STATISTICAL-MECHANICS OF UNSUPERVISED LEARNING

被引：28

作者：

BIEHL, M

MIETZNER, A

机构：

[1] Physikalisches Institut, Julius-Maximilians-Universität, Würzburg, D-9707h, Am Hubland

来源：

EUROPHYSICS LETTERS | 1993年 / 24卷 / 05期

关键词：

D O I：

10.1209/0295-5075/24/5/017

中图分类号：

O4 [物理学];

学科分类号：

0702 ;

摘要：

We study two different unsupervised learning strategies for a single-layer perceptron. The environment provides a set of unclassified training examples, which belong to two different classes, depending on their overlap with an N-dimensional concept vector. By means of a statistical-mechanics analysis, using the replica method, we investigate how well the perceptron infers the unknown structure from the input data.

引用

页码：421 / 426

页数：6

共 20 条

[1] ANLAUF J, UNPUB
[2] THE ADATRON - AN ADAPTIVE PERCEPTRON ALGORITHM
ANLAUF, JK
BIEHL, M
[J]. EUROPHYSICS LETTERS, 1989, 10 (07): : 687 - 692
[3] [Anonymous], 1991, INTRO THEORY NEURAL, DOI DOI 10.1201/9780429499661
[4] SCALING LAWS IN LEARNING OF CLASSIFICATION TASKS
BARKAI, N
SEUNG, HS
SOMPOLINSKY, H
[J]. PHYSICAL REVIEW LETTERS, 1993, 70 (20) : 3167 - 3170
[5] Duda R. O., 1973, PATTERN CLASSIFICATI, V3
[6] ENGELNMULLGES G, 1993, NUMERISCHE MATH INGE
[7] THE SPACE OF INTERACTIONS IN NEURAL NETWORK MODELS
GARDNER, E
[J]. JOURNAL OF PHYSICS A-MATHEMATICAL AND GENERAL, 1988, 21 (01): : 257 - 270
[8] KINZEL W, 1990, STATISTICAL MECHANIC
[9] LEARNING ALGORITHMS WITH OPTIMAL STABILITY IN NEURAL NETWORKS
KRAUTH, W
MEZARD, M
[J]. JOURNAL OF PHYSICS A-MATHEMATICAL AND GENERAL, 1987, 20 (11): : L745 - L752
[10] MEZARD M, 1987, SPIN GLASS THEORY BE

← 1 2 →