A new type of neural computation

Abstract

A new type of neural computation scheme is proposed, in which the numerical values are represented as the number of firing neurons of neural clusters. The scheme is quite classical in the sense that it is essentially based on McCulloch-Pitts model and Hebbian rule, but it is drastically new in the sense that it is free from any symbolism. The computational capability of this neural shceme is extensively discussed. © 2008 Springer London.