학술논문
Spin glass model for a neural network: associative memories stored with unequal weights.
Document Type
Journal
Author
Viana, L. (MEX-NAM-P) AMS Author Profile
Source
Subject
68 Computer science -- 68P Theory of data
68P20Information storage and retrieval
82Statistical mechanics, structure of matter -- 82A Statistical physics, structure of matter
82A57Disordered materials, including liquid crystals and spin glasses
92Biology and other natural sciences -- 92A Biology and behavioral sciences
92A09Physiology, biochemistry
68P20
82
82A57
92
92A09
Language
French
Abstract
Summary: ``A long-range spin glass model is considered in order to study a neural network in which a finite number $p$ of random patterns has been stored or memorised, each with unequal weight. This is done in order to reflect different degrees of training. The stability of these memories is studied and the possible existence of spurious stable states is investigated.''