Phase transitions in scale-free neural networks: Departure from the standard mean-field universality class
Por:
Aldana M., Larralde H.
Publicada:
1 ene 2004
Resumen:
The phase transitions from ordered to disordered states that occur in a family of neural network models with noise were considered. The existence and nature of the phase transition were strongly determined by the input probability distribution P(k). A stability analysis of the evolution equation predicted that the order parameter ultimately decays exponentially to its stationary value except at the critical noise, where a power law decay ensues. The existence of a phase transition even for the values of the scale-free exponent in the interval (1.5,2).
Filiaciones:
Aldana M.:
Centro de Ciencias Físicas, UNAM, Apdo. P. 48-3, Codigo Postal 62251, Cuernavaca, Morelos, Mexico
Consortium Americas Interdisc. Sci., University of New Mexico, Albuquerque, NM, United States
Larralde H.:
Centro de Ciencias Físicas, UNAM, Apdo. P. 48-3, Codigo Postal 62251, Cuernavaca, Morelos, Mexico
|