Exploiting within-clique factorizations in Junction-Tree Algorithms

Julian J. McAuley; Tibério S. Caetano

Exploiting within-clique factorizations in Junction-Tree Algorithms

Julian J. McAuley^*, Tibério S. Caetano

^*Corresponding author for this work

Research output: Contribution to journal › Conference article › peer-review

7 Citations (Scopus)

Abstract

We show that the expected computational complexity of the Junction-Tree Algorithm for maximum a posteriori inference in graphical models can be improved. Our results apply whenever the potentials over maximal cliques of the triangulated graph are factored over subcliques. This is common in many real applications, as we illustrate with several examples. The new algorithms are easily implemented, and experiments show substantial speed-ups over the classical Junction-Tree Algorithm. This enlarges the class of models for which exact inference is efficient.

Original language	English
Pages (from-to)	525-532
Number of pages	8
Journal	Journal of Machine Learning Research
Volume	9
Publication status	Published - 2010
Event	13th International Conference on Artificial Intelligence and Statistics, AISTATS 2010 - Sardinia, Italy Duration: 13 May 2010 → 15 May 2010

Cite this

@article{e9d6231400ab4807bf07669d84b5b608,

title = "Exploiting within-clique factorizations in Junction-Tree Algorithms",

abstract = "We show that the expected computational complexity of the Junction-Tree Algorithm for maximum a posteriori inference in graphical models can be improved. Our results apply whenever the potentials over maximal cliques of the triangulated graph are factored over subcliques. This is common in many real applications, as we illustrate with several examples. The new algorithms are easily implemented, and experiments show substantial speed-ups over the classical Junction-Tree Algorithm. This enlarges the class of models for which exact inference is efficient.",

author = "McAuley, {Julian J.} and Caetano, {Tib{\'e}rio S.}",

year = "2010",

language = "English",

volume = "9",

pages = "525--532",

journal = "Journal of Machine Learning Research",

issn = "1532-4435",

publisher = "Microtome Publishing",

note = "13th International Conference on Artificial Intelligence and Statistics, AISTATS 2010 ; Conference date: 13-05-2010 Through 15-05-2010",

}

TY - JOUR

T1 - Exploiting within-clique factorizations in Junction-Tree Algorithms

AU - McAuley, Julian J.

AU - Caetano, Tibério S.

PY - 2010

Y1 - 2010

N2 - We show that the expected computational complexity of the Junction-Tree Algorithm for maximum a posteriori inference in graphical models can be improved. Our results apply whenever the potentials over maximal cliques of the triangulated graph are factored over subcliques. This is common in many real applications, as we illustrate with several examples. The new algorithms are easily implemented, and experiments show substantial speed-ups over the classical Junction-Tree Algorithm. This enlarges the class of models for which exact inference is efficient.

AB - We show that the expected computational complexity of the Junction-Tree Algorithm for maximum a posteriori inference in graphical models can be improved. Our results apply whenever the potentials over maximal cliques of the triangulated graph are factored over subcliques. This is common in many real applications, as we illustrate with several examples. The new algorithms are easily implemented, and experiments show substantial speed-ups over the classical Junction-Tree Algorithm. This enlarges the class of models for which exact inference is efficient.

UR - http://www.scopus.com/inward/record.url?scp=84860172533&partnerID=8YFLogxK

M3 - Conference article

AN - SCOPUS:84860172533

SN - 1532-4435

VL - 9

SP - 525

EP - 532

JO - Journal of Machine Learning Research

JF - Journal of Machine Learning Research

T2 - 13th International Conference on Artificial Intelligence and Statistics, AISTATS 2010

Y2 - 13 May 2010 through 15 May 2010

ER -

Exploiting within-clique factorizations in Junction-Tree Algorithms

Abstract

Other files and links

Fingerprint

Cite this