The double loop as a model of a learning neural system

Christophe Lecerf

Equipe de Cognition Artificielle, Université Paris8 Vincennes St-Denis

15, rue Catulienne, F-93526 St-Denis. France

cl AT mire.net

 

 

Abstract

This paper presents the double loop feedback model, which is used for unsupervised learning and signal coupling in an artificial neural network. We first consider physiological arguments suggesting that loops and double loops are widely spread in the exchange flows of the central nervous system. We then demonstrate that the double loop pattern works as a functional memory unit and we describe the main properties of a double loop resonator built with the classical Hebb's law learning principle in a feedforward basis. The implementation of such a network is discussed, followed by the experimental results from unsupervised learning experiments. In this model, we show how signal coupling results from superimposition of flows originating from the nervous system, the organism and the external environment together. Emergent effects of Hebb's law, both bottom-up from cells to network and top-down from network to cells are finally brought into discussion.

 

Keywords: Unsupervised learning, memory, double loop resonator, emergent effects, signal coupling.

 

Article publié dans les actes (Proceedings) de la multi-conférence SCI'98, vol. 1, pp 587-594, Orlando, July 12-16, 1998.

Reprints sur demande à cl AT mire.net