Equivariant nonstationary source separation

被引：41

作者：

Choi, S ^{[1
]}

Cichocki, A ^{[1
]}

Amari, S ^{[1
]}

机构：

[1] Pohang Univ Sci & Technol, Dept Comp Sci & Engn, Nam Gu, Pohang 790784, South Korea

来源：

NEURAL NETWORKS | 2002年 / 15卷 / 01期

关键词：

blind source separation; decorrelation; independent component analysis; natural gradient; nonstationarity;

D O I：

10.1016/S0893-6080(01)00137-X

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Most of source separation methods focus on stationary sources, so higher-order statistics is necessary for successful separation, unless sources are temporally correlated. For nonstationary sources, however, it was shown [Neural Networks 8 (1995) 4111 that source separation could be achieved by second-order decorrelation. In this paper, we consider the cost function proposed by Matsuoka et al. [Neural Networks 8 (1995) 4111 and derive natural gradient learning algorithms for both fully connected recurrent network and feedforward network. Since our algorithms employ the natural gradient method, they possess the equivariant property and find a steepest descent direction unlike the algorithm [Neural Networks 8 (1995) 411]. We also show that our algorithms are always locally stable, regardless of probability distributions of nonstationary sources. (C) 2002 Elsevier Science Ltd. All rights reserved.

引用

页码：121 / 130

页数：10

共 50 条

[21] Nonstationary Source Separation Using Sequential and Variational Bayesian Learning
Chien, Jen-Tzung
Hsieh, Hsin-Lung
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, 2013, 24 (05) : 681 - 694
[22] A comparison of decorrelation criteria for the blind source separation of nonstationary signals
Fancourt, C
Parra, L
SAM2002: IEEE SENSOR ARRAY AND MULTICHANNEL SIGNAL PROCESSING WORKSHOP PROCEEDINGS, 2002, : 418 - 421
[23] TENSOR BASED SINGULAR SPECTRUM ANALYSIS FOR NONSTATIONARY SOURCE SEPARATION
Kouchaki, Samaneh
Sanei, Saeid
2013 IEEE INTERNATIONAL WORKSHOP ON MACHINE LEARNING FOR SIGNAL PROCESSING (MLSP), 2013,
[24] NONSTATIONARY AND TEMPORALLY CORRELATED SOURCE SEPARATION USING GAUSSIAN PROCESS
Hsieh, Hsin-Lung
Chien, Jen-Tzung
2011 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, 2011, : 2120 - 2123
[25] Adaptive Least-Squares method for blind source separation with an equivariant property
School of Electronic Engineering, Xidian Univ., Xi'an 710071, China
不详
Xi'an Dianzi Keji Daxue Xuebao, 2008, 6 (999-1002):
[26] A new variable step-size equivariant adaptive source separation algorithm
Xic, Xiaofu
Shi, Qingyan
Wu, Renbiao
2007 ASIA-PACIFIC CONFERENCE ON COMMUNICATIONS, 2007, : 479 - 482
[27] Blind source separation of nonstationary convolutively mixed signals in the subband domain
Russell, I
Xi, JT
Mertins, A
Chicharo, J
2004 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, VOL V, PROCEEDINGS: DESIGN AND IMPLEMENTATION OF SIGNAL PROCESSING SYSTEMS INDUSTRY TECHNOLOGY TRACKS MACHINE LEARNING FOR SIGNAL PROCESSING MULTIMEDIA SIGNAL PROCESSING SIGNAL PROCESSING FOR EDUCATION, 2004, : 481 - 484
[28] BLIND SOURCE SEPARATION FOR NONSTATIONARY TENSOR-VALUED TIME SERIES
Virta, J.
Nordhausen, K.
2017 IEEE 27TH INTERNATIONAL WORKSHOP ON MACHINE LEARNING FOR SIGNAL PROCESSING, 2017,
[29] Natural gradient learning for second-order nonstationary source separation
Choi, S
Cichocki, A
Amari, S
PROCEEDING OF THE 2002 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, VOLS 1-3, 2002, : 654 - 658
[30] A virtual instrument for efficient blind-source separation of nonstationary signals
Milanovic, Zeljka
Saulig, Nicoletta
Sucic, Victor
2016 INTERNATIONAL MULTIDISCIPLINARY CONFERENCE ON COMPUTER AND ENERGY SCIENCE (SPLITECH), 2016, : 169 - 172

← 1 2 3 4 5 →