Jeudi 23 Mai


Retour à la vue des calendrier
Jeudi 23 Mai
Heure: 14:00 - 15:30
Lieu: Amphi Ampere
Résumé: Universal and best approximation through Bayesian ARTMAP
Description: Lucian SASU Bayesian ARTMAP is an adaptive resonance-theory based model, which was extended to perform regression. The network is shown to have both universal and best approximation properties. The network architecture and learning algorithm are presented, along with sketches of proofs and open questions.