Sampling from Gaussian graphical models using subgraph perturbations

Ying Liu, Oliver Kosut, Alan S. Willsky

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

4 Citations (Scopus)

Abstract

The problem of efficiently drawing samples from a Gaussian graphical model or Gaussian Markov random field is studied. We introduce the subgraph perturbation sampling algorithm, which makes use of any pre-existing tractable inference algorithm for a subgraph by perturbing this algorithm so as to yield asymptotically exact samples for the intended distribution. The subgraph can have any structure for which efficient inference algorithms exist: for example, tree-structured, low tree-width, or having a small feedback vertex set. The experimental results demonstrate that this subgraph perturbation algorithm efficiently yields accurate samples for many graph topologies.

Original languageEnglish (US)
Title of host publicationIEEE International Symposium on Information Theory - Proceedings
Pages2498-2502
Number of pages5
DOIs
StatePublished - 2013
Event2013 IEEE International Symposium on Information Theory, ISIT 2013 - Istanbul, Turkey
Duration: Jul 7 2013Jul 12 2013

Other

Other2013 IEEE International Symposium on Information Theory, ISIT 2013
CountryTurkey
CityIstanbul
Period7/7/137/12/13

Fingerprint

Gaussian Model
Graphical Models
Subgraph
Sampling
Perturbation
Gaussian Markov Random Field
Feedback Vertex Set
Treewidth
Topology
Feedback
Experimental Results
Graph in graph theory
Demonstrate

ASJC Scopus subject areas

  • Applied Mathematics
  • Modeling and Simulation
  • Theoretical Computer Science
  • Information Systems

Cite this

Liu, Y., Kosut, O., & Willsky, A. S. (2013). Sampling from Gaussian graphical models using subgraph perturbations. In IEEE International Symposium on Information Theory - Proceedings (pp. 2498-2502). [6620676] https://doi.org/10.1109/ISIT.2013.6620676

Sampling from Gaussian graphical models using subgraph perturbations. / Liu, Ying; Kosut, Oliver; Willsky, Alan S.

IEEE International Symposium on Information Theory - Proceedings. 2013. p. 2498-2502 6620676.

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

Liu, Y, Kosut, O & Willsky, AS 2013, Sampling from Gaussian graphical models using subgraph perturbations. in IEEE International Symposium on Information Theory - Proceedings., 6620676, pp. 2498-2502, 2013 IEEE International Symposium on Information Theory, ISIT 2013, Istanbul, Turkey, 7/7/13. https://doi.org/10.1109/ISIT.2013.6620676
Liu Y, Kosut O, Willsky AS. Sampling from Gaussian graphical models using subgraph perturbations. In IEEE International Symposium on Information Theory - Proceedings. 2013. p. 2498-2502. 6620676 https://doi.org/10.1109/ISIT.2013.6620676
Liu, Ying ; Kosut, Oliver ; Willsky, Alan S. / Sampling from Gaussian graphical models using subgraph perturbations. IEEE International Symposium on Information Theory - Proceedings. 2013. pp. 2498-2502
@inproceedings{7592162d26444e978ac21922c80b4a22,
title = "Sampling from Gaussian graphical models using subgraph perturbations",
abstract = "The problem of efficiently drawing samples from a Gaussian graphical model or Gaussian Markov random field is studied. We introduce the subgraph perturbation sampling algorithm, which makes use of any pre-existing tractable inference algorithm for a subgraph by perturbing this algorithm so as to yield asymptotically exact samples for the intended distribution. The subgraph can have any structure for which efficient inference algorithms exist: for example, tree-structured, low tree-width, or having a small feedback vertex set. The experimental results demonstrate that this subgraph perturbation algorithm efficiently yields accurate samples for many graph topologies.",
author = "Ying Liu and Oliver Kosut and Willsky, {Alan S.}",
year = "2013",
doi = "10.1109/ISIT.2013.6620676",
language = "English (US)",
isbn = "9781479904464",
pages = "2498--2502",
booktitle = "IEEE International Symposium on Information Theory - Proceedings",

}

TY - GEN

T1 - Sampling from Gaussian graphical models using subgraph perturbations

AU - Liu, Ying

AU - Kosut, Oliver

AU - Willsky, Alan S.

PY - 2013

Y1 - 2013

N2 - The problem of efficiently drawing samples from a Gaussian graphical model or Gaussian Markov random field is studied. We introduce the subgraph perturbation sampling algorithm, which makes use of any pre-existing tractable inference algorithm for a subgraph by perturbing this algorithm so as to yield asymptotically exact samples for the intended distribution. The subgraph can have any structure for which efficient inference algorithms exist: for example, tree-structured, low tree-width, or having a small feedback vertex set. The experimental results demonstrate that this subgraph perturbation algorithm efficiently yields accurate samples for many graph topologies.

AB - The problem of efficiently drawing samples from a Gaussian graphical model or Gaussian Markov random field is studied. We introduce the subgraph perturbation sampling algorithm, which makes use of any pre-existing tractable inference algorithm for a subgraph by perturbing this algorithm so as to yield asymptotically exact samples for the intended distribution. The subgraph can have any structure for which efficient inference algorithms exist: for example, tree-structured, low tree-width, or having a small feedback vertex set. The experimental results demonstrate that this subgraph perturbation algorithm efficiently yields accurate samples for many graph topologies.

UR - http://www.scopus.com/inward/record.url?scp=84890407746&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84890407746&partnerID=8YFLogxK

U2 - 10.1109/ISIT.2013.6620676

DO - 10.1109/ISIT.2013.6620676

M3 - Conference contribution

SN - 9781479904464

SP - 2498

EP - 2502

BT - IEEE International Symposium on Information Theory - Proceedings

ER -