ИСТИНА |
Войти в систему Регистрация |
|
ИСТИНА ЦЭМИ РАН |
||
A new approach to simulation modeling of neural networks based on the use of stochastic cellular automata is suggested. The notion of two-dimensional Markov automata is introduced. The model of an excitable medium and the mechanism of inhibition are considered. Simple feedforward neural network, realizing the XOR function, is constructed and numerically analyzed.