Nonparametric uncertainty quantification for stochastic gradient flows
From MaRDI portal
diffusion mapsnonlinear filteringnonlinear responsenonparametric uncertainty quantificationstatistical predictionstochastic gradient systems
Prediction theory (aspects of stochastic processes) (60G25) Signal detection and filtering (aspects of stochastic processes) (60G35) Stochastic ordinary differential equations (aspects of stochastic analysis) (60H10) Diffusion processes (60J60) Inference from stochastic processes and prediction (62M20) Stochastic methods (Fokker-Planck, Langevin, etc.) applied to problems in time-dependent statistical mechanics (82C31) Filtering in stochastic control theory (93E11)
Abstract: This paper presents a nonparametric statistical modeling method for quantifying uncertainty in stochastic gradient systems with isotropic diffusion. The central idea is to apply the diffusion maps algorithm to a training data set to produce a stochastic matrix whose generator is a discrete approximation to the backward Kolmogorov operator of the underlying dynamics. The eigenvectors of this stochastic matrix, which we will refer to as the diffusion coordinates, are discrete approximations to the eigenfunctions of the Kolmogorov operator and form an orthonormal basis for functions defined on the data set. Using this basis, we consider the projection of three uncertainty quantification (UQ) problems (prediction, filtering, and response) into the diffusion coordinates. In these coordinates, the nonlinear prediction and response problems reduce to solving systems of infinite-dimensional linear ordinary differential equations. Similarly, the continuous-time nonlinear filtering problem reduces to solving a system of infinite-dimensional linear stochastic differential equations. Solving the UQ problems then reduces to solving the corresponding truncated linear systems in finitely many diffusion coordinates. By solving these systems we give a model-free algorithm for UQ on gradient flow systems with isotropic diffusion. We numerically verify these algorithms on a 1-dimensional linear gradient flow system where the analytic solutions of the UQ problems are known. We also apply the algorithm to a chaotically forced nonlinear gradient flow system which is known to be well approximated as a stochastically forced gradient flow.
Recommendations
- A nonparametric belief propagation method for uncertainty quantification with applications to flow in random porous media
- A backward SDE method for uncertainty quantification in deep learning
- Uncertainty quantification for stochastic approximation limits using chaos expansion
- Path-Based Divergence Rates and Lagrangian Uncertainty in Stochastic Flows
- A stochastic gradient method for a class of nonlinear PDE-constrained optimal control problems under uncertainty
Cites work
- A powerful numerical technique solving Zakai equation for nonlinear filtering
- Data-driven modeling and scientific computation. Methods for complex systems and big data
- Diffusion maps
- Diffusion Maps, Reduction Coordinates, and Low Dimensional Representation of Stochastic Systems
- Diffusion maps, spectral clustering and reaction coordinates of dynamical systems
- Dynamically orthogonal field equations for continuous stochastic dynamical systems
- Extracting macroscopic dynamics: model problems and algorithms
- Filtering complex turbulent systems.
- Forecasting turbulent modes with nonparametric diffusion models: learning from noisy data
- Fundamentals of stochastic filtering
- Information Theory and Stochastics for Multiscale Nonlinear Systems
- Laplacian Eigenmaps for Dimensionality Reduction and Data Representation
- Local kernels and the geometric structure of data
- Nonlinear Laplacian spectral analysis for time series with intermittency and low-frequency variability
- On Estimation of a Probability Density Function and Mode
- On Galerkin approximations for the Zakai equation with diffusive and point process observations
- On Locally Adaptive Density Estimation
- Principal component analysis.
- Reconstruction of diffusion using spectral data from time series
- Remarks on Some Nonparametric Estimates of a Density Function
- Spectral Methods for Uncertainty Quantification
- Statistically accurate low-order models for uncertainty quantification in turbulent dynamical systems
- Uncertainty Quantification and Polynomial Chaos Techniques in Computational Fluid Dynamics
- Uncertainty quantification. Theory, implementation, and applications
- Variable bandwidth diffusion kernels
- Variable kernel density estimation
Cited in
(8)- Reduced-space Gaussian process regression for data-driven probabilistic forecast of chaotic dynamical systems
- Data-driven efficient solvers for Langevin dynamics on manifold in high dimensions
- Forecasting turbulent modes with nonparametric diffusion models: learning from noisy data
- Coarse-gradient Langevin algorithms for dynamic data integration and uncertainty quantification
- A nonparametric belief propagation method for uncertainty quantification with applications to flow in random porous media
- On the Koopman operator of algorithms
- Data-driven probability density forecast for stochastic dynamical systems
- Semiparametric modeling: correcting low-dimensional model error in parametric models
This page was built for publication: Nonparametric uncertainty quantification for stochastic gradient flows
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2945163)
