A stochastic population approach to the problem of stable recruitment hierarchies in spiking neural networks

Recruitment learning in hierarchies is an inherently unstable process (Valiant, 1994). This paper presents conditions on parameters for a feedforward network to ensure stable recruitment hierarchies. The parameter analysis is conducted by using a stochastic population approach to model a spiking neural network. The resulting network converges to activate a desired number of units at each stage of the hierarchy. The original recruitment method is modified first by increasing feedforward connection density for ensuring sufficient activation, then by incorporating noisy feedforward delays for separating inputs temporally, and finally by limiting excess activation via lateral inhibition. The task of activating Correspondence to: A. S. Maida. 

• Type: Journal
• Year: 2006
• Where: BC
• Authors: Cengiz Günay, Anthony S. Maida
• Keywords: BC 2006, Feedforward, Spiking Neural Network, Stable Recruitment Hierarchies