Application of Cellular Automates in Some Models of Artificial Intelligence
Oleksandr S. Makarenko, Viacheslav Osaulenko · 2018
The problem of modeling many natural phenomena is too complex and requires interaction on different scales. Here we show that one of the possible and convenient way to approach this problem with hierarchical cellular automata. The peculiarity of this approach is that the state of the cell is modeled by another cellular automaton with as many recursive layers as needed. We make an example of how the “Smart cube”, as we called it, can be used for simulating such a complex system as the brain. Different levels from neurotransmitters to large-scale neural population interaction can be represented into one model of cellular automata. Furthermore, this allows to experiment with different learning rules and incorporate many interesting mechanisms like anticipation.