In recent years, neural network is playing an active part in various fields of ``Character recognition'' or ``Image processing''. The neural network is the nervous circuit in which the brain cells are linked together.
In 1943, the model of an artificial nervous cell was proposed by McCulloch and Pitts. It was thought of only fundamental part of nervous cell. However, it showed almighty nature by way of logic calculation system. Further, the concrete way was shown that learns artificially by Rosenblatt's perceptron.
However, in 1969, Minsky and Papert showed the extent of perceptron's capability. Soon the research of neural network became unpopular.
However, learning rule called back-propagation was proposed some time ago. Therefore, the neural network became popular again.
Recently, chaos data have been found in the activity of higher animals' cerebral. Consequently, ``chaos'' is added to the neuron model. The neural network composed of chaotic neurons (neurons considering chaos) is called ``chaotic neural network''.
In this research, I will make this neural network to learn cyclic pattern. Cyclic pattern is found in the sequence. Therefore, I will make up the middle layer with chaotic neuron or feedback. And I am aiming at finding the cyclic pattern under the neural network used back-propagation.