pith. sign in

arxiv: 1508.00947 · v2 · pith:EGEZCRFNnew · submitted 2015-08-05 · 🧮 math.ST · math.PR· stat.ME· stat.TH

MCMC-Based Inference in the Era of Big Data: A Fundamental Analysis of the Convergence Complexity of High-Dimensional Chains

classification 🧮 math.ST math.PRstat.MEstat.TH
keywords convergencemcmcsdemonstrategeometrichigh-dimensionalproblemsrateschains
0
0 comments X
read the original abstract

Markov chain Monte Carlo (MCMC) lies at the core of modern Bayesian methodology, much of which would be impossible without it. Thus, the convergence properties of MCMCs have received significant attention, and in particular, proving (geometric) ergodicity is of critical interest. Trust in the ability of MCMCs to sample from modern-day high-dimensional posteriors, however, has been limited by a widespread perception that these chains typically experience serious convergence problems. In this paper, we first demonstrate that contemporary methods for obtaining convergence rates have serious limitations when the dimension grows. We then propose a framework for rigorously establishing the convergence behavior of commonly used high-dimensional MCMCs. In particular, we demonstrate theoretically the precise nature and severity of the convergence problems of popular MCMCs when implemented in high dimensions, including phase transitions in the convergence rates in various $n$ and $p$ regimes, and a universality result across an entire spectrum of models. We also show that convergence problems effectively eliminate the apparent safeguard of geometric ergodicity. We then demonstrate theoretical principles by which MCMCs can be constructed and analyzed to yield bounded geometric convergence rates even as the dimension $p$ grows without bound. Additionally, we propose a diagnostic tool for establishing convergence.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Fast Mixing of Data Augmentation Algorithms: Bayesian Probit, Logit, and Lasso Regression

    math.ST 2024-12 unverdicted novelty 8.0

    The paper proves polynomial mixing time upper bounds for three data augmentation algorithms (ProbitDA, LogitDA, LassoDA) with explicit dependence on design matrix, prior, n, and d.