Abstract:Among the performance-enhancing procedures for Hopfield-type networks that implement associative memory, Hebbian Unlearning (or dreaming) strikes for its simplicity and its clear biological interpretation. Yet, it does not easily lend itself to a clear analytical understanding. Here we show how Hebbian Unlearning can be effectively described in terms of a simple evolution of the spectrum and the eigenvectors of the coupling matrix. We use these ideas to design new dreaming algorithms that are effective from a computational point of view, and are analytically far more transparent than the original scheme.
Abstract:We study numerically the memory which forgets, introduced in 1986 by Parisi by bounding the synaptic strength, with a mechanism which avoid confusion, allows to remember the pattern learned more recently and has a physiologically very well defined meaning. We analyze a number of features of the learning at finite number of neurons and finite number of patterns. We discuss how the system behaves in the large but finite N limit. We analyze the basin of attraction of the patterns that have been learned, and we show that it is exponentially small in the age of the pattern. This is a clearly non physiological feature of the model.