Optimal transport between determinantal point processes and application to fast simulation

DOI10.15559/21-VMSTA180MaRDI QIDQ2239803zbMATH OpenOpenAlexFDO

Authors Laurent Decreusefond, Guillaume Moroz

Publication date 5 November 2021

Published in Modern Stochastics. Theory and Applications (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2011.00822

zbMATH Keywords

optimal transportpoint processWasserstein distance

Mathematics Subject Classification ID

Optimal transportation (49Q22) Point processes (e.g., Poisson, Cox, Hawkes processes) (60G55)

Abstract: We analyze several optimal transportation problems between de-terminantal point processes. We show how to estimate some of the distances between distributions of DPP they induce. We then apply these results to evaluate the accuracy of a new and fast DPP simulation algorithm. We can now simulate in a reasonable amount of time more than ten thousands points.

Review text

In this paper, the authors study several problems related to distances between the distributions of point processes. The main focus is on determinantal point processes, which can be described in terms of a reference measure \(m\) and a Hilbert-Schmidt operator on \(L^2(X,m)\) with a (symmetric) kernel \(K\), which is locally trace-class and its spectrum lies in \([0,1]\). The authors investigate properties of two distances between point processes which arise from optimal transport. The first one, called the Kantorovich-Rubinstein distance, is defined in terms of evaluation of two measures on a single \(1\)-Lipschitz functions; the second one is the classical \(2\)-Wasserstein distance, which is the minimal cost in the corresponding optimal transport problem. The results are then applied to a simulation of the Ginibre process.

Recommendations

• On simulation of continuous determinantal point processes
• A note on the simulation of the Ginibre point process
• Optimal transport: fast probabilistic approximation with exact solvers
• A fast approach to optimal transport: the back-and-forth method
• Wasserstein distance on configuration space

Cites work

• A note on the simulation of the Ginibre point process
• An Introduction to the Theory of Point Processes
• Compound Poisson process approximation.
• Determinantal point process models and statistical inference
• Determinantal point processes for machine learning
• Determinantal processes and independence
• Five lectures on optimal transportation: geometry, regularity and applications
• scientific article; zbMATH DE number 3866301 (Why is no real title available?)
• scientific article; zbMATH DE number 1909499 (Why is no real title available?)
• Monte Carlo with determinantal point processes
• On Strassen's theorem on stochastic domination
• Optimal Transport
• Quasi-invariance and integration by parts for determinantal and permanental processes
• Rademacher's theorem on configuration spaces and applications
• Random point fields associated with certain Fredholm determinants. I: Fermion, Poisson and Boson point processes.
• Real Analysis and Probability
• Sampling unitary ensembles
• The coincidence approach to stochastic point processes
• The palm measure and the Voronoi tessellation for the Ginibre process
• The polyanalytic Ginibre ensembles
• The Stein-Chen method, point processes and compensators
• Upper bounds on rubinstein distances on configuration spaces and applications
• Wasserstein distance on configuration space
• Zeros of Gaussian analytic functions and determinantal point processes

Cited in

(2)

• A note on the simulation of the Ginibre point process
• On simulation of continuous determinantal point processes

Describes a project that uses

Uses Software

• spatstat

This page was built for publication: Optimal transport between determinantal point processes and application to fast simulation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2239803)