Storage capacity of a diluted neural network with Ising couplings
M. -C. Bouten, Andrzej Komoda, R. Serneels · Journal of Physics A Mathematical and General · 1990
The capacity for storing random patterns in a diluted neural network is determined following the method of Gardner (1988). The non-zero coupling coefficients are restricted to take on the Ising values +or-1. Different degrees of dilution are considered. The maximum values alpha c =1.17 for the storage capacity is obtained when each neuron is connected to 63% of the other neurons.