Exhaustive Thermodynamical Analysis of Boolean Learning Networks

摘要: The learning and generalization capabilities exhibited by a recently introduced model of self-organizing Boolean networks are explained interpreted studying the thermodynamics training process that model. thermodynamical analysis shows occur as direct consequence second law thermodynamics. complexity solving problem, for given architecture, can be precisely defined in terms entropy changes is related to amount specialization present architecture. We speculate our conclusions valid, some extent, not only model, but also more general complex systems, including biological systems.