Parametric and uncertainty computations with tensor product representations

Hermann G. Matthies; Alexander Litvinenko; Oliver Pajonk; Bojana V. Rosić; Elmar Zander

doi:10.1007/978-3-642-32677-6_9

Parametric and uncertainty computations with tensor product representations

Hermann G. Matthies^*, Alexander Litvinenko, Oliver Pajonk, Bojana V. Rosić, Elmar Zander

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

16 Citations (Scopus)

Abstract

Computational uncertainty quantification in a probabilistic setting is a special case of a parametric problem. Parameter dependent state vectors lead via association to a linear operator to analogues of covariance, its spectral decomposition, and the associated Karhunen-Loève expansion. From this one obtains a generalised tensor representation The parameter in question may be a tuple of numbers, a function, a stochastic process, or a random tensor field. The tensor factorisation may be cascaded, leading to tensors of higher degree. When carried on a discretised level, such factorisations in the form of low-rank approximations lead to very sparse representations of the high dimensional quantities involved. Updating of uncertainty for new information is an important part of uncertainty quantification. Formulated in terms or random variables instead of measures, the Bayesian update is a projection and allows the use of the tensor factorisations also in this case.

Original language	English
Title of host publication	Uncertainty Quantification in Scientific Computing
Subtitle of host publication	10th IFIP WG 2.5 Working Conference, WoCoUQ 2011, Revised Selected Papers
Editors	Andrew M. Dienstfrey, Ronald F. Boisvert
Place of Publication	Berlin, Heidelberg
Publisher	Springer
Pages	139-150
Number of pages	12
ISBN (Electronic)	978-3-642-32677-6
ISBN (Print)	978-3-642-32676-9
DOIs	https://doi.org/10.1007/978-3-642-32677-6_9
Publication status	Published - 1 Jan 2012
Externally published	Yes
Event	10th IFIP WG 2.5 Working Conference on Uncertainty Quantification in Scientific Computing, WoCoUQ 2011 - Boulder, United States Duration: 1 Aug 2011 → 4 Aug 2011 Conference number: 10

Publication series

Name	IFIP Advances in Information and Communication Technology
Publisher	Springer
Volume	377
ISSN (Print)	1868-4238
ISSN (Electronic)	1861-2288

Conference

Conference	10th IFIP WG 2.5 Working Conference on Uncertainty Quantification in Scientific Computing, WoCoUQ 2011
Abbreviated title	WoCoUQ
Country/Territory	United States
City	Boulder
Period	1/08/11 → 4/08/11

Keywords

Bayesian updating
low-rank tensor approximation
parametric problems
uncertainty quantification

Access to Document

10.1007/978-3-642-32677-6_9

Cite this

Matthies, H. G., Litvinenko, A., Pajonk, O., Rosić, B. V., & Zander, E. (2012). Parametric and uncertainty computations with tensor product representations. In A. M. Dienstfrey, & R. F. Boisvert (Eds.), Uncertainty Quantification in Scientific Computing: 10th IFIP WG 2.5 Working Conference, WoCoUQ 2011, Revised Selected Papers (pp. 139-150). (IFIP Advances in Information and Communication Technology; Vol. 377). Springer. https://doi.org/10.1007/978-3-642-32677-6_9

Matthies, Hermann G. ; Litvinenko, Alexander ; Pajonk, Oliver et al. / Parametric and uncertainty computations with tensor product representations. Uncertainty Quantification in Scientific Computing: 10th IFIP WG 2.5 Working Conference, WoCoUQ 2011, Revised Selected Papers. editor / Andrew M. Dienstfrey ; Ronald F. Boisvert. Berlin, Heidelberg : Springer, 2012. pp. 139-150 (IFIP Advances in Information and Communication Technology).

@inproceedings{442c6287ccba4cb898e71652ed2b2a0e,

title = "Parametric and uncertainty computations with tensor product representations",

abstract = "Computational uncertainty quantification in a probabilistic setting is a special case of a parametric problem. Parameter dependent state vectors lead via association to a linear operator to analogues of covariance, its spectral decomposition, and the associated Karhunen-Lo{\`e}ve expansion. From this one obtains a generalised tensor representation The parameter in question may be a tuple of numbers, a function, a stochastic process, or a random tensor field. The tensor factorisation may be cascaded, leading to tensors of higher degree. When carried on a discretised level, such factorisations in the form of low-rank approximations lead to very sparse representations of the high dimensional quantities involved. Updating of uncertainty for new information is an important part of uncertainty quantification. Formulated in terms or random variables instead of measures, the Bayesian update is a projection and allows the use of the tensor factorisations also in this case.",

keywords = "Bayesian updating, low-rank tensor approximation, parametric problems, uncertainty quantification",

author = "Matthies, {Hermann G.} and Alexander Litvinenko and Oliver Pajonk and Rosi{\'c}, {Bojana V.} and Elmar Zander",

year = "2012",

month = jan,

day = "1",

doi = "10.1007/978-3-642-32677-6_9",

language = "English",

isbn = "978-3-642-32676-9",

series = "IFIP Advances in Information and Communication Technology",

publisher = "Springer",

pages = "139--150",

editor = "Dienstfrey, {Andrew M.} and Boisvert, {Ronald F.}",

booktitle = "Uncertainty Quantification in Scientific Computing",

address = "Germany",

note = "10th IFIP WG 2.5 Working Conference on Uncertainty Quantification in Scientific Computing, WoCoUQ 2011, WoCoUQ ; Conference date: 01-08-2011 Through 04-08-2011",

}

Matthies, HG, Litvinenko, A, Pajonk, O, Rosić, BV & Zander, E 2012, Parametric and uncertainty computations with tensor product representations. in AM Dienstfrey & RF Boisvert (eds), Uncertainty Quantification in Scientific Computing: 10th IFIP WG 2.5 Working Conference, WoCoUQ 2011, Revised Selected Papers. IFIP Advances in Information and Communication Technology, vol. 377, Springer, Berlin, Heidelberg, pp. 139-150, 10th IFIP WG 2.5 Working Conference on Uncertainty Quantification in Scientific Computing, WoCoUQ 2011, Boulder, Colorado, United States, 1/08/11. https://doi.org/10.1007/978-3-642-32677-6_9

Parametric and uncertainty computations with tensor product representations. / Matthies, Hermann G.; Litvinenko, Alexander; Pajonk, Oliver et al.
Uncertainty Quantification in Scientific Computing: 10th IFIP WG 2.5 Working Conference, WoCoUQ 2011, Revised Selected Papers. ed. / Andrew M. Dienstfrey; Ronald F. Boisvert. Berlin, Heidelberg: Springer, 2012. p. 139-150 (IFIP Advances in Information and Communication Technology; Vol. 377).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Parametric and uncertainty computations with tensor product representations

AU - Matthies, Hermann G.

AU - Litvinenko, Alexander

AU - Pajonk, Oliver

AU - Rosić, Bojana V.

AU - Zander, Elmar

N1 - Conference code: 10

PY - 2012/1/1

Y1 - 2012/1/1

N2 - Computational uncertainty quantification in a probabilistic setting is a special case of a parametric problem. Parameter dependent state vectors lead via association to a linear operator to analogues of covariance, its spectral decomposition, and the associated Karhunen-Loève expansion. From this one obtains a generalised tensor representation The parameter in question may be a tuple of numbers, a function, a stochastic process, or a random tensor field. The tensor factorisation may be cascaded, leading to tensors of higher degree. When carried on a discretised level, such factorisations in the form of low-rank approximations lead to very sparse representations of the high dimensional quantities involved. Updating of uncertainty for new information is an important part of uncertainty quantification. Formulated in terms or random variables instead of measures, the Bayesian update is a projection and allows the use of the tensor factorisations also in this case.

AB - Computational uncertainty quantification in a probabilistic setting is a special case of a parametric problem. Parameter dependent state vectors lead via association to a linear operator to analogues of covariance, its spectral decomposition, and the associated Karhunen-Loève expansion. From this one obtains a generalised tensor representation The parameter in question may be a tuple of numbers, a function, a stochastic process, or a random tensor field. The tensor factorisation may be cascaded, leading to tensors of higher degree. When carried on a discretised level, such factorisations in the form of low-rank approximations lead to very sparse representations of the high dimensional quantities involved. Updating of uncertainty for new information is an important part of uncertainty quantification. Formulated in terms or random variables instead of measures, the Bayesian update is a projection and allows the use of the tensor factorisations also in this case.

KW - Bayesian updating

KW - low-rank tensor approximation

KW - parametric problems

KW - uncertainty quantification

UR - http://www.scopus.com/inward/record.url?scp=84868322658&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-32677-6_9

DO - 10.1007/978-3-642-32677-6_9

M3 - Conference contribution

AN - SCOPUS:84868322658

SN - 978-3-642-32676-9

T3 - IFIP Advances in Information and Communication Technology

SP - 139

EP - 150

BT - Uncertainty Quantification in Scientific Computing

A2 - Dienstfrey, Andrew M.

A2 - Boisvert, Ronald F.

PB - Springer

CY - Berlin, Heidelberg

T2 - 10th IFIP WG 2.5 Working Conference on Uncertainty Quantification in Scientific Computing, WoCoUQ 2011

Y2 - 1 August 2011 through 4 August 2011

ER -

Matthies HG, Litvinenko A, Pajonk O, Rosić BV, Zander E. Parametric and uncertainty computations with tensor product representations. In Dienstfrey AM, Boisvert RF, editors, Uncertainty Quantification in Scientific Computing: 10th IFIP WG 2.5 Working Conference, WoCoUQ 2011, Revised Selected Papers. Berlin, Heidelberg: Springer. 2012. p. 139-150. (IFIP Advances in Information and Communication Technology). doi: 10.1007/978-3-642-32677-6_9

Parametric and uncertainty computations with tensor product representations

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this