Title: A dependent stochastic process in Bayesian nonparametrics
Authors: Mame Diarra Fall - University of Orleans and CNRS (France) [presenting]
Abstract: Dirichlet processes (DP) and Dirichlet processes mixtures (DPM) have emerged as cornerstones in Bayesian nonparametric models. The former can be used as a prior on a probability mass function, while the latter is a suitable prior for a probability density function. The focus is on a special case of dependent Dirichlet process mixtures (DDPM). We show how this can be used to handle challenging problems in image reconstruction.