Please use this identifier to cite or link to this item: https://doi.org/10.15480/882.1798
DC FieldValueLanguage
dc.contributor.authorStark, Maximilian-
dc.contributor.authorLewandowsky, Jan-
dc.contributor.authorBauch, Gerhard-
dc.date.accessioned2018-10-25T12:49:20Z-
dc.date.available2018-10-25T12:49:20Z-
dc.date.issued2018-10-11-
dc.identifierdoi: 10.3390/app8101884-
dc.identifier.citationApplied Sciences 8 (10): 1884 (2018)de_DE
dc.identifier.issn2076-3417de_DE
dc.identifier.urihttp://hdl.handle.net/11420/1801-
dc.description.abstractIn high-throughput applications, low-complexity and low-latency channel decoders are inevitable. Hence, for low-density parity-check (LDPC) codes, message passing decoding has to be implemented with coarse quantization—that is, the exchanged beliefs are quantized with a small number of bits. This can result in a significant performance degradation with respect to decoding with high-precision messages. Recently, so-called <i>information-bottleneck</i> decoders were proposed which leverage a machine learning framework (i.e., the information bottleneck method) to design coarse-precision decoders with error-correction performance close to high-precision belief-propagation decoding. In these decoders, all conventional arithmetic operations are replaced by look-up operations. Irregular LDPC codes for next-generation fiber optical communication systems are characterized by high code rates and large maximum node degrees. Consequently, the implementation complexity is mainly influenced by the memory required to store the look-up tables. In this paper, we show that the complexity of <i>information-bottleneck</i> decoders remains manageable for irregular LDPC codes if our proposed construction approach is deployed. Furthermore, we reveal that in order to design information bottleneck decoders for arbitrary degree distributions, an intermediate construction step which we call <i>message alignment</i> has to be included. Exemplary numerical simulations show that incorporating message alignment in the construction yields a 4-bit information bottleneck decoder which performs only <inline-formula> <math display="inline"> <semantics> <mrow> <mn>0.15</mn> </mrow> </semantics> </math> </inline-formula> dB worse than a double-precision belief propagation decoder and outperforms a min-sum decoder.en
dc.language.isoende_DE
dc.publisherMultidisciplinary Digital Publishing Institutede_DE
dc.relation.ispartofApplied Sciencesde_DE
dc.rightsCC BY 4.0de_DE
dc.rightsinfo:eu-repo/semantics/openAccess-
dc.subjectchannel codingde_DE
dc.subjectlow-density parity-check codesde_DE
dc.subjectiterative decodingde_DE
dc.subjectinformation-bottleneck signal processingde_DE
dc.subjectclusteringde_DE
dc.subjectmachine learningde_DE
dc.subject.ddc620: Ingenieurwissenschaftende_DE
dc.titleInformation-Bottleneck Decoding of High-Rate Irregular LDPC Codes for Optical Communication Using Message Alignmentde_DE
dc.typeArticlede_DE
dc.date.updated2018-10-24T08:16:01Z-
dc.identifier.urnurn:nbn:de:gbv:830-882.06286-
dc.identifier.doi10.15480/882.1798-
dc.type.diniarticle-
dc.subject.ddccode620-
dcterms.DCMITypeText-
tuhh.identifier.urnurn:nbn:de:gbv:830-882.06286de_DE
tuhh.oai.showtruede_DE
dc.identifier.hdl11420/1801-
tuhh.abstract.englishIn high-throughput applications, low-complexity and low-latency channel decoders are inevitable. Hence, for low-density parity-check (LDPC) codes, message passing decoding has to be implemented with coarse quantization—that is, the exchanged beliefs are quantized with a small number of bits. This can result in a significant performance degradation with respect to decoding with high-precision messages. Recently, so-called <i>information-bottleneck</i> decoders were proposed which leverage a machine learning framework (i.e., the information bottleneck method) to design coarse-precision decoders with error-correction performance close to high-precision belief-propagation decoding. In these decoders, all conventional arithmetic operations are replaced by look-up operations. Irregular LDPC codes for next-generation fiber optical communication systems are characterized by high code rates and large maximum node degrees. Consequently, the implementation complexity is mainly influenced by the memory required to store the look-up tables. In this paper, we show that the complexity of <i>information-bottleneck</i> decoders remains manageable for irregular LDPC codes if our proposed construction approach is deployed. Furthermore, we reveal that in order to design information bottleneck decoders for arbitrary degree distributions, an intermediate construction step which we call <i>message alignment</i> has to be included. Exemplary numerical simulations show that incorporating message alignment in the construction yields a 4-bit information bottleneck decoder which performs only <inline-formula> <math display="inline"> <semantics> <mrow> <mn>0.15</mn> </mrow> </semantics> </math> </inline-formula> dB worse than a double-precision belief propagation decoder and outperforms a min-sum decoder.de_DE
tuhh.publisher.doi10.3390/app8101884-
tuhh.publication.instituteNachrichtentechnik E-8de_DE
tuhh.identifier.doi10.15480/882.1798-
tuhh.type.opus(wissenschaftlicher) Artikelde
tuhh.institute.germanNachrichtentechnik E-8de
tuhh.institute.englishNachrichtentechnik E-8de_DE
tuhh.gvk.hasppnfalse-
tuhh.hasurnfalse-
openaire.rightsinfo:eu-repo/semantics/openAccessde_DE
dc.type.driverarticle-
dc.rights.ccbyde_DE
dc.rights.ccversion4.0de_DE
dc.type.casraiJournal Articleen
tuhh.container.volume8.2018de_DE
tuhh.container.startpage1884de_DE
tuhh.container.endpage17 Seitende_DE
dc.relation.projectOpen Access Publizieren 2018 - 2019 / TU Hamburgde_DE
dc.rights.nationallicensefalsede_DE
item.fulltextWith Fulltext-
item.creatorOrcidStark, Maximilian-
item.creatorOrcidLewandowsky, Jan-
item.creatorOrcidBauch, Gerhard-
item.languageiso639-1other-
item.creatorGNDStark, Maximilian-
item.creatorGNDLewandowsky, Jan-
item.creatorGNDBauch, Gerhard-
item.grantfulltextopen-
crisitem.author.deptNachrichtentechnik E-8-
crisitem.author.deptNachrichtentechnik E-8-
crisitem.author.deptNachrichtentechnik E-8-
crisitem.author.orcid0000-0002-1750-5895-
crisitem.author.orcid0000-0001-7945-3528-
crisitem.author.orcid0000-0002-0050-2604-
crisitem.author.parentorgStudiendekanat Elektrotechnik, Informatik und Mathematik-
crisitem.author.parentorgStudiendekanat Elektrotechnik, Informatik und Mathematik-
crisitem.author.parentorgStudiendekanat Elektrotechnik, Informatik und Mathematik-
Appears in Collections:Publications (tub.dok)
Files in This Item:
File Description SizeFormat
applsci-08-01884-v2.pdf832,82 kBAdobe PDFView/Open
Show simple item record

Page view(s)

106
Last Week
4
Last month
25
checked on Jun 16, 2019

Download(s)

54
checked on Jun 16, 2019

Google ScholarTM

Check

Export

This item is licensed under a Creative Commons License Creative Commons