LEARNING IN A 2-LAYER NEURAL NETWORK OF EDGE DETECTORS

被引：16

作者：

SOMPOLINSKY, H ^{[1
]}

TISHBY, N ^{[1
]}

机构：

[1] HEBREW UNIV JERUSALEM,RACAH INST PHYS,IL-91904 JERUSALEM,ISRAEL

来源：

EUROPHYSICS LETTERS | 1990年 / 13卷 / 06期

关键词：

D O I：

10.1209/0295-5075/13/6/016

中图分类号：

O4 [物理学];

学科分类号：

0702 ;

摘要：

Learning from examples to count domains in one-dimensional patterns is studied. Increasing the number of examples used for training a network to perform the task is equivalent to the annealing of a one-dimensional Ising model. The generalization error falls off exponentially with the number of examples per weight. The related contiguity problem, where the network discriminates between patterns with small and large number of domains, exhibits a first-order phase transition to perfect generalization at all temperatures. Monte Carlo simulations of both models are in very good agreement with the theoretical predictions. © 1990 IOP Publishing Ltd.

引用

页码：567 / 572

页数：6

共 14 条

[1] THE FERRIER LECTURE, 1980 - CRITICAL LIMITING FACTORS IN THE DESIGN OF THE EYE AND VISUAL-CORTEX [J].

BARLOW, HB .

PROCEEDINGS OF THE ROYAL SOCIETY SERIES B-BIOLOGICAL SCIENCES, 1981, 212 (1186) :1-34

[2] What Size Net Gives Valid Generalization? [J].

Baum, Eric B. ;

Haussler, David .

NEURAL COMPUTATION, 1989, 1 (01) :151-160

[3] 3 UNFINISHED WORKS ON THE OPTIMAL STORAGE CAPACITY OF NETWORKS [J].

GARDNER, E ;

DERRIDA, B .

JOURNAL OF PHYSICS A-MATHEMATICAL AND GENERAL, 1989, 22 (12) :1983-1994

[4] 1ST-ORDER TRANSITION TO PERFECT GENERALIZATION IN A NEURAL NETWORK WITH BINARY SYNAPSES [J].

GYORGYI, G .

PHYSICAL REVIEW A, 1990, 41 (12) :7097-7100

[5]

GYORGYI G, 1990, NEURAL NETWORKS SPIN

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

HANSEL, D ;

SOMPOLINSKY, H .

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

[7]

Horn B., 1986, ROBOT VISION, DOI DOI 10.1137/1030032

[8]

Krauth W., 1988, Complex Systems, V2, P387

[9]

LEVIN E, 1990, P IEEE

[10]

Mezard M., 1987, SPIN GLASS THEORY IN, V9

← 1 2 →