AN ARTIFICIAL MODULAR NEURAL-NETWORK AND ITS BASIC DYNAMICAL CHARACTERISTICS

This work contains a proposition of an artificial modular neural netwo rk (MNN) in which every module network exchanges input/output informat ion with others simultaneously. It further studies the basic dynamical characteristics of this network through both computer simulations and analytical considerations. A notable feature of this model is that it has generic representation with regard to the number of composed modu les, network topologies, and classes of introduced interactions. The i nformation processing of the MNN is described as the minimization of a total-energy function that consists of partial-energy functions for m odules and their interactions, and the activity and weight dynamics ar e derived from the total-energy function under the Lyapunov stability condition. This concept was realized by Cross-Coupled Hopfield Nets (C CHN) that one of the authors proposed. In this paper, in order to inve stigate the basic dynamical properties of CCHN, we offer a representat ive model called Cross-Coupled Hopfield Nets with Local And Global Int eractions (CCHN-LAGI) to which two distinct classes of interactions - local and global interactions are introduced. Through a conventional t est for associative memories, it is confirmed that our energy-function -based approach gives us proper dynamics of CCHN-LAGI even if the netw orks have different modularity. We also discuss the contribution of a single interaction and the joint contribution of the two distinct inte ractions through the eigenvalue analysis of connection matrices.