1: \begin{abstract}

2:   This document presents methods to remove the \emph{initialization} or

3: \emph{burn-in} bias from Markov chain Monte Carlo (MCMC) estimates, with consequences on parallel computing,

4: convergence diagnostics and performance assessment. The document is written as an introduction to these methods for MCMC users. Some theoretical results are mentioned, but the focus is on the methodology.

5: \end{abstract}

6: