Abstract: Based on current research on applications of chaotic neurons network for information processing, a new auto-associative matrix is devised for artificial neural network composed of chaotic neurons. And then, a chaotic neurons network for associative memory is built up. From numerical simulation experiments, it has proved that its associative success rate and recall speed are improved by comparing with existing methods. Moveover, the associative recalling process of the network is theoretically analyzed in detail and explanations of improvement are given.

Key words: neural networks, chaotic map, associative memory