Regularization with a pruning prior

被引：26

作者：

Goutte, C

Hansen, LK

机构：

[1] Department of Mathematical Modelling, Technical University of Denmark

来源：

NEURAL NETWORKS | 1997年 / 10卷 / 06期

关键词：

regularization; generalization method; Bayesian learning; evidence framework; Laplace prior; comparison with weight decay;

D O I：

10.1016/S0893-6080(97)00027-0

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

We investigate the use of a regularization prior and its pruning properties. We illustrate the behavior of this prior by conducting analyses both using a Bayesian framework and with the generalization method, on a simple toy problem. Results are thoroughly compared with those obtained with a traditional weight decay. (C) 1997 Elsevier Science Ltd.

引用

页码：1053 / 1059

页数：7

共 10 条

[1] FITTING AUTOREGRESSIVE MODELS FOR PREDICTION [J].

AKAIKE, H .

ANNALS OF THE INSTITUTE OF STATISTICAL MATHEMATICS, 1969, 21 (02) :243-&

[2]

[Anonymous], 1992, SMR

[3]

Buntine W. L., 1991, Complex Systems, V5, P603

[4]

GOUTTE C, 1996, NEURAL NETWORKS SIGN, V6

[5] PRUNING FROM ADAPTIVE REGULARIZATION [J].

HANSEN, LK ;

RASMUSSEN, CE .

NEURAL COMPUTATION, 1994, 6 (06) :1223-1232

[6]

JAYNES ET, 1985, MAXIMUM ENTROPY BAYE, P1

[7]

KROGH A, 1992, NIPS, V4

[8]

MACKAY DJC, 1992, NEURAL COMPUT, V4, P415, DOI [10.1162/neco.1992.4.3.415, 10.1162/neco.1992.4.3.448]

[9] BAYESIAN REGULARIZATION AND PRUNING USING A LAPLACE PRIOR [J].

WILLIAMS, PM .

NEURAL COMPUTATION, 1995, 7 (01) :117-143

[10]

WOLPERT D, 1995, 1993 MAXIMUM ENTROPY

← 1 →