- Date: 27 Feb. (Thu.)
- Place: W.W. 6th-floor, Colloquium Room
- Time: 16:30-18:00
- Speaker: Pierre Bras (École normale supérieure Paris)
- Title: Bayesian inference and Metropolis-Hastings algorithms (provisional title)
- Abstract:
Markov chains Monte Carlo (MCMC) methods are simulation methods for sampling from a probability distribution from which direct sampling is difficult, and are particulary used in bayesian learning. The Metropolis-Hastings algorithm is one of the most popular. I will present the algorithm, then prove convergence results, and present the adaptation of the algorithm to stochastic optimization problems. Please come and join us.