A Hierarchical Classification of First-Order Recurrent Neural Networks

Abstract : We provide a refined hierarchical classification of first-order recurrent neural networks made up of McCulloch and Pitts cells. The classification is achieved by first proving the equivalence between the expressive powers of such neural networks and Muller automata, and then translating the Wadge classification theory from the automata-theoretic to the neural network context. The obtained hierarchical classification of neural networks consists of a decidable pre-well ordering of width 2 and height !!, and a decidability procedure of this hierarchy is provided. Notably, this classification is shown to be intimately related to the attractive properties of the networks, and hence provides a new refined measurement of the computational power of these networks in terms of their attractive behaviours.
Type de document :
Communication dans un congrès
Dediu AH; Fernau H; Martin-Vide C. 4th International Conference on Language and Automata Theory and Applications, May 2011, Trier, Germany. 6031, pp.142-153, 2011, 〈10.1007/978-3-642-13089-2_12〉
Liste complète des métadonnées

Littérature citée [19 références]  Voir  Masquer  Télécharger

http://www.hal.inserm.fr/inserm-00624083
Contributeur : Jean-Paul Issartel <>
Soumis le : jeudi 15 septembre 2011 - 16:49:07
Dernière modification le : jeudi 1 février 2018 - 01:11:39
Document(s) archivé(s) le : vendredi 16 décembre 2011 - 02:27:02

Fichier

Cabessa_2010_A-hierarchical_Au...
Fichiers produits par l'(les) auteur(s)

Identifiants

Collections

U836 | UGA

Citation

Jérémie Cabessa, Alessandro Villa. A Hierarchical Classification of First-Order Recurrent Neural Networks. Dediu AH; Fernau H; Martin-Vide C. 4th International Conference on Language and Automata Theory and Applications, May 2011, Trier, Germany. 6031, pp.142-153, 2011, 〈10.1007/978-3-642-13089-2_12〉. 〈inserm-00624083〉

Partager

Métriques

Consultations de la notice

178

Téléchargements de fichiers

117