Sampling conditioned diffusions via Pathspace Projected Monte Carlo

Fuente: arXiv
Saved in:
Bibliographic Details
Main Author: Grafke, Tobias
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912439474323456
author Grafke, Tobias
author_facet Grafke, Tobias
contents We present an algorithm to sample stochastic differential equations conditioned on rather general constraints, including integral constraints, endpoint constraints, and stochastic integral constraints. The algorithm is a pathspace Metropolis-adjusted manifold sampling scheme, which samples stochastic paths on the submanifold of realizations that adhere to the conditioning constraint. We demonstrate the effectiveness of the algorithm by sampling a dynamical condensation phase transition, conditioning a random walk on a fixed Levy stochastic area, conditioning a stochastic nonlinear wave equation on high amplitude waves, and sampling a stochastic partial differential equation model of turbulent pipe flow conditioned on relaminarization events.
format Preprint
id arxiv_https___arxiv_org_abs_2506_15743
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Sampling conditioned diffusions via Pathspace Projected Monte Carlo
Grafke, Tobias
Machine Learning
Probability
Statistics Theory
We present an algorithm to sample stochastic differential equations conditioned on rather general constraints, including integral constraints, endpoint constraints, and stochastic integral constraints. The algorithm is a pathspace Metropolis-adjusted manifold sampling scheme, which samples stochastic paths on the submanifold of realizations that adhere to the conditioning constraint. We demonstrate the effectiveness of the algorithm by sampling a dynamical condensation phase transition, conditioning a random walk on a fixed Levy stochastic area, conditioning a stochastic nonlinear wave equation on high amplitude waves, and sampling a stochastic partial differential equation model of turbulent pipe flow conditioned on relaminarization events.
title Sampling conditioned diffusions via Pathspace Projected Monte Carlo
topic Machine Learning
Probability
Statistics Theory
url https://arxiv.org/abs/2506.15743