Abstract
A statistical mechanical approach for neural networks in which couplings are the slow dynamical variables is considered. The couplings are assumed to be confined in a restricted subspace near the Hebb-rule structure corresponding to quenched patterns. We study the situation when the couplings thermalize at a temperature different from that of the spin degrees of freedom, which makes it possible to treat the system in terms of the traditional replica approach with a finite number of replicas. The structure of the model is such that the effective evolution of the couplings tends to deepen the free energy minima corresponding to the learnt patterns. The phase diagram obtained exhibits a substantial increase of the retrieval region.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
