Positive and negative circuits in discrete neural networks

被引：56

作者：

Aracena, J ^{[1
]}

Demongeot, J

Goles, E

机构：

[1] Univ Chile, DIM, Santiago 1703, Chile

[2] Univ Chile, Ctr Modelamiento Matemat, Santiago 1703, Chile

[3] IAB Fac Med, UMR CNRS 5525, IMAG, TIMC, F-38706 La Tronche, France

来源：

IEEE TRANSACTIONS ON NEURAL NETWORKS | 2004年 / 15卷 / 01期

关键词：

discrete neural networks (DNNs); fixed points; graph; positive and negative circuits; positive feedback vertex sets;

D O I：

10.1109/TNN.2003.821555

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 [模式识别与智能系统]; 0812 [计算机科学与技术]; 0835 [软件工程]; 1405 [智能科学与技术];

摘要：

We study the relationships between the positive and negative circuits of the connection graph and the fixed points of discrete neural networks (DNNs). As main results, we give necessary conditions and sufficient conditions for the existence of fixed points in a DNN. Moreover, we exhibit an upper bound for the number of fixed points in terms of the structure and number of positive circuits in the connection graph. This allows the determination of the maximum capacity for storing vectors in DNNs as fixed points, depending on the architecture of the network.

引用

页码：77 / 83

页数：7

共 19 条

[1]

BERGER C, 1974, GRAPHES HYPERGRAPHES

[2]

Positive and negative feedback: Striking a balance between necessary antagonists [J].