MCM’Ski / Adap’ski
MCM’ski is a conference organized jointly by ISBA and IMS. This year it takes place in Utah, in a ski resort (not bad, heh!). The theme is Markov Chain Monte Carlo methods and their applications. Adap’ski is a satellite workshop taking place just before, focusing on adaptive MCMC methods. Hence, these four days of talks and posters are focused on a specific topic of Computational Statistics, with a reasonably small audience, whereas the last conference I’ve been to, the Valencia Meeting, was a huge gathering of all kinds of (Bayesian) statisticians. I’m looking forward to this different format, hoping that it would lead to fruitful conversations with the world’s specialists on MCMC.
I will present a poster on an article I’ve talked about here, and available on arXiv here. It describes a method to make the most of each target density evaluation in a Independent Metropolis-Hastings algorithm, in order to reduce the variance of estimators based on the Markov chain generated by the algorithm.