Expansion of the Variational Garrote to a Multiple Measurement Vectors Model

Sofie Therese Hansen; Carsten Stahlhut; Lars Kai Hansen

doi:10.3233/978-1-61499-330-8-105

Expansion of the Variational Garrote to a Multiple Measurement Vectors Model

Sofie Therese Hansen, Carsten Stahlhut, Lars Kai Hansen

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

3 Downloads (Orbit)

Abstract

The recovery of sparse signals in underdetermined systems is the focus of this paper. We propose an expanded version of the Variational Garrote, originally presented by Kappen (2011), which can use multiple measurement vectors (MMVs) to further improve source retrieval performance. We show its superiority compared to the original formulation and demonstrate its ability to correctly estimate both the sources’ location and their magnitude. Finally evidence is given of the high performance of the proposed algorithm compared to other MMV models.

Original language	English
Title of host publication	Twelfth Scandinavian Conference on Artificial Intelligence
Editors	M. Jaeger
Publisher	IOS Press
Publication date	2013
Pages	105-114
ISBN (Print)	978-1-61499-329-2
ISBN (Electronic)	978-1-61499-330-8
DOIs	https://doi.org/10.3233/978-1-61499-330-8-105
Publication status	Published - 2013
Event	12th Scandinavian Conference on Artificial Intelligence (SCAI 2013) - Aalborg, Denmark Duration: 20 Nov 2013 → 22 Nov 2013 http://scai2013.cs.aau.dk/

Conference

Conference	12th Scandinavian Conference on Artificial Intelligence (SCAI 2013)
Country/Territory	Denmark
City	Aalborg
Period	20/11/2013 → 22/11/2013
Internet address	http://scai2013.cs.aau.dk/

Series	Frontiers in Artificial Intelligence and Applications
Volume	257
ISSN	0922-6389

Access to Document

10.3233/978-1-61499-330-8-105

OpenUrl availability

Full text

Cite this

@inproceedings{8b05d964a73546bf911f9a3c7fb897cb,

title = "Expansion of the Variational Garrote to a Multiple Measurement Vectors Model",

abstract = "The recovery of sparse signals in underdetermined systems is the focus of this paper. We propose an expanded version of the Variational Garrote, originally presented by Kappen (2011), which can use multiple measurement vectors (MMVs) to further improve source retrieval performance. We show its superiority compared to the original formulation and demonstrate its ability to correctly estimate both the sources{\textquoteright} location and their magnitude. Finally evidence is given of the high performance of the proposed algorithm compared to other MMV models.",

author = "Hansen, \{Sofie Therese\} and Carsten Stahlhut and Hansen, \{Lars Kai\}",

year = "2013",

doi = "10.3233/978-1-61499-330-8-105",

language = "English",

isbn = "978-1-61499-329-2",

series = "Frontiers in Artificial Intelligence and Applications",

publisher = "IOS Press",

pages = "105--114",

editor = "M. Jaeger",

booktitle = "Twelfth Scandinavian Conference on Artificial Intelligence",

note = "12th Scandinavian Conference on Artificial Intelligence (SCAI 2013) ; Conference date: 20-11-2013 Through 22-11-2013",

url = "http://scai2013.cs.aau.dk/",

}

Hansen, ST, Stahlhut, C & Hansen, LK 2013, Expansion of the Variational Garrote to a Multiple Measurement Vectors Model. in M Jaeger (ed.), Twelfth Scandinavian Conference on Artificial Intelligence. IOS Press, Frontiers in Artificial Intelligence and Applications, vol. 257, pp. 105-114, 12th Scandinavian Conference on Artificial Intelligence (SCAI 2013), Aalborg, Denmark, 20/11/2013. https://doi.org/10.3233/978-1-61499-330-8-105

Expansion of the Variational Garrote to a Multiple Measurement Vectors Model. / Hansen, Sofie Therese; Stahlhut, Carsten; Hansen, Lars Kai.
Twelfth Scandinavian Conference on Artificial Intelligence. ed. / M. Jaeger. IOS Press, 2013. p. 105-114 (Frontiers in Artificial Intelligence and Applications, Vol. 257).

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

TY - GEN

T1 - Expansion of the Variational Garrote to a Multiple Measurement Vectors Model

AU - Hansen, Sofie Therese

AU - Stahlhut, Carsten

AU - Hansen, Lars Kai

PY - 2013

Y1 - 2013

N2 - The recovery of sparse signals in underdetermined systems is the focus of this paper. We propose an expanded version of the Variational Garrote, originally presented by Kappen (2011), which can use multiple measurement vectors (MMVs) to further improve source retrieval performance. We show its superiority compared to the original formulation and demonstrate its ability to correctly estimate both the sources’ location and their magnitude. Finally evidence is given of the high performance of the proposed algorithm compared to other MMV models.

AB - The recovery of sparse signals in underdetermined systems is the focus of this paper. We propose an expanded version of the Variational Garrote, originally presented by Kappen (2011), which can use multiple measurement vectors (MMVs) to further improve source retrieval performance. We show its superiority compared to the original formulation and demonstrate its ability to correctly estimate both the sources’ location and their magnitude. Finally evidence is given of the high performance of the proposed algorithm compared to other MMV models.

U2 - 10.3233/978-1-61499-330-8-105

DO - 10.3233/978-1-61499-330-8-105

M3 - Article in proceedings

SN - 978-1-61499-329-2

T3 - Frontiers in Artificial Intelligence and Applications

SP - 105

EP - 114

BT - Twelfth Scandinavian Conference on Artificial Intelligence

A2 - Jaeger, M.

PB - IOS Press

T2 - 12th Scandinavian Conference on Artificial Intelligence (SCAI 2013)

Y2 - 20 November 2013 through 22 November 2013

ER -

Expansion of the Variational Garrote to a Multiple Measurement Vectors Model

Abstract

Conference

Access to Document

OpenUrl availability

Fingerprint

Cite this