Self-organized annealing in laterally inhibited neural networks shows power law decay

dc.creatorEmmert-Streib, Frank
dc.date2004-01-30
dc.date2005-07-15
dc.date.accessioned2026-07-07T02:56:12Z
dc.date.available2026-07-07T02:56:12Z
dc.descriptionIn this paper we present a method which assigns to each layer of a multilayer neural network, whose network dynamics is governed by a noisy winner-take-all mechanism, a neural temperature. This neural temperature is obtained by a least mean square fit of the probability distribution of the noisy winner-take-all mechanism to the distribution of a softmax mechanism, which has a well defined temperature as free parameter. We call this approximated temperature resulting from the optimization step the neural temperature. We apply this method to a multilayer neural network during learning the XOR-problem with a Hebb-like learning rule and show that after a transient the neural temperature decreases in each layer according to a power law. This indicates a self-organized annealing behavior induced by the learning rule itself instead of an external adjustment of a control parameter as in physically motivated optimization methods e.g. simulated annealing.
dc.description10 pages, 4 figures
dc.identifierhttps://arxiv.org/abs/cond-mat/0401633
dc.identifierhttp://arxiv.org/abs/cond-mat/0401633
dc.identifier.urihttp://salesiana.dossiersoluciones.com/handle/123456789/23228
dc.subjectDisordered Systems and Neural Networks
dc.titleSelf-organized annealing in laterally inhibited neural networks shows power law decay
dc.typetext

Files

Collections