Cette thèse porte sur le problème de l'inférence en grande dimension.Nous proposons différentes méthodes pour l'estimation de constantes de normalisation et l'échantillonnage de distributions complexes.Dans une première partie, nous développons plusieurs méthodes de Monte Carlo par chaînes de Markov.D'une part, nous développons une nouvelle approche pour des noyaux non-réversibles. D'autre part, nous proposons deux méthodes massivement parallélisables combinant des propriétés locales et globales des méthodes de Monte Carlo par chaînes de Markov, en particulier en se basant sur un nouvel estimateur de constante de normalisation.Nous appliquons ces méthodes à une tâche d'inférence approchée de distribution emph{a posteriori} de réseaux de ne...
Computational Bayesian statistics builds approximations to the posterior distribution either bysampl...
Monte Carlo methods are are an ubiquitous tool in modern statistics. Under the Bayesian paradigm, th...
Computational Bayesian statistics builds approximations to the posterior distribution either bysampl...
This thesis addresses the problem of high dimensional inference.We propose different methods for est...
This thesis addresses the problem of high dimensional inference.We propose different methods for est...
This thesis addresses the problem of high dimensional inference.We propose different methods for est...
National audienceBayesian posterior distributions can be numerically intractable, even by the means ...
Automatic decision making and pattern recognition under uncertainty are difficult tasks that are ubi...
Automatic decision making and pattern recognition under uncertainty are difficult tasks that are ubi...
We propose a new class of learning algorithms that combines variational approximation and Markov cha...
La statistique bayésienne computationnelle construit des approximations de la distribution a posteri...
Variational inference is one of the tools that now lies at the heart of the modern data analysis lif...
Computational Bayesian statistics builds approximations to the posterior distribution either bysampl...
Computational Bayesian statistics builds approximations to the posterior distribution either bysampl...
Computational Bayesian statistics builds approximations to the posterior distribution either bysampl...
Computational Bayesian statistics builds approximations to the posterior distribution either bysampl...
Monte Carlo methods are are an ubiquitous tool in modern statistics. Under the Bayesian paradigm, th...
Computational Bayesian statistics builds approximations to the posterior distribution either bysampl...
This thesis addresses the problem of high dimensional inference.We propose different methods for est...
This thesis addresses the problem of high dimensional inference.We propose different methods for est...
This thesis addresses the problem of high dimensional inference.We propose different methods for est...
National audienceBayesian posterior distributions can be numerically intractable, even by the means ...
Automatic decision making and pattern recognition under uncertainty are difficult tasks that are ubi...
Automatic decision making and pattern recognition under uncertainty are difficult tasks that are ubi...
We propose a new class of learning algorithms that combines variational approximation and Markov cha...
La statistique bayésienne computationnelle construit des approximations de la distribution a posteri...
Variational inference is one of the tools that now lies at the heart of the modern data analysis lif...
Computational Bayesian statistics builds approximations to the posterior distribution either bysampl...
Computational Bayesian statistics builds approximations to the posterior distribution either bysampl...
Computational Bayesian statistics builds approximations to the posterior distribution either bysampl...
Computational Bayesian statistics builds approximations to the posterior distribution either bysampl...
Monte Carlo methods are are an ubiquitous tool in modern statistics. Under the Bayesian paradigm, th...
Computational Bayesian statistics builds approximations to the posterior distribution either bysampl...