Open Access
Translator Disclaimer
June 2020 Function-Specific Mixing Times and Concentration Away from Equilibrium
Maxim Rabinovich, Aaditya Ramdas, Michael I. Jordan, Martin J. Wainwright
Bayesian Anal. 15(2): 505-532 (June 2020). DOI: 10.1214/19-BA1151


Slow mixing is the central hurdle is applications of Markov chains, especially those used for Monte Carlo approximations (MCMC). In the setting of Bayesian inference, it is often only of interest to estimate the stationary expectations of a small set of functions, and so the usual definition of mixing based on total variation convergence may be too conservative. Accordingly, we introduce function-specific analogs of mixing times and spectral gaps, and use them to prove Hoeffding-like function-specific concentration inequalities. These results show that it is possible for empirical expectations of functions to concentrate long before the underlying chain has mixed in the classical sense, and we show that the concentration rates we achieve are optimal up to constants. We use our techniques to derive confidence intervals that are sharper than those implied by both classical Markov-chain Hoeffding bounds and Berry-Esseen-corrected central limit theorem (CLT) bounds. For applications that require testing, rather than point estimation, we show similar improvements over recent sequential testing results for MCMC. We conclude by applying our framework to real-data examples of MCMC, providing evidence that our theory is both accurate and relevant to practice.


Download Citation

Maxim Rabinovich. Aaditya Ramdas. Michael I. Jordan. Martin J. Wainwright. "Function-Specific Mixing Times and Concentration Away from Equilibrium." Bayesian Anal. 15 (2) 505 - 532, June 2020.


Published: June 2020
First available in Project Euclid: 13 June 2019

MathSciNet: MR4078723
Digital Object Identifier: 10.1214/19-BA1151

Primary: 60J10
Secondary: 62M02 , 62M05

Keywords: Concentration inequalities , confidence intervals , Markov chain Monte Carlo , Markov chains , Probability , sequential testing , statistics


Vol.15 • No. 2 • June 2020
Back to Top