Weights Convergence and Spikes Correlation in an Adaptive Neural Network Implemented on VLSI - Archive ouverte HAL
Communication Dans Un Congrès Année : 2008

Weights Convergence and Spikes Correlation in an Adaptive Neural Network Implemented on VLSI

Résumé

This paper presents simulations of a conductance-based neural network implemented on a mixed hardwaresoftware simulation system. Synaptic connections follow a bio-realistic STDP rule. Neurons receive correlated input noise patterns, resulting in a weights convergence in a confined range of conductance values. The correlation of the output spike trains depends on the correlation degree of the input patterns

Domaines

Electronique
Fichier non déposé

Dates et versions

hal-00288431 , version 1 (16-06-2008)

Identifiants

  • HAL Id : hal-00288431 , version 1

Citer

Adel Daouzli, Sylvain Saïghi, Laure Buhry, Yannick Bornat, Sylvie Renaud. Weights Convergence and Spikes Correlation in an Adaptive Neural Network Implemented on VLSI. Bio-inspired Systems and Signal Processing (BIOSIGNALS), Jan 2008, France. pp.286-291. ⟨hal-00288431⟩
141 Consultations
0 Téléchargements

Partager

More