Showing 1–2 of 2 results for author: Stier, T D

Entropic transfer operators for stochastic systems

Authors: Hancheng Bi, Clément Sarrazin, Bernhard Schmitzer, Thilo D. Stier

Abstract: Dynamical systems can be analyzed via their Frobenius-Perron transfer operator and its estimation from data is an active field of research. Recently entropic transfer operators have been introduced to estimate the operator of deterministic systems. The approach is based on the regularizing properties of entropic optimal transport plans. In this article we generalize the method to stochastic and no… ▽ More Dynamical systems can be analyzed via their Frobenius-Perron transfer operator and its estimation from data is an active field of research. Recently entropic transfer operators have been introduced to estimate the operator of deterministic systems. The approach is based on the regularizing properties of entropic optimal transport plans. In this article we generalize the method to stochastic and non-stationary systems and give a quantitative convergence analysis of the empirical operator as the available samples increase. We introduce a way to extend the operator's eigenfunctions to previously unseen samples, such that they can be efficiently included into a spectral embedding. The practicality and numerical scalability of the method are demonstrated on a real-world fluid dynamics experiment. △ Less

Submitted 7 March, 2025; originally announced March 2025.

A Dimension-Independent Bound on the Wasserstein Contraction Rate of a Geodesic Random Walk on the Sphere

Authors: Philip Schär, Thilo D. Stier

Abstract: We theoretically analyze the properties of a geodesic random walk on the Euclidean $d$-sphere. Specifically, we prove that the random walk's transition kernel is Wasserstein contractive with a contraction rate which can be bounded from above independently of the dimension $d$. Our result is of particular interest due to its implications regarding the potential for dimension-independent performance… ▽ More We theoretically analyze the properties of a geodesic random walk on the Euclidean $d$-sphere. Specifically, we prove that the random walk's transition kernel is Wasserstein contractive with a contraction rate which can be bounded from above independently of the dimension $d$. Our result is of particular interest due to its implications regarding the potential for dimension-independent performance of both geodesic slice sampling on the sphere and Gibbsian polar slice sampling, which are Markov chain Monte Carlo methods for approximate sampling from essentially arbitrary distributions on their respective state spaces. △ Less

Submitted 16 June, 2024; v1 submitted 16 September, 2023; originally announced September 2023.

Comments: 11 pages, 1 figure

MSC Class: 60G50 (Primary) 60J05; 65C05 (Secondary)

Journal ref: Electronic Communications in Probability 29, article 62, 2024