Investigation of Associative Memory by Small World Hopfield Neural Network with Characteristics of Social Network

In the Hopfield Neural Network (HNN), each neuron is connected to every other neuron. Thereby, the HNN causes high cost to generate the network in terms of implementation. Small World Hopfield Neural Network (SWHNN) improves ability for transmission by introducing the shortcut connections into the sparse regular network. However the storage ability of the SWHNN decreases for associative memory because the SWHNN has sparse connectivity. In this study, we propose the SWHNN with characteristic of Local Bridge (SWHNN-LB) using “local bridge”. And we explore performance of the SWHNN-LB using associative memory.