Partenaires

CNRS
Logo tutelle
Logo tutelle
Logo tutelle


Rechercher

Sur ce site

Sur le Web du CNRS


Accueil du site > Séminaires > Probabilités Statistiques et réseaux de neurones > Dependency exploration

Vendredi 18 juin 2004, 10h-12h

Dependency exploration

Samuel Kaski (University of Helsinki)

Résumé : Canonical correlation analysis is a traditional linear method for finding what is common in two data sets. For normally distributed data it maximizes mutual information, and recently several groups have been working on other kinds of methods for the same problem, maximizing mutual information between data sets. I will discuss a new Bayesian principle for this task, and new clustering and component models for exploring the dependencies between data sets.

Dans la même rubrique :