Parallel processing algorithm for Bayesian network inference

Gil Kaspi; Joel Ratsaby

doi:10.1109/EEEI.2012.6377114

Parallel processing algorithm for Bayesian network inference

Gil Kaspi, Joel Ratsaby

Department of Electrical and Electronics Engineering

Ariel University

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

Abstract

We introduce an algorithm for Bayesian network inference using parallel computations that perform variable-elimination over multiple threads of execution. The algorithm can be implemented on a collection of parallel execution entities on a single FPGA. Each execution entity performs addition and multiplication. Relative to the standard bucket elimination, the parallel algorithm reduces the computational time by an amount that depends on the coupling (probabilistic dependency) of the network and on the evidence available at time of prediction query.

Original language	American English
Title of host publication	2012 IEEE 27th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2012
DOIs	https://doi.org/10.1109/EEEI.2012.6377114
State	Published - 1 Dec 2012
Event	2012 IEEE 27th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2012 - Eilat, Israel Duration: 14 Nov 2012 → 17 Nov 2012

Publication series

Name	2012 IEEE 27th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2012

Conference

Conference	2012 IEEE 27th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2012
Country/Territory	Israel
City	Eilat
Period	14/11/12 → 17/11/12

All Science Journal Classification (ASJC) codes

Electrical and Electronic Engineering

Access to Document

10.1109/EEEI.2012.6377114

Cite this

@inproceedings{1e747b3e65e44eb3ab0214207f0ea148,

title = "Parallel processing algorithm for Bayesian network inference",

abstract = "We introduce an algorithm for Bayesian network inference using parallel computations that perform variable-elimination over multiple threads of execution. The algorithm can be implemented on a collection of parallel execution entities on a single FPGA. Each execution entity performs addition and multiplication. Relative to the standard bucket elimination, the parallel algorithm reduces the computational time by an amount that depends on the coupling (probabilistic dependency) of the network and on the evidence available at time of prediction query.",

author = "Gil Kaspi and Joel Ratsaby",

year = "2012",

month = dec,

day = "1",

doi = "10.1109/EEEI.2012.6377114",

language = "American English",

isbn = "9781467346801",

series = "2012 IEEE 27th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2012",

booktitle = "2012 IEEE 27th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2012",

note = "2012 IEEE 27th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2012 ; Conference date: 14-11-2012 Through 17-11-2012",

}

Kaspi, G & Ratsaby, J 2012, Parallel processing algorithm for Bayesian network inference. in 2012 IEEE 27th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2012., 6377114, 2012 IEEE 27th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2012, 2012 IEEE 27th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2012, Eilat, Israel, 14/11/12. https://doi.org/10.1109/EEEI.2012.6377114

Parallel processing algorithm for Bayesian network inference. / Kaspi, Gil; Ratsaby, Joel.
2012 IEEE 27th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2012. 2012. 6377114 (2012 IEEE 27th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2012).

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

TY - GEN

T1 - Parallel processing algorithm for Bayesian network inference

AU - Kaspi, Gil

AU - Ratsaby, Joel

PY - 2012/12/1

Y1 - 2012/12/1

N2 - We introduce an algorithm for Bayesian network inference using parallel computations that perform variable-elimination over multiple threads of execution. The algorithm can be implemented on a collection of parallel execution entities on a single FPGA. Each execution entity performs addition and multiplication. Relative to the standard bucket elimination, the parallel algorithm reduces the computational time by an amount that depends on the coupling (probabilistic dependency) of the network and on the evidence available at time of prediction query.

AB - We introduce an algorithm for Bayesian network inference using parallel computations that perform variable-elimination over multiple threads of execution. The algorithm can be implemented on a collection of parallel execution entities on a single FPGA. Each execution entity performs addition and multiplication. Relative to the standard bucket elimination, the parallel algorithm reduces the computational time by an amount that depends on the coupling (probabilistic dependency) of the network and on the evidence available at time of prediction query.

UR - http://www.scopus.com/inward/record.url?scp=84871980029&partnerID=8YFLogxK

U2 - 10.1109/EEEI.2012.6377114

DO - 10.1109/EEEI.2012.6377114

M3 - Conference contribution

SN - 9781467346801

T3 - 2012 IEEE 27th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2012

BT - 2012 IEEE 27th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2012

T2 - 2012 IEEE 27th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2012

Y2 - 14 November 2012 through 17 November 2012

ER -

Parallel processing algorithm for Bayesian network inference

Abstract

Publication series

Conference

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this