STATISTICAL PROPERTIES OF RANDOM ASYMMETRICAL NEURAL NETWORKS

A fully connected set of formal neurons that has not been subject to a ny training algorithm is studied. The thresholds and couplings are ran dom variables chosen from Gaussian distributions. The dynamics of the model can be studied within a mean field approximation. Our results sh ow a change of behaviour from a monostable to a bistable regime as the parameters are modified. A nonequilibrium potential is introduced to describe the model, and an analogy with a usual phase transition can b e drawn. This analogy suggests using a thermodynamical approach to stu dy the problem. Both the dynamical and the thermodynamical approaches give the same phase diagram for the model. The mean value of the thres hold controls the order of the phase transition between the monostable and bistable regimes.