I researched into a neural network that consisted of a neuron which is the model of a cerebral nerve cell.
The neuron compares the amount of the input signal whose threshold value, when the input signal is large, is excited and will then output a signal. The prior state is called the ignition state.
The chaotic neural network which is added to the view of chaos is known as ``noncycle vibration which a within determinism system makes''.
This chaotic neural network was first made to
learn about unknown patterns within this experiment; in other words,
it is made to learn the wrong.
In the case of this research, the noise is put on the unknown
pattern.
It is then investigated whether the right pattern could be
recollected from there or not.
Moreover, it experimented with changes in value of the parameter 
as well.
When changing these two, we also discover the change in the number of study
successes and failures.
When there was a noise, the result with it better ``no to reset''
was obtained.
When there is a noise, it is easy to memorize the way with front information.
Moreover, When the number of noises is ten, and the number of times
of study was increased, it turns out that the number of study success
becomes fewer.
This neural network demonstrates power in a character and recognition of language. Unlike the conventional computer, this neural network can do flexible processing. Therefore, development will be expected in future.