COGNITIVE AND BEHAVIORAL INTELLIGENT ARTIFICIAL LIFE

摘要

This paper presents an idea how to construct cognitive and behavioral intelligent artificial life. Anrnintelligent information processing system like human brain was developed for a mechanical systemrnwhich is able to both recognize and track a moving object in real time. This information system isrncalled the artificial brain constructed by a 32-bit microcomputer NEC PC-9801, the RICOHrnneurocomputer RN-2000 and their interface operating these two different type computers. However, thernfunction of the artificial brain is still primitive because the decisions on the structure, number of layers,rnnumber of neurons in layers in the artificial brain are still made by engineers from experiments in thernenvironment. In order to construct the cognitive and behavioral intelligent artificial life we have tornimbed the function of self-organization into the neural networks in its artificial brain. Thus it becomesrnpossible for the artificial life to learn autonomously from the interaction between the artificial life andrnthe environment. The emergence of behaviors from the artificial life is realized from the neural networkrnself-organized by learning the results of interactions between them. In this paper, the mechanism ofrnself-organization in the neural networks is analyzed from a point of view of realizing the emergence ofrnbehaviors in the cognitive and behavioral intelligent artificial life. Based on the analyses we try tornconstruct the artificial brain with the function of self-organization in its neural network. Also, we tryrnto design the intelligent artificial life that supports and substitutes human's work.