Independent subspace analysis on innovations

Barnabás Póczos, Bálint Takács, A. Lőrincz

Research output: Chapter in Book/Report/Conference proceedingConference contribution

6 Citations (Scopus)

Abstract

Independent subspace analysis (ISA) that deals with multidimensional independent sources, is a generalization of independent component analysis (ICA). However, all known ISA algorithms may become ineffective when the sources possess temporal structure. The innovation process instead of the original mixtures has been proposed to solve ICA problems with temporal dependencies. Here we show that this strategy can be applied to ISA as well. We demonstrate the idea on a mixture of 3D processes and also on a mixture of facial pictures used as two-dimensional deterministic sources. ISA on innovations was able to find the original subspaces, while plain ISA was not.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages698-706
Number of pages9
Volume3720 LNAI
Publication statusPublished - 2005
Event16th European Conference on Machine Learning, ECML 2005 - Porto, Portugal
Duration: Oct 3 2005Oct 7 2005

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume3720 LNAI
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other16th European Conference on Machine Learning, ECML 2005
CountryPortugal
CityPorto
Period10/3/0510/7/05

Fingerprint

Innovation
Subspace
Independent component analysis
Independent Component Analysis
Algorithm Analysis
Demonstrate

ASJC Scopus subject areas

  • Computer Science(all)
  • Biochemistry, Genetics and Molecular Biology(all)
  • Theoretical Computer Science

Cite this

Póczos, B., Takács, B., & Lőrincz, A. (2005). Independent subspace analysis on innovations. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 3720 LNAI, pp. 698-706). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 3720 LNAI).

Independent subspace analysis on innovations. / Póczos, Barnabás; Takács, Bálint; Lőrincz, A.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 3720 LNAI 2005. p. 698-706 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 3720 LNAI).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Póczos, B, Takács, B & Lőrincz, A 2005, Independent subspace analysis on innovations. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 3720 LNAI, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 3720 LNAI, pp. 698-706, 16th European Conference on Machine Learning, ECML 2005, Porto, Portugal, 10/3/05.
Póczos B, Takács B, Lőrincz A. Independent subspace analysis on innovations. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 3720 LNAI. 2005. p. 698-706. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
Póczos, Barnabás ; Takács, Bálint ; Lőrincz, A. / Independent subspace analysis on innovations. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 3720 LNAI 2005. pp. 698-706 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{097f8a6c12924f82a3d82d4d6c8fd38b,
title = "Independent subspace analysis on innovations",
abstract = "Independent subspace analysis (ISA) that deals with multidimensional independent sources, is a generalization of independent component analysis (ICA). However, all known ISA algorithms may become ineffective when the sources possess temporal structure. The innovation process instead of the original mixtures has been proposed to solve ICA problems with temporal dependencies. Here we show that this strategy can be applied to ISA as well. We demonstrate the idea on a mixture of 3D processes and also on a mixture of facial pictures used as two-dimensional deterministic sources. ISA on innovations was able to find the original subspaces, while plain ISA was not.",
author = "Barnab{\'a}s P{\'o}czos and B{\'a}lint Tak{\'a}cs and A. Lőrincz",
year = "2005",
language = "English",
isbn = "3540292438",
volume = "3720 LNAI",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "698--706",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Independent subspace analysis on innovations

AU - Póczos, Barnabás

AU - Takács, Bálint

AU - Lőrincz, A.

PY - 2005

Y1 - 2005

N2 - Independent subspace analysis (ISA) that deals with multidimensional independent sources, is a generalization of independent component analysis (ICA). However, all known ISA algorithms may become ineffective when the sources possess temporal structure. The innovation process instead of the original mixtures has been proposed to solve ICA problems with temporal dependencies. Here we show that this strategy can be applied to ISA as well. We demonstrate the idea on a mixture of 3D processes and also on a mixture of facial pictures used as two-dimensional deterministic sources. ISA on innovations was able to find the original subspaces, while plain ISA was not.

AB - Independent subspace analysis (ISA) that deals with multidimensional independent sources, is a generalization of independent component analysis (ICA). However, all known ISA algorithms may become ineffective when the sources possess temporal structure. The innovation process instead of the original mixtures has been proposed to solve ICA problems with temporal dependencies. Here we show that this strategy can be applied to ISA as well. We demonstrate the idea on a mixture of 3D processes and also on a mixture of facial pictures used as two-dimensional deterministic sources. ISA on innovations was able to find the original subspaces, while plain ISA was not.

UR - http://www.scopus.com/inward/record.url?scp=33646394047&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33646394047&partnerID=8YFLogxK

M3 - Conference contribution

SN - 3540292438

SN - 9783540292432

VL - 3720 LNAI

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 698

EP - 706

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

ER -