2007 International Symposium on Nonlinear Theory and its Applications
Relation between Recall Frequency of Stored Patterns and their Basin Size in Chaotic Neural Network
Kazuki HIROZAWA, Yuko OSANA,
PDF download (606.5KB)
In the associative memories composed of chaotic neurons, stored patterns can be recalled dynamically. In the dynamic associations of the chaotic neural network, some patterns are recalled frequently and the other patterns are not recalled or recalled only a few times. However, it is not clear what makes such di?erence on the recall frequency. In this paper, we examine the number of local minimums in the energy space and the basin size in the Hopfield network and the relation between the recall frequency of stored patterns and their basin size in the chaotic neural network.