Abstract
Several scenarios of mutually interacting neural networks were studied. Using perceptrons with well-known on-line training algorithms in the limit of infinite system size, exact equations of motion were derived for the dynamics of order parameters that describe the properties of the system. The information given in this study open the road to a general understanding of interacting adaptive systems with possible applications in biology, computer science, and economics.
Original language | English |
---|---|
Pages (from-to) | 2555-2565 |
Number of pages | 11 |
Journal | Physical Review E |
Volume | 62 |
Issue number | 2 B |
DOIs | |
State | Published - Aug 2000 |