Neural Network Model of Associative Memory: To Visualize Solutions in Weight Space

We apply some variants of evolutionary computations to the Hopfield model of associative memory. In the model, a number of patterns can be stored in the network as attractors if synaptic weights are determined appropriately. One of our goals of this study is to learn the number and distribution of these Solutions in weight space, which is still an open problem. To address this issue, we test a method to visualize Solutions in high-dimensional space in this paper