Applicability of single- and two-hidden-layer neural networks in decoding linear block codes

Srdan Brkic; Predrag Ivanis; Bane Vasic

doi:10.1109/TELFOR52709.2021.9653357

Applicability of single- and two-hidden-layer neural networks in decoding linear block codes

Srdan Brkic, Predrag Ivanis, Bane Vasic

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

Abstract

In this paper, we analyze applicability of single- and two-hidden-layer feed-forward artificial neural networks, SLFNs and TLFNs, respectively, in decoding linear block codes. Based on the provable capability of SLFNs and TLFNs to approximate discrete functions, we discuss sizes of the network capable to perform maximum likelihood decoding. Furthermore, we propose a decoding scheme, which use artificial neural networks (ANNs) to lower the error-floors of low-density parity-check (LDPC) codes. By learning a small number of error patterns, uncorrectable with typical decoders of LDPC codes, ANN can lower the error-floor by an order of magnitude, with only marginal average complexity incense.

Original language	English (US)
Title of host publication	2021 29th Telecommunications Forum, TELFOR 2021 - Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9781665425841
DOIs	https://doi.org/10.1109/TELFOR52709.2021.9653357
State	Published - 2021
Event	29th Telecommunications Forum, TELFOR 2021 - Virtual, Belgrade, Serbia Duration: Nov 23 2021 → Nov 24 2021

Publication series

Name	2021 29th Telecommunications Forum, TELFOR 2021 - Proceedings

Conference

Conference	29th Telecommunications Forum, TELFOR 2021
Country/Territory	Serbia
City	Virtual, Belgrade
Period	11/23/21 → 11/24/21

Keywords

Error-floors
Linear block codes
Low-density parity-check codes
ML decoding
Neural networks

ASJC Scopus subject areas

Computer Networks and Communications
Hardware and Architecture
Safety, Risk, Reliability and Quality

Access to Document

10.1109/TELFOR52709.2021.9653357

Cite this

Brkic, S., Ivanis, P., & Vasic, B. (2021). Applicability of single- and two-hidden-layer neural networks in decoding linear block codes. In 2021 29th Telecommunications Forum, TELFOR 2021 - Proceedings (2021 29th Telecommunications Forum, TELFOR 2021 - Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/TELFOR52709.2021.9653357

Applicability of single- and two-hidden-layer neural networks in decoding linear block codes. / Brkic, Srdan; Ivanis, Predrag; Vasic, Bane.

2021 29th Telecommunications Forum, TELFOR 2021 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2021. (2021 29th Telecommunications Forum, TELFOR 2021 - Proceedings).

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

Brkic, S, Ivanis, P & Vasic, B 2021, Applicability of single- and two-hidden-layer neural networks in decoding linear block codes. in 2021 29th Telecommunications Forum, TELFOR 2021 - Proceedings. 2021 29th Telecommunications Forum, TELFOR 2021 - Proceedings, Institute of Electrical and Electronics Engineers Inc., 29th Telecommunications Forum, TELFOR 2021, Virtual, Belgrade, Serbia, 11/23/21. https://doi.org/10.1109/TELFOR52709.2021.9653357

@inproceedings{505682e425ab444da6eb5f7828cfd015,

title = "Applicability of single- and two-hidden-layer neural networks in decoding linear block codes",

abstract = "In this paper, we analyze applicability of single- and two-hidden-layer feed-forward artificial neural networks, SLFNs and TLFNs, respectively, in decoding linear block codes. Based on the provable capability of SLFNs and TLFNs to approximate discrete functions, we discuss sizes of the network capable to perform maximum likelihood decoding. Furthermore, we propose a decoding scheme, which use artificial neural networks (ANNs) to lower the error-floors of low-density parity-check (LDPC) codes. By learning a small number of error patterns, uncorrectable with typical decoders of LDPC codes, ANN can lower the error-floor by an order of magnitude, with only marginal average complexity incense. ",

keywords = "Error-floors, Linear block codes, Low-density parity-check codes, ML decoding, Neural networks",

author = "Srdan Brkic and Predrag Ivanis and Bane Vasic",

note = "Publisher Copyright: {\textcopyright} 2021 IEEE.; 29th Telecommunications Forum, TELFOR 2021 ; Conference date: 23-11-2021 Through 24-11-2021",

year = "2021",

doi = "10.1109/TELFOR52709.2021.9653357",

language = "English (US)",

series = "2021 29th Telecommunications Forum, TELFOR 2021 - Proceedings",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

booktitle = "2021 29th Telecommunications Forum, TELFOR 2021 - Proceedings",

}

TY - GEN

T1 - Applicability of single- and two-hidden-layer neural networks in decoding linear block codes

AU - Brkic, Srdan

AU - Ivanis, Predrag

AU - Vasic, Bane

PY - 2021

Y1 - 2021

N2 - In this paper, we analyze applicability of single- and two-hidden-layer feed-forward artificial neural networks, SLFNs and TLFNs, respectively, in decoding linear block codes. Based on the provable capability of SLFNs and TLFNs to approximate discrete functions, we discuss sizes of the network capable to perform maximum likelihood decoding. Furthermore, we propose a decoding scheme, which use artificial neural networks (ANNs) to lower the error-floors of low-density parity-check (LDPC) codes. By learning a small number of error patterns, uncorrectable with typical decoders of LDPC codes, ANN can lower the error-floor by an order of magnitude, with only marginal average complexity incense.

AB - In this paper, we analyze applicability of single- and two-hidden-layer feed-forward artificial neural networks, SLFNs and TLFNs, respectively, in decoding linear block codes. Based on the provable capability of SLFNs and TLFNs to approximate discrete functions, we discuss sizes of the network capable to perform maximum likelihood decoding. Furthermore, we propose a decoding scheme, which use artificial neural networks (ANNs) to lower the error-floors of low-density parity-check (LDPC) codes. By learning a small number of error patterns, uncorrectable with typical decoders of LDPC codes, ANN can lower the error-floor by an order of magnitude, with only marginal average complexity incense.

KW - Error-floors

KW - Linear block codes

KW - Low-density parity-check codes

KW - ML decoding

KW - Neural networks

UR - http://www.scopus.com/inward/record.url?scp=85124623530&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85124623530&partnerID=8YFLogxK

U2 - 10.1109/TELFOR52709.2021.9653357

DO - 10.1109/TELFOR52709.2021.9653357

M3 - Conference contribution

AN - SCOPUS:85124623530

T3 - 2021 29th Telecommunications Forum, TELFOR 2021 - Proceedings

BT - 2021 29th Telecommunications Forum, TELFOR 2021 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 29th Telecommunications Forum, TELFOR 2021

Y2 - 23 November 2021 through 24 November 2021

ER -

Applicability of single- and two-hidden-layer neural networks in decoding linear block codes

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this