Forgetful Memories
- 1 April 1988
- journal article
- Published by IOP Publishing in Europhysics Letters
- Vol. 5 (7), 663-668
- https://doi.org/10.1209/0295-5075/5/7/016
Abstract
Iterative learning schemes for fully connected neural nets are solved analytically. The key to the solution is a Markov chain representation of the nonlinear, iterative, encoding procedure, whose asymptotics can be determined exactly. Forgetting is an intrinsic property of the network, induced by a first-order transition of the retrieval quality as a function of the storage ancestry. The storage capacity, which can be obtained analytically, is extensive. Numerical simulations are found to be in excellent agreement with our results.Keywords
This publication has 13 references indexed in Scilit:
- Nonlinear neural networks. II. Information processingJournal of Statistical Physics, 1988
- Nonlinear neural networks. I. General theoryJournal of Statistical Physics, 1988
- Nonlinear neural networks near saturationPhysical Review A, 1987
- Statistical mechanics of neural networks near saturationAnnals of Physics, 1987
- Nonlinear Neural NetworksPhysical Review Letters, 1986
- A memory which forgetsJournal of Physics A: General Physics, 1986
- Spin glass model of learning by selection.Proceedings of the National Academy of Sciences, 1986
- Spin-glass models of neural networksPhysical Review A, 1985
- Neurons with graded response have collective computational properties like those of two-state neurons.Proceedings of the National Academy of Sciences, 1984
- Neural networks and physical systems with emergent collective computational abilities.Proceedings of the National Academy of Sciences, 1982