The multilevel method is a class of algorithms that allowed the sampling probability distribution, these methods are based on (overdamped) Langevin approximation. The purpose is to sample an approximation of Bayesian estimator with a controlled cost, in particular with the dimension and the required precision. After an introduction of the statistical issues we will present multilevel methods for sampling a Gibbs measure and the complexity of these algorithms.
Multilevel methods for Gibbs sampling.
- Se connecter pour poster des commentaires
Nom de l'orateur
Maxime Egéa
Etablissement de l'orateur
LAREMA (Université d'Angers)
Date et heure de l'exposé
Lieu de l'exposé
Salle Éole