English

Mixing under monotone censoring

Probability 2013-12-03 v2 Combinatorics

Abstract

We initiate the study of mixing times of Markov chain under monotone censoring. Suppose we have some Markov Chain MM on a state space Ω\Omega with stationary distribution π\pi and a monotone set AΩA \subset \Omega. We consider the chain MM' which is the same as the chain MM started at some xAx \in A except that moves of MM of the form xyx \to y where xAx \in A and yAy \notin A are {\em censored} and replaced by the move xxx \to x. If MM is ergodic and AA is connected, the new chain converges to π\pi conditional on AA. In this paper we are interested in the mixing time of the chain MM' in terms of properties of MM and AA. Our results are based on new connections with the field of property testing. A number of open problems are presented.

Keywords

Cite

@article{arxiv.1311.5945,
  title  = {Mixing under monotone censoring},
  author = {Jian Ding and Elchanan Mossel},
  journal= {arXiv preprint arXiv:1311.5945},
  year   = {2013}
}

Comments

6 pages

R2 v1 2026-06-22T02:13:28.761Z