ON THE USE OF THE BINOMIAL MODEL OF SYNAPTIC TRANSMISSION IN FORMAL NEURAL NETWORKS

Due to a lack of neural realism, formal neural networks may be of limited use in representing the real nervous system (RNS). In this paper, we consider here the possibility of incorporating the binomial model of synaptic transmission in the definition of connection weights of formal neural networks. This may significantly improve the value of formal representations of two basic features of the RNS: plasticity and noise. On the other hand, the binomial model, in its simplest forms, suffers from a serious limitation: the lack of temporal dimension.