A Log-Likelihood Ratio based Generalized Belief Propagation

Alexandru Amaricai, Mohsem Bahrami, Bane Vasić

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

Abstract

In this paper, we propose a reduced complexity Generalized Belief Propagation (GBP) that propagates messages in Log-Likelihood Ratio (LLR) domain. The key novelties of the proposed LLR-GBP are: (i) reduced fixed point precision for messages instead of computational complex floating point format, (ii) operations performed in logarithm domain, thus eliminating the need for multiplications and divisions, (iii) usage of message ratios that leads to simple hard decision mechanisms. We demonstrated the validity of LLR-GBP on reconstruction of images passed through binary-input two-dimensional Gaussian channels with memory and affected by additive white Gaussian noise.

Original languageEnglish (US)
Title of host publicationEUROCON 2019 - 18th International Conference on Smart Technologies
EditorsBoris Dumnic, Marko Delimar, Cedomir Stefanovic
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781538693018
DOIs
StatePublished - Jul 2019
Event18th International Conference on Smart Technologies, EUROCON 2019 - Novi Sad, Serbia
Duration: Jul 1 2019Jul 4 2019

Publication series

NameEUROCON 2019 - 18th International Conference on Smart Technologies

Conference

Conference18th International Conference on Smart Technologies, EUROCON 2019
CountrySerbia
CityNovi Sad
Period7/1/197/4/19

Fingerprint

Log-likelihood Ratio
Belief Propagation
Data storage equipment
Computer-Assisted Image Processing
Floating point
Gaussian White Noise
Logarithm
Division
Multiplication
Fixed point
Binary
Likelihood ratio
Propagation

Keywords

  • generalized belief propagation (GBP)
  • graphical models
  • Probabilistic inference

ASJC Scopus subject areas

  • Artificial Intelligence
  • Signal Processing
  • Management Science and Operations Research
  • Energy Engineering and Power Technology
  • Electrical and Electronic Engineering
  • Computational Mathematics
  • Control and Optimization
  • Health Informatics

Cite this

Amaricai, A., Bahrami, M., & Vasić, B. (2019). A Log-Likelihood Ratio based Generalized Belief Propagation. In B. Dumnic, M. Delimar, & C. Stefanovic (Eds.), EUROCON 2019 - 18th International Conference on Smart Technologies [8861528] (EUROCON 2019 - 18th International Conference on Smart Technologies). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/EUROCON.2019.8861528

A Log-Likelihood Ratio based Generalized Belief Propagation. / Amaricai, Alexandru; Bahrami, Mohsem; Vasić, Bane.

EUROCON 2019 - 18th International Conference on Smart Technologies. ed. / Boris Dumnic; Marko Delimar; Cedomir Stefanovic. Institute of Electrical and Electronics Engineers Inc., 2019. 8861528 (EUROCON 2019 - 18th International Conference on Smart Technologies).

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

Amaricai, A, Bahrami, M & Vasić, B 2019, A Log-Likelihood Ratio based Generalized Belief Propagation. in B Dumnic, M Delimar & C Stefanovic (eds), EUROCON 2019 - 18th International Conference on Smart Technologies., 8861528, EUROCON 2019 - 18th International Conference on Smart Technologies, Institute of Electrical and Electronics Engineers Inc., 18th International Conference on Smart Technologies, EUROCON 2019, Novi Sad, Serbia, 7/1/19. https://doi.org/10.1109/EUROCON.2019.8861528
Amaricai A, Bahrami M, Vasić B. A Log-Likelihood Ratio based Generalized Belief Propagation. In Dumnic B, Delimar M, Stefanovic C, editors, EUROCON 2019 - 18th International Conference on Smart Technologies. Institute of Electrical and Electronics Engineers Inc. 2019. 8861528. (EUROCON 2019 - 18th International Conference on Smart Technologies). https://doi.org/10.1109/EUROCON.2019.8861528
Amaricai, Alexandru ; Bahrami, Mohsem ; Vasić, Bane. / A Log-Likelihood Ratio based Generalized Belief Propagation. EUROCON 2019 - 18th International Conference on Smart Technologies. editor / Boris Dumnic ; Marko Delimar ; Cedomir Stefanovic. Institute of Electrical and Electronics Engineers Inc., 2019. (EUROCON 2019 - 18th International Conference on Smart Technologies).
@inproceedings{ec91d293af66478cafe4c0770d7e1f57,
title = "A Log-Likelihood Ratio based Generalized Belief Propagation",
abstract = "In this paper, we propose a reduced complexity Generalized Belief Propagation (GBP) that propagates messages in Log-Likelihood Ratio (LLR) domain. The key novelties of the proposed LLR-GBP are: (i) reduced fixed point precision for messages instead of computational complex floating point format, (ii) operations performed in logarithm domain, thus eliminating the need for multiplications and divisions, (iii) usage of message ratios that leads to simple hard decision mechanisms. We demonstrated the validity of LLR-GBP on reconstruction of images passed through binary-input two-dimensional Gaussian channels with memory and affected by additive white Gaussian noise.",
keywords = "generalized belief propagation (GBP), graphical models, Probabilistic inference",
author = "Alexandru Amaricai and Mohsem Bahrami and Bane Vasić",
year = "2019",
month = "7",
doi = "10.1109/EUROCON.2019.8861528",
language = "English (US)",
series = "EUROCON 2019 - 18th International Conference on Smart Technologies",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
editor = "Boris Dumnic and Marko Delimar and Cedomir Stefanovic",
booktitle = "EUROCON 2019 - 18th International Conference on Smart Technologies",

}

TY - GEN

T1 - A Log-Likelihood Ratio based Generalized Belief Propagation

AU - Amaricai, Alexandru

AU - Bahrami, Mohsem

AU - Vasić, Bane

PY - 2019/7

Y1 - 2019/7

N2 - In this paper, we propose a reduced complexity Generalized Belief Propagation (GBP) that propagates messages in Log-Likelihood Ratio (LLR) domain. The key novelties of the proposed LLR-GBP are: (i) reduced fixed point precision for messages instead of computational complex floating point format, (ii) operations performed in logarithm domain, thus eliminating the need for multiplications and divisions, (iii) usage of message ratios that leads to simple hard decision mechanisms. We demonstrated the validity of LLR-GBP on reconstruction of images passed through binary-input two-dimensional Gaussian channels with memory and affected by additive white Gaussian noise.

AB - In this paper, we propose a reduced complexity Generalized Belief Propagation (GBP) that propagates messages in Log-Likelihood Ratio (LLR) domain. The key novelties of the proposed LLR-GBP are: (i) reduced fixed point precision for messages instead of computational complex floating point format, (ii) operations performed in logarithm domain, thus eliminating the need for multiplications and divisions, (iii) usage of message ratios that leads to simple hard decision mechanisms. We demonstrated the validity of LLR-GBP on reconstruction of images passed through binary-input two-dimensional Gaussian channels with memory and affected by additive white Gaussian noise.

KW - generalized belief propagation (GBP)

KW - graphical models

KW - Probabilistic inference

UR - http://www.scopus.com/inward/record.url?scp=85074205839&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85074205839&partnerID=8YFLogxK

U2 - 10.1109/EUROCON.2019.8861528

DO - 10.1109/EUROCON.2019.8861528

M3 - Conference contribution

AN - SCOPUS:85074205839

T3 - EUROCON 2019 - 18th International Conference on Smart Technologies

BT - EUROCON 2019 - 18th International Conference on Smart Technologies

A2 - Dumnic, Boris

A2 - Delimar, Marko

A2 - Stefanovic, Cedomir

PB - Institute of Electrical and Electronics Engineers Inc.

ER -