UCL Discovery
UCL home » Library Services » Electronic resources » UCL Discovery

Undercomplete Blind Subspace Deconvolution via Linear Prediction

Szabo, Z; Póczos, B; Lőrincz, A; (2007) Undercomplete Blind Subspace Deconvolution via Linear Prediction. In: Kok, JN and Koronacki, J and Mantaras, RL and Matwin, S and Mladenic, D and Skowron, A, (eds.) Machine Learning: ECML 2007. Proceedings of the 18th European Conference on Machine Learning, Warsaw, Poland, September 17-21, 2007. (pp. 740 - 747). Springer-Verlag, Berlin Heidelberg Green open access

[thumbnail of szabo07undercomplete_LPA.pdf]
Preview
PDF
szabo07undercomplete_LPA.pdf
Available under License : See the attached licence file.

Download (507kB)
[thumbnail of szabo07undercomplete_LPA_poster.pdf]
Preview
PDF
szabo07undercomplete_LPA_poster.pdf
Available under License : See the attached licence file.

Download (360kB)
[thumbnail of szabo07undercomplete_LPA_poster_highlight.pdf]
Preview
PDF
szabo07undercomplete_LPA_poster_highlight.pdf
Available under License : See the attached licence file.

Download (112kB)

Abstract

We present a novel solution technique for the blind subspace deconvolution (BSSD) problem, where temporal convolution of multidimensional hidden independent components is observed and the task is to uncover the hidden components using the observation only. We carry out this task for the undercomplete case (uBSSD): we reduce the original uBSSD task via linear prediction to independent subspace analysis (ISA), which we can solve. As it has been shown recently, applying temporal concatenation can also reduce uBSSD to ISA, but the associated ISA problem can easily become 'high dimensional' [1]. The new reduction method circumvents this dimensionality problem. We perform detailed studies on the efficiency of the proposed technique by means of numerical simulations. We have found several advantages: our method can achieve high quality estimations for smaller number of samples and it can cope with deeper temporal convolutions.

Type: Proceedings paper
Title: Undercomplete Blind Subspace Deconvolution via Linear Prediction
Event: European Conference on Machine Learning (ECML)
Location: Warsaw, Poland
Dates: 2007-09-17 - 2007-09-21
ISBN-13: 978-3-540-74957-8
Open access status: An open access version is available from UCL Discovery
DOI: 10.1007/978-3-540-74958-5_75
Publisher version: http://dx.doi.org/10.1007/978-3-540-74958-5_75
Language: English
Additional information: This is the authors' accepted version of this published article. The final publication is available at Springer via http://dx.doi.org/10.1007/978-3-540-74958-5_75
UCL classification: UCL
UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences
UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences > Faculty of Life Sciences
URI: https://discovery.ucl.ac.uk/id/eprint/1433229
Downloads since deposit
265Downloads
Download activity - last month
Download activity - last 12 months
Downloads by country - last 12 months

Archive Staff Only

View Item View Item