next up previous contents
Next: 目次

概要:

In recent years, neural network is used in many fields, such as image processing and pattern recognition.

In this study, dynamical recall is studied using a chaotic neural network. A chaotic neural network is learned with incremental learning. In addition, dynamical recall is the important part of search access.

First, when recall, signal was inputted from outside. As a result of the study, dynamical recall was improved by inputting the constant. In addition, it was more improved by inputting random values.

Next, the improvement of dynamical recall was tried with the method of improving learning state. A chaotic neural network was able to be learned more strongly by changing the learning frequency. However, dynamical recall was not improved.

Our future tasks are to input another type signal to a neural network, to change parameters of chaotic neuron and so on.





出口研究室へ

Toshinori DEGUCHI
2005年 4月 1日 金曜日 16時36分09秒 JST