IMPROVING A NETWORK GENERALIZATION ABILITY BY SELECTING EXAMPLES

被引：80

作者：

KINZEL, W ^{[1
]}

RUJAN, P ^{[1
]}

机构：

[1] UNIV OLDENBURG,FACHBEREICH PHYS 8,W-2900 OLDENBURG,GERMANY

来源：

EUROPHYSICS LETTERS | 1990年 / 13卷 / 05期

关键词：

D O I：

10.1209/0295-5075/13/5/016

中图分类号：

O4 [物理学];

学科分类号：

0702 ;

摘要：

We show that the generalization ability of simple Pereeptron-like devices is strongly enhanced by allowing the network itself to select the training examples. Analytic and numerical results are obtained for the Hebb and for the optimal Perceptron learning rule, respectively. © 1990 IOP Publishing Ltd.

引用

页码：473 / 477

页数：5

共 11 条

[1]

BIEHL M, 1990, THESISU GIESSEN

[2] THE SPACE OF INTERACTIONS IN NEURAL NETWORK MODELS [J].

GARDNER, E .

JOURNAL OF PHYSICS A-MATHEMATICAL AND GENERAL, 1988, 21 (01) :257-270

[3]

GYORGYI G, 1990, NEURAL NETWORKS SPIN

[4] LEARNING FROM EXAMPLES IN A SINGLE-LAYER NEURAL NETWORK [J].

HANSEL, D ;

SOMPOLINSKY, H .

EUROPHYSICS LETTERS, 1990, 11 (07) :687-692

[5]

KINZEL W, 1990, PHYSICS NEURAL NETWO

[6] LEARNING ALGORITHMS WITH OPTIMAL STABILITY IN NEURAL NETWORKS [J].

KRAUTH, W ;

MEZARD, M .

JOURNAL OF PHYSICS A-MATHEMATICAL AND GENERAL, 1987, 20 (11) :L745-L752

[7]

LAMBERT PF, 1969, METHODOLOGIES PATTER

[8] ON THE ABILITY OF THE OPTIMAL PERCEPTRON TO GENERALIZE [J].

OPPER, M ;

KINZEL, W ;

KLEINZ, J ;

NEHL, R .

JOURNAL OF PHYSICS A-MATHEMATICAL AND GENERAL, 1990, 23 (11) :L581-L586

[9]

RUJAN P, IN PRESS 11 P SITG C

[10] A THEORY OF THE LEARNABLE [J].

VALIANT, LG .

COMMUNICATIONS OF THE ACM, 1984, 27 (11) :1134-1142

← 1 2 →