Increased storage capacity for hierarchically structured information in a neural network of Ising type
L. B. Ioffe, Reiner Kühn, J. Leo van Hemmen · Journal of Physics A Mathematical and General · 1989
Modelling formal neurons by Ising spins, the authors describe a simple two-neuron interaction which allows optimal storage capacity for hierarchically structured information. This takes care both of the low-activity limit in simple, Hopfield-type, networks and of the correlations which occur inside the classes of information hierarchies.