Normalized natural gradient in independent component analysis

被引：18

作者：

Tang, Ying ^{[1
]}

Li, Jianping ^{[1
]}

机构：

[1] Univ Elect Sci & Technol China, Sch Engn & Comp Sci, Chengdu 610054, Sichuan, Peoples R China

来源：

SIGNAL PROCESSING | 2010年 / 90卷 / 09期

关键词：

Independent component analysis (ICA); Normalized natural gradient; Matrix exponential; Fully multiplicative iteration; STABILITY ANALYSIS; LEARNING ALGORITHMS; MATRIX; SEPARATION; INFOMAX; RULES;

D O I：

10.1016/j.sigpro.2010.03.015

中图分类号：

TM [电工技术]; TN [电子技术、通信技术];

学科分类号：

0808 ; 0809 ;

摘要：

This letter establishes the normalized natural gradient based algorithm in independent component analysis by employing simple matrix algebra. Moreover, another algorithm tightly related to the normalized natural gradient is introduced based on the truncated version of matrix exponentials, strictly restricting solutions in the space of non-singular matrices. Experimental results demonstrate the superior convergence of the proposed algorithms. (C) Elsevier B.V. All rights reserved.

引用

页码：2773 / 2777

页数：5

共 26 条

[1] Amari, 2002, ADAPTIVE BLIND SIGNA
[2] Amari S, 1996, ADV NEUR IN, V8, P757
[3] Nonholonomic orthogonal learning algorithms for blind source separation
Amari, S
Chen, TP
Cichocki, A
[J]. NEURAL COMPUTATION, 2000, 12 (06) : 1463 - 1484
[4] Adaptive method of realizing natural gradient learning for multilayer perceptrons
Amari, S
Park, H
Fukumizu, K
[J]. NEURAL COMPUTATION, 2000, 12 (06) : 1399 - 1409
[5] Stability analysis of learning algorithms for blind source separation
Amari, S
Chen, TP
Cichocki, A
[J]. NEURAL NETWORKS, 1997, 10 (08) : 1345 - 1351
[6] Natural gradient works efficiently in learning
Amari, S
[J]. NEURAL COMPUTATION, 1998, 10 (02) : 251 - 276
[7] Amari S, 1997, ADV NEUR IN, V9, P127
[8] [Anonymous], 2012, Differential-geometrical methods in statistics
[9] AN INFORMATION MAXIMIZATION APPROACH TO BLIND SEPARATION AND BLIND DECONVOLUTION
BELL, AJ
SEJNOWSKI, TJ
[J]. NEURAL COMPUTATION, 1995, 7 (06) : 1129 - 1159
[10] Infomax and maximum likelihood for blind source separation
Cardoso, JF
[J]. IEEE SIGNAL PROCESSING LETTERS, 1997, 4 (04) : 112 - 114

← 1 2 3 →