Adaptive thresholds for neural networks with synaptic noise

被引:9
作者
Bolle, D. [1 ]
Heylen, R. [1 ]
机构
[1] Katholieke Univ Leuven, Inst Theoret Phys, B-3001 Louvain, Belgium
关键词
layered networks; fully connected networks; adaptive threshold; retrieval dynamics; autonomous functioning;
D O I
10.1142/S012906570700110X
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
The inclusion of a macroscopic adaptive threshold is studied for the retrieval dynamics of both layered feedforward and fully connected neural network models with synaptic noise. These two types of architectures require a different method to be solved numerically. In both cases it is shown that, if the threshold is chosen appropriately as a function of the cross-talk noise and of the activity of the stored patterns, adapting itself automatically in the course of the recall process, an autonomous functioning of the network is guaranteed. This self-control mechanism considerably improves the quality of retrieval, in particular the storage capacity, the basins of attraction and the mutual information content.
引用
收藏
页码:241 / 252
页数:12
相关论文
共 26 条
  • [1] STATISTICAL NEURODYNAMICS OF ASSOCIATIVE MEMORY
    AMARI, S
    MAGINU, K
    [J]. NEURAL NETWORKS, 1988, 1 (01) : 63 - 73
  • [2] NEURAL THEORY OF ASSOCIATION AND CONCEPT-FORMATION
    AMARI, SI
    [J]. BIOLOGICAL CYBERNETICS, 1977, 26 (03) : 175 - 185
  • [3] INFORMATION-STORAGE IN NEURAL NETWORKS WITH LOW-LEVELS OF ACTIVITY
    AMIT, DJ
    GUTFREUND, H
    SOMPOLINSKY, H
    [J]. PHYSICAL REVIEW A, 1987, 35 (05) : 2293 - 2303
  • [4] Blahut R. E., 1990, PRINCIPLES PRACTICE
  • [5] Mutual information and self-control of a fully-connected low-activity neural network
    Bollé, D
    Carreta, DD
    [J]. PHYSICA A, 2000, 286 (3-4): : 401 - 416
  • [6] Mutual information of sparsely coded associative memory with self-control and ternary neurons
    Bollé, D
    Dominguez, DRC
    Amari, S
    [J]. NEURAL NETWORKS, 2000, 13 (4-5) : 455 - 462
  • [7] Thresholds in layered neural networks with variable activity
    Bollé, D
    Massolo, G
    [J]. JOURNAL OF PHYSICS A-MATHEMATICAL AND GENERAL, 2000, 33 (14): : 2597 - 2609
  • [8] Bollé D, 2004, IEEE IJCNN, P3195
  • [9] The signal-to-noise analysis of the Little-Hopfield model revisited
    Bollé, D
    Blanco, JB
    Verbeiren, T
    [J]. JOURNAL OF PHYSICS A-MATHEMATICAL AND GENERAL, 2004, 37 (06): : 1951 - 1969
  • [10] BOLLE D, 2004, ADV CONDENSED MATTER, P321