Spectral telescope: Convergence rate bounds for random-scan Gibbs samplers based on a hierarchical structure
The Annals of Applied Probability · 01.02.2024 · 34 (1B) · 1319-1349
DOI: 10.1214/23-aap1992 ↗Kurz gefasst
Fachartikel in der Zeitschrift The Annals of Applied Probability, erschienen am 01.02.2024. Verfasst von Qian Qin und 1 weiteren Autor:innen. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 35 Literaturangaben.
Abstract (Originalsprache)
Random-scan Gibbs samplers possess a natural hierarchical structure. The structure connects Gibbs samplers targeting higher-dimensional distributions to those targeting lower-dimensional ones. This leads to a quasi-telescoping property of their spectral gaps. Based on this property, we derive three new bounds on the spectral gaps and convergence rates of Gibbs samplers on general domains. The three bounds relate a chain’s spectral gap to, respectively, the correlation structure of the target distribution, a class of random walk chains, and a collection of influence matrices. Notably, one of our results generalizes the technique of spectral independence, which has received considerable attention for its success on finite domains, to general state spaces. We illustrate our methods through a sampler targeting the uniform distribution on a corner of an n-cube.
Zitieren
APA
Qin, Q., & Wang, G. (2024). Spectral telescope: Convergence rate bounds for random-scan Gibbs samplers based on a hierarchical structure. The Annals of Applied Probability, 34(1B), 1319-1349. https://doi.org/10.1214/23-aap1992
Vancouver
Qin Q, Wang G. Spectral telescope: Convergence rate bounds for random-scan Gibbs samplers based on a hierarchical structure. The Annals of Applied Probability. 2024;34(1B):1319-1349. doi:10.1214/23-aap1992
BibTeX
@article{Qin2024,
author = {Qin, Qian and Wang, Guanyang},
title = {Spectral telescope: Convergence rate bounds for random-scan Gibbs samplers based on a hierarchical structure},
journal = {The Annals of Applied Probability},
year = {2024},
volume = {34},
number = {1B},
pages = {1319--1349},
doi = {10.1214/23-aap1992}
}
EUVIMED