Derivative-Informed Dimension Reduction of Stochastic Differential Equations: A Path-Space Sensitivity Approach

DOI10.48550/ARXIV.2610.33210MaRDI QIDQ7408843arXivFDO

Authors Han Cheng Lie, Youssef M. Marzouk, Joanna Zou, Youssef Marzouk

Publication date 23 October 2026

Published in arXiv

Tldr We develop an approach to dimension reduction of stochastic differential equations by identifying low-dimensional structure in the drift-to-statistical-observable map using path-space sensitivity operators constructed from Fréchet derivatives, enabling improved tractability of uncertainty quantification with certified error bounds.

Full work available at URL https://doi.org/10.48550/ARXIV.2610.33210

arXiv classification

math.NAmath.PRstat.CO

Cites work

• A computational framework for infinite-dimensional Bayesian inverse problems. I: The linearized case, with application to global seismic inversion
• A Recursive Polynomial Chaos Evolution Method for Stochastic Differential Equations
• Active subspaces. Emerging ideas for dimension reduction in parameter studies
• An explicit link between Gaussian fields and Gaussian Markov random fields: the stochastic partial differential equation approach
• An operator learning perspective on parameter-to-observable maps
• Automated solution of differential equations by the finite element method. The FEniCS book
• Bayesian non-linear statistical inverse problems
• Data-driven approximation of the Koopman generator: model reduction, system identification, and control
• Data-driven model reduction and transfer operator approximation
• Dimension reduction for derivative-informed operator learning: An analysis of approximation errors
• Efficient Rare Event Simulation by Optimal Nonequilibrium Forcing
• Efficient Rare Event Simulation by Optimal Nonequilibrium Forcing
• Error bounds of the invariant statistics in machine learning of ergodic Itô diffusions
• Gradient-based dimension reduction of multivariate vector-valued functions
• Koopman operator-based model reduction for switched-system control of PDEs
• Likelihood-informed dimension reduction for nonlinear inverse problems
• Multi-Fidelity Covariance Estimation in the Log-Euclidean Geometry
• Solving high-dimensional parabolic PDEs using the tensor train format
• Spectral properties of dynamical systems, model reduction and decompositions
• The Mathematical Theory of Finite Element Methods

This page was built for publication: Derivative-Informed Dimension Reduction of Stochastic Differential Equations: A Path-Space Sensitivity Approach

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7408843)