DYNAMICS OF DILUTED ATTRACTOR NEURAL NETWORKS WITH DELAYS

We present an analysis of the attractors of a deterministic dynamics i n formal neural networks characterized by binary threshold units and a nonsymmetric connectivity. It is shown that in these networks a store d pattern or a pattern sequence is represented by a cloud of attractor s rather than by a single attractor. Dilution, which we describe by a power-law scaling, and delayed couplings are shown to equip this type of network with a dynamic behaviour that is interesting enough for sim plified models of biological motor systems.