Shin Ishii
Eliminating spurious memories using
a network of chaotic elements
Abstract:A Globally Coupled Map (GCM) model is a network of chaotic
elements that are globally coupled with each other. We have already
proposed an associative memory system based on the GCM, which has
a better ability than the Hopfield network. This success is obtained
through the mechanism that a network state can escape from spurious
memories with its chaotic dynamics. Therefore, our approach is not
to reduce spurious memories, rather, it is to escape from them.
In this paper, we propose a modified associative memory system,
in which spurious memories are noticeably reduced. This is achieved
by modifying the chaotic dynamics of the system, and not by modifying
its learning rule. With this improvement, our system's memory
capacity and the basin volume are expanded in a great deal. Some
experimental results in comparison with those of a neural network
employing a nonmonotonic output function are also shown.