Latency-based probabilistic information processing in a learning feedback hierarchy

Alexander Gepperth 1, 2
1 Flowers - Flowing Epigenetic Robots and Systems
Inria Bordeaux - Sud-Ouest, U2IS - Unité d'Informatique et d'Ingénierie des Systèmes
Abstract : In this article, we study a three-layer neural hierarchy composed of bi-directionally connected recurrent layers which is trained to perform a synthetic object recognition task. The main feature of this network is its ability to represent, transmit and fuse probabilistic information, and thus to take near-optimal decisions when inputs are contradictory, noisy or missing. This is achieved by a neural space-latency code which is a natural consequence of the simple recurrent dynamics in each layer. Furthermore, the network possesses a feedback mechanism that is compatible with the space-latency code by making use of the attractor properties of neural layers. We show that this feedback mechanism can resolve/correct ambiguities at lower levels. As the fusion of feedback information in each layer is achieved in a probabilistically coherent fashion, feedback only has an effect if low-level inputs are ambiguous.
Type de document :
Communication dans un congrès
International Joint Conference on Neural Networks (IJCNN), Jun 2014, Beijing, China. pp.3031 - 3037, 2014, 〈10.1109/IJCNN.2014.6889919〉
Liste complète des métadonnées

Littérature citée [29 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/hal-01098704
Contributeur : Alexander Gepperth <>
Soumis le : dimanche 28 décembre 2014 - 16:00:53
Dernière modification le : mardi 10 avril 2018 - 09:46:08
Document(s) archivé(s) le : dimanche 29 mars 2015 - 10:15:21

Fichier

gepperth2014latency.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

Collections

Citation

Alexander Gepperth. Latency-based probabilistic information processing in a learning feedback hierarchy. International Joint Conference on Neural Networks (IJCNN), Jun 2014, Beijing, China. pp.3031 - 3037, 2014, 〈10.1109/IJCNN.2014.6889919〉. 〈hal-01098704〉

Partager

Métriques

Consultations de la notice

118

Téléchargements de fichiers

65