Equivariant nonstationary source separation

Seungjin Choi, Andrzej Cichocki, Shunichi Amari

Research output: Contribution to journalArticlepeer-review

39 Citations (Scopus)

Abstract

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) 411] 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) 411] 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.

Original languageEnglish
Pages (from-to)121-130
Number of pages10
JournalNeural Networks
Volume15
Issue number1
DOIs
Publication statusPublished - 2002
Externally publishedYes

Keywords

  • Blind source separation
  • Decorrelation
  • Independent component analysis
  • Natural gradient
  • Nonstationarity

Fingerprint

Dive into the research topics of 'Equivariant nonstationary source separation'. Together they form a unique fingerprint.

Cite this