Bistable Gradient Networks II: Storage Capacity and Behaviour Near Saturation
arXiv:cond-mat/0205465 · doi:10.1103/PhysRevE.67.016119
Abstract
We examine numerically the storage capacity and the behaviour near saturation of an attractor neural network consisting of bistable elements with an adjustable coupling strength, the Bistable Gradient Network (BGN). For strong coupling, we find evidence of a first-order "memory blackout" phase transition as in the Hopfield network. For weak coupling, on the other hand, there is no evidence of such a transition and memorized patterns can be stable even at high levels of loading. The enhanced storage capacity comes, however, at the cost of imperfect retrieval of the patterns from corrupted versions.
15 pages, 12 eps figures. Submitted to Phys. Rev. E. Sequel to cond-mat/0203568