G. Berthiau et al., LEARNING OF NEURAL NETWORKS APPROXIMATING CONTINUOUS-FUNCTIONS THROUGH CIRCUIT SIMULATOR SPICE-PAC DRIVEN BY SIMULATED ANNEALING, International journal of electronics, 76(3), 1994, pp. 437-441
Simulated annealing (SA) adapted to continuous variables is used to de
termine the synaptic coefficients of an analogue multilayer neural net
work, approximating any continuous function of one or several variable
s. The 'open' electrical simulator SPICE-PAC driven by SA produces a g
lobally optimal set of synaptic weights, in a reasonable time and with
out requiring heavy and inaccurate gradient computations. We illustrat
e and improve our weights tuning strategy through simple examples.