The aim is to generate minima on the energy surface, in order that at an input the network can converge to to them. As with many other network paradigms, a collection P of training patterns p ϵ (1,−1)|K|, which represents the minima of our energy surface. Here we do not look for the minima of an unidentified error function but describe minima on such type of a function. The purpose is the fact that the network shall automatically take the closest minimum when the input is presented. The training of a Hopfield network is done by training each and every training pattern exactly once.