Multilevel-Langevin pathwise average for Bayesian estimate
1 : Université dÁngers
Université d\'Angers, Université d'Angers, Université d\'Angers, Université d\'Angers
The study of Langevin diffusion is one way to estimate the posterior distribution of a Bayesian estimator. In this talk we will present the multilevel methods (pathwise average version) and their advantages in terms of complexity. In order to work with biological data, particular attention will be paid on the dependence in the dimension of the problem.