Controlling the chaotic dynamics of neural networks
Francisco Welington Lima, Tarik Hadzibeganovic
Last modified: 2010-04-17
Abstract
Roughly the past decade of research on chaos control in neural networks will be reviewed and discussed, with a focus on the most actively developing directions and applications. A particular attention will be given to a method using local control, i.e. pinnings acting only on a small subset of neuronal units. Pertinent simulation results employing this method will be compared against other competing approaches. As it will be demonstrated, there are many practical advantages in knowing how to take a system with a large number of degrees of freedom from a chaotic to a fixed or periodic behavior.