A controlled chaotic neural network for storing and retrieving information
Alessandro Londei, C. Mazur · 2002
We propose a new neural network composed of identical chaotic oscillators mutually coupled in cellular way by time delay connections. This feature is motivated by biological observations where the cognitive state of a brain is shown by periodic synchronicity of neurons. We show how such connections are able to force a couple of chaotic systems on a periodic orbit and we use such information to set the network parameters. Finally the network's ability to retrieve a stored pattern and to decorrelate an ambiguous input is shown.