Saved in:
Bibliographic Details
Main Authors: Garg, Jhanvi, Balasubramanian, Krishna, Zhou, Quan
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2505.15059
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912624714711040
author Garg, Jhanvi
Balasubramanian, Krishna
Zhou, Quan
author_facet Garg, Jhanvi
Balasubramanian, Krishna
Zhou, Quan
contents Simulated tempering is a widely used strategy for sampling from multimodal distributions. In this paper, we consider simulated tempering combined with an arbitrary local Markov chain Monte Carlo sampler and present a new decomposition theorem that provides a lower bound on the restricted spectral gap of the algorithm for sampling from mixture distributions. By working with the restricted spectral gap, the applicability of our results is extended to broader settings such as when the usual spectral gap is difficult to bound or becomes degenerate. We demonstrate the application of our theoretical results by analyzing simulated tempering combined with random walk Metropolis--Hastings for sampling from mixtures of Gaussian distributions. Our complexity bound scales polynomially with the separation between modes, logarithmically with $1/\varepsilon$, where $\varepsilon$ denotes the target accuracy in total variation distance, and exponentially with the dimension $d$.
format Preprint
id arxiv_https___arxiv_org_abs_2505_15059
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Restricted Spectral Gap Decomposition for Simulated Tempering Targeting Mixture Distributions
Garg, Jhanvi
Balasubramanian, Krishna
Zhou, Quan
Statistics Theory
Probability
Computation
Machine Learning
Simulated tempering is a widely used strategy for sampling from multimodal distributions. In this paper, we consider simulated tempering combined with an arbitrary local Markov chain Monte Carlo sampler and present a new decomposition theorem that provides a lower bound on the restricted spectral gap of the algorithm for sampling from mixture distributions. By working with the restricted spectral gap, the applicability of our results is extended to broader settings such as when the usual spectral gap is difficult to bound or becomes degenerate. We demonstrate the application of our theoretical results by analyzing simulated tempering combined with random walk Metropolis--Hastings for sampling from mixtures of Gaussian distributions. Our complexity bound scales polynomially with the separation between modes, logarithmically with $1/\varepsilon$, where $\varepsilon$ denotes the target accuracy in total variation distance, and exponentially with the dimension $d$.
title Restricted Spectral Gap Decomposition for Simulated Tempering Targeting Mixture Distributions
topic Statistics Theory
Probability
Computation
Machine Learning
url https://arxiv.org/abs/2505.15059