Replica-symmetry breaking is studied in fully connected neural networks with modified pseudo-inverse interactions. The interaction matrix has an intermediate form between the Hebb learning rule and the pseudo-inverse one. At low temperature there is a region of parameters where the replica-symmetric solution is stable while its entropy is negative. It indicates the existence of the alternative solution in which the replica symmetry is broken. A one-step replica-symmetry breaking solution is found and its properties are analyzed.
REPLICA-SYMMETRY BREAKING IN NEURAL NETWORKS / V. S., Dotsenko; Tirozzi, Benedetto. - In: PHYSICA. A. - ISSN 0378-4371. - 185:1-4(1992), pp. 385-394. (Intervento presentato al convegno INTERNATIONAL CONF ON COMPLEX SYSTEMS : FRACTALS, SPIN GLASSES AND NEURAL NETWORKS tenutosi a TRIESTE, ITALY nel JUL 02-06, 1991) [10.1016/0378-4371(92)90479-a].
REPLICA-SYMMETRY BREAKING IN NEURAL NETWORKS
TIROZZI, Benedetto
1992
Abstract
Replica-symmetry breaking is studied in fully connected neural networks with modified pseudo-inverse interactions. The interaction matrix has an intermediate form between the Hebb learning rule and the pseudo-inverse one. At low temperature there is a region of parameters where the replica-symmetric solution is stable while its entropy is negative. It indicates the existence of the alternative solution in which the replica symmetry is broken. A one-step replica-symmetry breaking solution is found and its properties are analyzed.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
