density snapshot in time. We demonstrate that this approach can approximate probability density function evolutions in time from observed sampled data for systems driven by both Brownian and Lévy noise. We present examples with two- and three-dimensional, uni- and multimodal distributions to validate the method.

References in corpus (37)

• Discovering governing equations from data: Sparse identification of nonlinear dynamical systems
• A Data-Driven Approximation of the Koopman Operator: Extending Dynamic Mode Decomposition
• Variational Inference with Normalizing Flows
• Normalizing Flows: An Introduction and Review of Current Methods
• Data-driven discovery of coordinates and governing equations
• Deep learning-based numerical methods for high-dimensional parabolic partial differential equations and backward stochastic differential equations
• Density estimation using Real NVP
• Neural Ordinary Differential Equations
• Normalizing Flows for Probabilistic Modeling and Inference
• Data-driven approximation of the Koopman generator: Model reduction, system identification, and control
• Data-driven model reduction and transfer operator approximation
• Coarse Molecular Dynamics of a Peptide Fragment: Free Energy, Kinetics, and Long-Time Dynamics Computations
• Masked Autoregressive Flow for Density Estimation
• On the numerical approximation of the Perron-Frobenius and Koopman operator
• Scalable Gradients for Stochastic Differential Equations
• Neural Jump Stochastic Differential Equations
• Subspace dynamic mode decomposition for stochastic Koopman analysis
• Galerkin Approximation of Dynamical Quantities using Trajectory Data
• Coarse-Grained Kinetic Computations for Rare Events: Application to Micelle Formation
• Estimating long term behavior of flows without trajectory integration: the infinitesimal generator approach
• A Data-Driven Approach for Discovering Stochastic Dynamical Systems with Non-Gaussian Levy Noise
• Coarse-scale PDEs from fine-scale observations via machine learning
• An early warning indicator for atmospheric blocking events using transfer operators
• Physics-Informed Generative Adversarial Networks for Stochastic Differential Equations
• Neural Stochastic Differential Equations: Deep Latent Gaussian Models in the Diffusion Limit
• Stochastic Normalizing Flows
• Non-parametric Estimation of Stochastic Differential Equations with Sparse Gaussian Processes
• Detecting the maximum likelihood transition path from data of stochastic dynamic systems
• Discovering transition phenomena from data of stochastic dynamical systems with Lévy noise
• Learning the temporal evolution of multivariate densities via normalizing flows
• Operator Fitting for Parameter Estimation of Stochastic Differential Equations
• Towards Robust and Stable Deep Learning Algorithms for Forward Backward Stochastic Differential Equations
• Solving Inverse Stochastic Problems from Discrete Particle Observations Using the Fokker-Planck Equation and Physics-informed Neural Networks
• Temporal Normalizing Flows
• Bayesian Optical Flow with Uncertainty Quantification
• Equation-Free Dynamic Renormalization: Self-Similarity in Multidimensional Particle System Dynamics
• Optimal Transportation of Vector-Valued Measures

Cited by in corpus (6)

• Parsimonious Physics-Informed Random Projection Neural Networks for Initial-Value Problems of ODEs and index-1 DAEs
• Solving time dependent Fokker-Planck equations via temporal normalizing flow
• Learning the temporal evolution of multivariate densities via normalizing flows
• RandONet: Shallow-Networks with Random Projections for learning linear and nonlinear operators
• Generative modeling of time-dependent densities via optimal transport and projection pursuit
• Least Squares and Marginal Log-Likelihood Model Predictive Control using Normalizing Flows