Am activation function adapting training algorithm for sigmoidal feedforward networks

被引：50

作者：

Chandra, P ^{[1
]}

Singh, Y ^{[1
]}

机构：

[1] GGS Indraprastha Univ, Sch Informat Technol, Delhi 110006, India

来源：

NEUROCOMPUTING | 2004年 / 61卷

关键词：

feedforward artificial neural networks; sigmoidal activation; squashing function; self-adaptation;

D O I：

10.1016/j.neucom.2004.04.001

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

The universal approximation results for sigmoidal feedforward artificial neural networks do not recommend a preferred activation function. In this paper a new activation function adapting algorithm is proposed for sigmoidal feedforward neural network training. The algorithm is compared against the backpropagation algorithm on four function approximation tasks. The results demonstrate that the proposed algorithm can be an order of magnitude faster than the backpropagation algorithm. (C) 2004 Elsevier B.V. All rights reserved.

引用

页码：429 / 437

页数：9

共 50 条

[41] Extended least squares based algorithm for training feedforward networks
Yam, JYF
Chow, TWS
[J]. IEEE TRANSACTIONS ON NEURAL NETWORKS, 1997, 8 (03): : 806 - 810
[42] Training feedforward neural networks: An algorithm giving improved generalization
Lee, CW
[J]. NEURAL NETWORKS, 1997, 10 (01) : 61 - 68
[43] A hybrid linear/nonlinear training algorithm for feedforward neural networks
McLoone, S
Brown, MD
Irwin, G
Lightbody, G
[J]. IEEE TRANSACTIONS ON NEURAL NETWORKS, 1998, 9 (04): : 669 - 684
[44] AN IMPROVED GENETIC ALGORITHM FOR TRAINING LAYERED FEEDFORWARD NEURAL NETWORKS
刘平
程翼宇
[J]. Journal of Zhejiang University-Science A(Applied Physics & Engineering), 2000, (03) : 85 - 89
[45] A Second Order Training Algorithm for Multilayer Feedforward Neural Networks
谭营
何振亚
邓超
[J]. Journal of Southeast University(English Edition), 1997, (01) : 32 - 36
[46] A New Weight Initialization Method for Sigmoidal Feedforward Artificial Neural Networks
Sodhi, Sartaj Singh
Chandra, Pravin
Tanwar, Sharad
[J]. PROCEEDINGS OF THE 2014 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS (IJCNN), 2014, : 291 - 298
[47] On PAC learning of functions with smoothness properties using feedforward sigmoidal networks
Oak Ridge Natl Lab, Oak Ridge, United States
[J]. Proc IEEE, 10 (1562-1568):
[48] On PAC learning of functions with smoothness properties using feedforward sigmoidal networks
Rao, NSV
Protopopescu, VA
[J]. PROCEEDINGS OF THE IEEE, 1996, 84 (10) : 1562 - 1569
[49] MULTILAYER FEEDFORWARD NETWORKS WITH A NONPOLYNOMIAL ACTIVATION FUNCTION CAN APPROXIMATE ANY FUNCTION
LESHNO, M
LIN, VY
PINKUS, A
SCHOCKEN, S
[J]. NEURAL NETWORKS, 1993, 6 (06) : 861 - 867
[50] A local linearized least squares algorithm for training feedforward neural networks
Stan, O
Kamen, E
[J]. IEEE TRANSACTIONS ON NEURAL NETWORKS, 2000, 11 (02): : 487 - 495

← 1 2 3 4 5 →