Polarized Dropout: A Novel Deep Joint Source Channel Coding Scheme for Erasure Channels

This letter investigates the challenges encountered by deep joint source-channel coding in erasure channels. We explore the effectiveness of the widely adopted dropout technique in endowing deep neural networks with resilience against erasures. However, directly applying dropout at the channel layer...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE communications letters Jg. 28; H. 9; S. 1986 - 1990
Hauptverfasser: Ren, Zichang, Wang, Yiru, Zhao, Yuping
Format: Journal Article
Sprache:Englisch
Veröffentlicht: New York IEEE 01.09.2024
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Schlagworte:
ISSN:1089-7798, 1558-2558
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Abstract This letter investigates the challenges encountered by deep joint source-channel coding in erasure channels. We explore the effectiveness of the widely adopted dropout technique in endowing deep neural networks with resilience against erasures. However, directly applying dropout at the channel layer introduces uncertainty into the neural network's training process, leading to performance degradation. To address this issue, we introduce the Polarized Dropout scheme and a novel network architecture that encodes analog symbols using the Walsh-Hadamard transform based on real-number field computation. Leveraging the polarization of symbol recovery probabilities, for a given erasure rate, a determined set of neurons will be assigned a dropout rate of 1, while the remainder are assigned a dropout rate of 0. Simulation results indicate a maximum enhancement of nearly 6dB in communication performance.
AbstractList This letter investigates the challenges encountered by deep joint source-channel coding in erasure channels. We explore the effectiveness of the widely adopted dropout technique in endowing deep neural networks with resilience against erasures. However, directly applying dropout at the channel layer introduces uncertainty into the neural network's training process, leading to performance degradation. To address this issue, we introduce the Polarized Dropout scheme and a novel network architecture that encodes analog symbols using the Walsh-Hadamard transform based on real-number field computation. Leveraging the polarization of symbol recovery probabilities, for a given erasure rate, a determined set of neurons will be assigned a dropout rate of 1, while the remainder are assigned a dropout rate of 0. Simulation results indicate a maximum enhancement of nearly 6dB in communication performance.
Author Ren, Zichang
Wang, Yiru
Zhao, Yuping
Author_xml – sequence: 1
  givenname: Zichang
  orcidid: 0009-0006-5041-5708
  surname: Ren
  fullname: Ren, Zichang
  email: vutoc_rzc@163.com
  organization: School of Electronics, Peking University, Beijing, China
– sequence: 2
  givenname: Yiru
  orcidid: 0000-0002-0686-1026
  surname: Wang
  fullname: Wang, Yiru
  email: yiruwang@stu.pku.edu.cn
  organization: School of Electronics, Peking University, Beijing, China
– sequence: 3
  givenname: Yuping
  orcidid: 0000-0003-1199-9318
  surname: Zhao
  fullname: Zhao, Yuping
  email: yuping.zhao@pku.edu.cn
  organization: School of Electronics, Peking University, Beijing, China
BookMark eNpNkFtLw0AQhRepYK3-AfFhwefUvWRvvpW03mitUMXHJU0mNqXNxt1E0F9vaov4MjMw58wZvlPUq1wFCF1QMqSUmOtpMp_NhoyweMhjxqiSR6hPhdAR60qvm4k2kVJGn6DTENaEEM0E7aO3Z7dJffkNOR57V7u2ucEj_OQ-YYPHADV-dGXV4IVrfQY4WaVV1W0Sl5fVO15kK9gCLpzHE5-G1v8pwhk6LtJNgPNDH6DX28lLch9N53cPyWgaZSxWTaSymORxkfE4V1IonlKyXHKmKahUC05SWRAqmCKy0EbnquAZKAEFi7nJqMn5AF3t79befbQQGrvuXq26SMspYUZKIVWnYntV5l0IHgpb-3Kb-i9Lid0BtL8A7Q6gPQDsTJd7UwkA_wzCSGIo_wHqh2zR
CODEN ICLEF6
Cites_doi 10.1109/TCCN.2019.2919300
10.1109/TCCN.2022.3151935
10.1109/TSP.2021.3071210
10.1109/GLOBECOM54140.2023.10436819
10.1109/TCCN.2017.2758370
10.1109/GLOBECOM54140.2023.10437020
10.1109/TCOMM.2023.3258487
10.1109/JSAC.2022.3223408
10.1109/JSAC.2023.3288252
10.1016/j.dsp.2021.103207
10.1109/TIT.2009.2021379
ContentType Journal Article
Copyright Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2024
Copyright_xml – notice: Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2024
DBID 97E
RIA
RIE
AAYXX
CITATION
7SP
8FD
L7M
DOI 10.1109/LCOMM.2024.3422176
DatabaseName IEEE All-Society Periodicals Package (ASPP) 2005–Present
IEEE All-Society Periodicals Package (ASPP) 1998–Present
IEEE Electronic Library (IEL)
CrossRef
Electronics & Communications Abstracts
Technology Research Database
Advanced Technologies Database with Aerospace
DatabaseTitle CrossRef
Technology Research Database
Advanced Technologies Database with Aerospace
Electronics & Communications Abstracts
DatabaseTitleList
Technology Research Database
Database_xml – sequence: 1
  dbid: RIE
  name: IEEE Xplore
  url: https://ieeexplore.ieee.org/
  sourceTypes: Publisher
DeliveryMethod fulltext_linktorsrc
Discipline Engineering
EISSN 1558-2558
EndPage 1990
ExternalDocumentID 10_1109_LCOMM_2024_3422176
10596091
Genre orig-research
GrantInformation_xml – fundername: National Key Research and Development Program of China
  grantid: 2018YFB1801402; 2018YFB1801403
  funderid: 10.13039/501100012166
GroupedDBID -~X
0R~
29I
4.4
5GY
5VS
6IK
97E
AAJGR
AARMG
AASAJ
AAWTH
ABAZT
ABQJQ
ABVLG
ACGFO
ACIWK
AENEX
AETIX
AGQYO
AGSQL
AHBIQ
AI.
AIBXA
AKJIK
AKQYR
ALLEH
ALMA_UNASSIGNED_HOLDINGS
ATWAV
AZLTO
BEFXN
BFFAM
BGNUA
BKEBE
BPEOZ
CS3
DU5
EBS
EJD
HZ~
H~9
IES
IFIPE
IFJZH
IPLJI
JAVBF
LAI
M43
O9-
OCL
P2P
RIA
RIE
RNS
TN5
VH1
AAYXX
CITATION
7SP
8FD
L7M
ID FETCH-LOGICAL-c247t-7c40d4fc34d76573a10bb3281e7a8530a6f0152706f898d7f3ce75ef2439c19d3
IEDL.DBID RIE
ISICitedReferencesCount 0
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001314323800006&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 1089-7798
IngestDate Mon Jun 30 10:10:30 EDT 2025
Sat Nov 29 03:56:15 EST 2025
Wed Aug 27 02:01:31 EDT 2025
IsPeerReviewed true
IsScholarly true
Issue 9
Language English
License https://ieeexplore.ieee.org/Xplorehelp/downloads/license-information/IEEE.html
https://doi.org/10.15223/policy-029
https://doi.org/10.15223/policy-037
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-c247t-7c40d4fc34d76573a10bb3281e7a8530a6f0152706f898d7f3ce75ef2439c19d3
Notes ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ORCID 0009-0006-5041-5708
0000-0003-1199-9318
0000-0002-0686-1026
PQID 3102966567
PQPubID 85419
PageCount 5
ParticipantIDs ieee_primary_10596091
proquest_journals_3102966567
crossref_primary_10_1109_LCOMM_2024_3422176
PublicationCentury 2000
PublicationDate 2024-09-01
PublicationDateYYYYMMDD 2024-09-01
PublicationDate_xml – month: 09
  year: 2024
  text: 2024-09-01
  day: 01
PublicationDecade 2020
PublicationPlace New York
PublicationPlace_xml – name: New York
PublicationTitle IEEE communications letters
PublicationTitleAbbrev LCOMM
PublicationYear 2024
Publisher IEEE
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Publisher_xml – name: IEEE
– name: The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
References Hinton (ref11) 2012
ref13
ref10
ref2
ref1
ref8
ref7
ref9
ref4
ref3
ref6
ref5
Krizhevsky (ref14) 2009
Srivastava (ref12) 2014; 15
References_xml – ident: ref1
  doi: 10.1109/TCCN.2019.2919300
– ident: ref5
  doi: 10.1109/TCCN.2022.3151935
– ident: ref3
  doi: 10.1109/TSP.2021.3071210
– ident: ref10
  doi: 10.1109/GLOBECOM54140.2023.10436819
– ident: ref4
  doi: 10.1109/TCCN.2017.2758370
– ident: ref7
  doi: 10.1109/GLOBECOM54140.2023.10437020
– ident: ref8
  doi: 10.1109/TCOMM.2023.3258487
– year: 2012
  ident: ref11
  article-title: Improving neural networks by preventing co-adaptation of feature detectors
  publication-title: arXiv:1207.0580
– year: 2009
  ident: ref14
  article-title: Learning multiple layers of features from tiny images
– ident: ref2
  doi: 10.1109/JSAC.2022.3223408
– ident: ref6
  doi: 10.1109/JSAC.2023.3288252
– volume: 15
  start-page: 1929
  issue: 1
  year: 2014
  ident: ref12
  article-title: Dropout: A simple way to prevent neural networks from overfitting
  publication-title: J. Mach. Learn. Res.
– ident: ref9
  doi: 10.1016/j.dsp.2021.103207
– ident: ref13
  doi: 10.1109/TIT.2009.2021379
SSID ssj0008251
Score 2.435873
Snippet This letter investigates the challenges encountered by deep joint source-channel coding in erasure channels. We explore the effectiveness of the widely adopted...
SourceID proquest
crossref
ieee
SourceType Aggregation Database
Index Database
Publisher
StartPage 1986
SubjectTerms Artificial neural networks
Channels
Coding
Decoding
deep learning
erasure channel
Joint source-channel coding
Number theory
Performance degradation
polar code
Receivers
Symbols
Training
Transfer functions
Transforms
Vectors
Title Polarized Dropout: A Novel Deep Joint Source Channel Coding Scheme for Erasure Channels
URI https://ieeexplore.ieee.org/document/10596091
https://www.proquest.com/docview/3102966567
Volume 28
WOSCitedRecordID wos001314323800006&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
journalDatabaseRights – providerCode: PRVIEE
  databaseName: IEEE Xplore
  customDbUrl:
  eissn: 1558-2558
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0008251
  issn: 1089-7798
  databaseCode: RIE
  dateStart: 19970101
  isFulltext: true
  titleUrlDefault: https://ieeexplore.ieee.org/
  providerName: IEEE
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV3PS8MwFA4qHvTgz4nTKTl4k7osTZPW29gcIjoFf-1WsuQFBtqOrfPgX2-StjIRD94KTUp5X5v3veR97yF0pm0YIaWCQJpYBgy4OyQkJqDC2AlRQo2vwPdyK4bDeDRKHiqxutfCAIBPPoMLd-nP8nWuFm6rrO24ACdOq74qBC_FWt_LrtNgltn0iaWMSVwrZEjSvu3d393ZWJCyi5BRS8L5Dy_k26r8Wou9gxls__PVdtBWxSRxt4R-F61Atoc2l-oL7qPXBxe4Tj5B477rhrAoLnEXD_MPeMN9gCm-ySdZgR_9Bj52QoPM3unlzp_hRwvnO2BLavHVzG8k1iPmDfQ8uHrqXQdVJ4VAUSaKQChGNDMqZFrwSISyQ8bjkMYdENL6ayK5Ia6_LeEmTmItTKhARGCopSuqk-jwAK1leQaHCDMwkQXXErVIMy3VWNqYyVqWS0og5FETndeWTadlwYzUBxokST0OqcMhrXBoooaz5dLI0oxN1KrRSKufap5aJkptdBZxcfTHtGO04Z5e5oC10FoxW8AJWlcfxWQ-O_Xfyxc9N7ys
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV3JTsMwELUQIAEHdkRZfeCGAq7j2Am3qoBY2lKJ9Ra59liqBEnVpj3w9dhOiooQB26RYivRvMTzxp43g9CJtmGElAoCaWIZMODukJCYgApjJ0QJNb4C30tLdDrx21vSrcTqXgsDAD75DM7cpT_L17kau62yc8cFOHFa9QXXOquSa30vvE6FWebTJ5Y0JvFUI0OS81bzod220SBlZyGjlobzH37IN1b5tRp7F3O99s-XW0erFZfEjRL8DTQH2SZamakwuIVeuy507X-CxpeuH8K4uMAN3Mkn8I4vAQb4Lu9nBX70W_jYSQ0ye6eZO4-GHy2gH4AtrcVXQ7-VOB0x2kbP11dPzZug6qUQKMpEEQjFiGZGhUwLHolQ1kmvF9K4DkJaj00kN8R1uCXcxEmshQkViAgMtYRF1RMd7qD5LM9gF2EGJrLwWqoWaaal6kkbNVnLckkJhDyqodOpZdNBWTIj9aEGSVKPQ-pwSCscamjb2XJmZGnGGjqYopFWv9UotVyU2vgs4mLvj2nHaOnmqd1KW7ed-3207J5UZoQdoPliOIZDtKgmRX80PPLfzhdBXL_1
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Polarized+Dropout%3A+A+Novel+Deep+Joint+Source+Channel+Coding+Scheme+for+Erasure+Channels&rft.jtitle=IEEE+communications+letters&rft.au=Ren%2C+Zichang&rft.au=Wang%2C+Yiru&rft.au=Zhao%2C+Yuping&rft.date=2024-09-01&rft.issn=1089-7798&rft.eissn=1558-2558&rft.volume=28&rft.issue=9&rft.spage=1986&rft.epage=1990&rft_id=info:doi/10.1109%2FLCOMM.2024.3422176&rft.externalDBID=n%2Fa&rft.externalDocID=10_1109_LCOMM_2024_3422176
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1089-7798&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1089-7798&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1089-7798&client=summon