RS MODEL FOR NEURAL NETWORKS

After a general discussion of the architecture and dynamics of neural networks, we present a new model Hamiltonian that considers interactions between pairs, triads, …, n-adics of neurons. The Hopfield model is a limit case when the patterns are uncorrelated and the load parameter, α, is low. We apply the usual statistical mechanics techniques to obtain the main properties of the new model. The capacity is greatly improved compared with previous models. We demonstrate that only states that overlap with just one memory are stable at low temperatures.

[1]  M. Usher,et al.  Capacities of multiconnected memory models , 1988 .

[2]  Sompolinsky,et al.  Information storage in neural networks with low levels of activity. , 1987, Physical review. A, General physics.

[3]  Kanter,et al.  Associative recall of memory without errors. , 1987, Physical review. A, General physics.

[4]  Sompolinsky,et al.  Storing infinite numbers of patterns in a spin-glass model of neural networks. , 1985, Physical review letters.

[5]  Abbott,et al.  Storage capacity of generalized networks. , 1987, Physical review. A, General physics.

[6]  D. Amit,et al.  Statistical mechanics of neural networks near saturation , 1987 .

[7]  Sompolinsky,et al.  Spin-glass models of neural networks. , 1985, Physical review. A, General physics.

[8]  J J Hopfield,et al.  Neural networks and physical systems with emergent collective computational abilities. , 1982, Proceedings of the National Academy of Sciences of the United States of America.

[9]  Jose Roberto Iglesias,et al.  An alternative model for neural networks , 1990 .