Joint Source-Channel Coding for a Multivariate Gaussian over a Gaussian MAC using Variational Domain Adaptation
With the development of the distributed learning and edge computing, servers must often receive information from multiple terminal devices; thus, the importance of source-channel coding for distributed sources over multiple access channels (MACs) becomes more and more significant. This study present...
Saved in:
| Published in: | IEEE transactions on cognitive communications and networking Vol. 9; no. 6; p. 1 |
|---|---|
| Main Authors: | , , |
| Format: | Journal Article |
| Language: | English |
| Published: |
Piscataway
IEEE
01.12.2023
The Institute of Electrical and Electronics Engineers, Inc. (IEEE) |
| Subjects: | |
| ISSN: | 2332-7731, 2332-7731 |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Abstract | With the development of the distributed learning and edge computing, servers must often receive information from multiple terminal devices; thus, the importance of source-channel coding for distributed sources over multiple access channels (MACs) becomes more and more significant. This study presents a deep joint source-channel coding (JSCC) design for a multivariate Gaussian source over a Gaussian MAC. The widely used autoencoder based deep-JSCC cannot perform stably under such conditions due to their easiness to fall into local optimum. Therefore we propose the variational domain adaptation (VDA)-JSCC scheme. Firstly, the loss function with an additional regularization term is introduced through variational analysis. The crucial prior distribution related to this item is obtained by domain adaptation, which is a transfer learning method. The proposed fine-tuning technique during the training process yields further performance improvement. Experiment results show that VDA-JSCC can always learn reasonable coding structures without artificial design and outperforms other state-of-the-art methods under different channel signal-to-noise ratios (CSNRs). We have also analyzed the reason why the performance of VDA-JSCC deteriorates in high CSNR range and then replace the encoder of VDA-JSCC with Mixture-of-Experts to improve its performance in high CSNR range. Finally, VDA-JSCC exhibits considerable robustness when the channel quality or correlation coefficient varies. |
|---|---|
| AbstractList | With the development of the distributed learning and edge computing, servers must often receive information from multiple terminal devices; thus, the importance of source-channel coding for distributed sources over multiple access channels (MACs) becomes more and more significant. This letter presents a deep joint source-channel coding (JSCC) design for a multivariate Gaussian source over a Gaussian MAC. The widely used autoencoder based deep-JSCC cannot perform stably under such conditions due to their easiness to fall into local optimum. Therefore we propose the variational domain adaptation (VDA)-JSCC scheme. Firstly, the loss function with an additional regularization term is introduced through variational analysis. The crucial prior distribution related to this item is obtained by domain adaptation, which is a transfer learning method. The proposed fine-tuning technique during the training process yields further performance improvement. Experiment results show that VDA-JSCC can always learn reasonable coding structures without artificial design and outperforms other state-of-the-art methods under different channel signal-to-noise ratios (CSNRs). We have also analyzed the reason why the performance of VDA-JSCC deteriorates in high CSNR range and then replace the encoder of VDA-JSCC with Mixture-of-Experts to improve its performance in high CSNR range. Finally, VDA-JSCC exhibits considerable robustness when the channel quality or correlation coefficient varies. With the development of the distributed learning and edge computing, servers must often receive information from multiple terminal devices; thus, the importance of source-channel coding for distributed sources over multiple access channels (MACs) becomes more and more significant. This study presents a deep joint source-channel coding (JSCC) design for a multivariate Gaussian source over a Gaussian MAC. The widely used autoencoder based deep-JSCC cannot perform stably under such conditions due to their easiness to fall into local optimum. Therefore we propose the variational domain adaptation (VDA)-JSCC scheme. Firstly, the loss function with an additional regularization term is introduced through variational analysis. The crucial prior distribution related to this item is obtained by domain adaptation, which is a transfer learning method. The proposed fine-tuning technique during the training process yields further performance improvement. Experiment results show that VDA-JSCC can always learn reasonable coding structures without artificial design and outperforms other state-of-the-art methods under different channel signal-to-noise ratios (CSNRs). We have also analyzed the reason why the performance of VDA-JSCC deteriorates in high CSNR range and then replace the encoder of VDA-JSCC with Mixture-of-Experts to improve its performance in high CSNR range. Finally, VDA-JSCC exhibits considerable robustness when the channel quality or correlation coefficient varies. |
| Author | Chen, Xuechen Li, Yishen Deng, Xiaoheng |
| Author_xml | – sequence: 1 givenname: Yishen surname: Li fullname: Li, Yishen organization: School of Computer Science and Engineering, Central South University, Changsha, China – sequence: 2 givenname: Xuechen orcidid: 0000-0002-7683-2933 surname: Chen fullname: Chen, Xuechen organization: School of Computer Science and Engineering, Central South University, Changsha, China – sequence: 3 givenname: Xiaoheng orcidid: 0000-0003-2740-8025 surname: Deng fullname: Deng, Xiaoheng organization: School of Computer Science and Engineering, Central South University, Changsha, China |
| BookMark | eNp9kM1OwzAQhC1UJErpAyBxsMQ5xX-J7WMVoIBaOFC4Rm7igKvULnZSibcnaSpUceDktXe-Xc-cg4F1VgNwidEEYyRvlmn6PCGI0AklkvGYnYAhoZREnFM8OKrPwDiENUIIJyRJBBsC9-SMreGra3yuo_RTWasrmLrC2A9YOg8VXDRVbXbKG1VrOFNNCEZZ6Ha6a_7eF9MUNqGj3vdS46yq4K3bKGPhtFDbev92AU5LVQU9Ppwj8HZ_t0wfovnL7DGdzqO8dVBHK1HwokAcCRFzjVmxwnkiNMWSE0mLPC-l1EwQqqlKNCMYr5BICoGULFuopCNw3c_devfV6FBn69Zi-6WQESElS7hgcavCvSr3LgSvy2zrzUb57wyjrIs266LNumizQ7Qtw_8wuem91V6Z6l_yqieN1vpoExYIxZL-ADoXiBI |
| CODEN | ITCCG7 |
| CitedBy_id | crossref_primary_10_1109_TCCN_2024_3384500 |
| Cites_doi | 10.1109/IGARSS.2018.8518078 10.1109/CISS.2015.7086847 10.1109/DCC.2019.00057 10.1109/TCCN.2019.2919300 10.1109/ICNNSP.2003.1279235 10.1109/LCOMM.2013.120413.132088 10.1109/SPAWC48557.2020.9154331 10.1007/s10462-021-10067-3 10.1109/ICC42927.2021.9500692 10.1109/TCOMM.2011.081711.090298 10.1109/TKDE.2009.191 10.1162/neco.1991.3.1.79 10.1609/aaai.v31i1.10983 10.1109/HPCC-SmartCity-DSS50907.2020.00162 10.1109/TPAMI.2021.3062644 10.1109/TIT.1965.1053821 10.1109/JSAIT.2020.2987203 10.1109/JSAC.2021.3078489 10.1109/TCOMM.2012.071912.110456 10.1109/TCSVT.2021.3082521 10.1109/SPCOM50965.2020.9179539 10.1109/TNNLS.2019.2900734 10.1109/TCOMM.2017.2726065 10.1109/TIT.2010.2044058 10.1007/s10994-009-5152-4 10.1109/SPAWC.2019.8815402 10.5555/2946645.2946704 10.1109/TIT.1980.1056273 10.1109/TCOMM.2015.2410774 10.1109/JSAC.2022.3180802 |
| ContentType | Journal Article |
| Copyright | Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2023 |
| Copyright_xml | – notice: Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2023 |
| DBID | 97E RIA RIE AAYXX CITATION 7SP 8FD L7M |
| DOI | 10.1109/TCCN.2023.3294754 |
| 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 Electronic Library (IEL) url: https://ieeexplore.ieee.org/ sourceTypes: Publisher |
| DeliveryMethod | fulltext_linktorsrc |
| Discipline | Engineering |
| EISSN | 2332-7731 |
| EndPage | 1 |
| ExternalDocumentID | 10_1109_TCCN_2023_3294754 10180059 |
| Genre | orig-research |
| GrantInformation_xml | – fundername: National Natural Science Foundation of China grantid: Nos. 62172441, 61772553 and 61301181. funderid: 10.13039/501100001809 |
| GroupedDBID | 0R~ 6IK 97E AAJGR AARMG AASAJ AAWTH ABAZT ABJNI ABQJQ ABVLG ACGFS AGQYO AHBIQ AKJIK AKQYR ALMA_UNASSIGNED_HOLDINGS ATWAV BEFXN BFFAM BGNUA BKEBE BPEOZ EBS IES IFIPE IPLJI JAVBF M43 OCL RIA RIE AAYXX AGSQL CITATION EJD O9- 7SP 8FD L7M |
| ID | FETCH-LOGICAL-c294t-b8d7dd0708857e14db1c68e3197293dccf99e4823e3a6e4211b086d80a9f708f3 |
| IEDL.DBID | RIE |
| ISICitedReferencesCount | 1 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001122336700016&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| ISSN | 2332-7731 |
| IngestDate | Mon Jun 30 04:44:32 EDT 2025 Tue Nov 18 22:18:31 EST 2025 Sat Nov 29 03:02:26 EST 2025 Wed Aug 27 02:21:38 EDT 2025 |
| IsPeerReviewed | false |
| IsScholarly | true |
| Issue | 6 |
| 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-c294t-b8d7dd0708857e14db1c68e3197293dccf99e4823e3a6e4211b086d80a9f708f3 |
| Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14 |
| ORCID | 0000-0002-7683-2933 0000-0003-2740-8025 0009-0009-5128-1221 |
| PQID | 2899467845 |
| PQPubID | 4437218 |
| PageCount | 1 |
| ParticipantIDs | ieee_primary_10180059 proquest_journals_2899467845 crossref_primary_10_1109_TCCN_2023_3294754 crossref_citationtrail_10_1109_TCCN_2023_3294754 |
| PublicationCentury | 2000 |
| PublicationDate | 2023-12-01 |
| PublicationDateYYYYMMDD | 2023-12-01 |
| PublicationDate_xml | – month: 12 year: 2023 text: 2023-12-01 day: 01 |
| PublicationDecade | 2020 |
| PublicationPlace | Piscataway |
| PublicationPlace_xml | – name: Piscataway |
| PublicationTitle | IEEE transactions on cognitive communications and networking |
| PublicationTitleAbbrev | TCCN |
| PublicationYear | 2023 |
| 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 | ref13 ref35 ref12 ref15 Song (ref20) 2020 ref14 Choi (ref19) ref31 ref30 ref11 ref33 Sønderby (ref29) ref2 ref1 ref17 Kingma (ref37) ref39 Floor (ref10) ref16 ref38 ref18 Kingma (ref28) 2014 ref24 ref23 ref26 Glorot (ref34) ref25 Loshchilov (ref36) ref22 ref21 ref27 ref8 ref7 ref9 ref4 ref3 ref6 ref5 Ghosh (ref32) |
| References_xml | – ident: ref39 doi: 10.1109/IGARSS.2018.8518078 – ident: ref5 doi: 10.1109/CISS.2015.7086847 – start-page: 2084 volume-title: Proc. 19th Eur. Signal Process. Conf. (EUSIPCO) ident: ref10 article-title: Distributed zero-delay joint source-channel coding for a bi-variate Gaussian on a Gaussian MAC – ident: ref35 doi: 10.1109/DCC.2019.00057 – ident: ref2 doi: 10.1109/TCCN.2019.2919300 – ident: ref16 doi: 10.1109/ICNNSP.2003.1279235 – start-page: 249 volume-title: Proc. 13th Int. Conf. Artif. Intell. Statist. ident: ref34 article-title: Understanding the difficulty of training deep feedforward neural networks – start-page: 3745 volume-title: Proc. 30th Int. Conf. Neural Inf. Process. Syst. ident: ref29 article-title: Ladder variational autoencoders – start-page: 1 volume-title: Proc. Int. Conf. Learn. Represent. ident: ref36 article-title: SGDR: Stochastic gradient descent with warm restarts – ident: ref12 doi: 10.1109/LCOMM.2013.120413.132088 – ident: ref23 doi: 10.1109/SPAWC48557.2020.9154331 – ident: ref14 doi: 10.1007/s10462-021-10067-3 – ident: ref18 doi: 10.1109/ICC42927.2021.9500692 – ident: ref4 doi: 10.1109/TCOMM.2011.081711.090298 – ident: ref24 doi: 10.1109/TKDE.2009.191 – ident: ref33 doi: 10.1162/neco.1991.3.1.79 – start-page: 1182 volume-title: Proc. Int. Conf. Mach. Learn. ident: ref19 article-title: NECST: Neural joint source-channel coding – ident: ref30 doi: 10.1609/aaai.v31i1.10983 – start-page: 1 volume-title: Proc. Int. Conf. Learn. Represent. ident: ref37 article-title: Adam: A method for stochastic optimization – ident: ref38 doi: 10.1109/HPCC-SmartCity-DSS50907.2020.00162 – ident: ref26 doi: 10.1109/TPAMI.2021.3062644 – ident: ref6 doi: 10.1109/TIT.1965.1053821 – ident: ref13 doi: 10.1109/JSAIT.2020.2987203 – ident: ref22 doi: 10.1109/JSAC.2021.3078489 – year: 2014 ident: ref28 article-title: Auto-encoding variational bayes publication-title: arXiv:1312.6114 – ident: ref3 doi: 10.1109/TCOMM.2012.071912.110456 – ident: ref15 doi: 10.1109/TCSVT.2021.3082521 – start-page: 1 volume-title: Proc. Int. Conf. Learn. Represent. ident: ref32 article-title: From variational to deterministic autoencoders – ident: ref17 doi: 10.1109/SPCOM50965.2020.9179539 – ident: ref31 doi: 10.1109/TNNLS.2019.2900734 – year: 2020 ident: ref20 article-title: Improving unsupervised domain adaptation with variational information bottleneck publication-title: arXiv:1911.09310 – ident: ref1 doi: 10.1109/TCOMM.2017.2726065 – ident: ref9 doi: 10.1109/TIT.2010.2044058 – ident: ref27 doi: 10.1007/s10994-009-5152-4 – ident: ref7 doi: 10.1109/SPAWC.2019.8815402 – ident: ref25 doi: 10.5555/2946645.2946704 – ident: ref8 doi: 10.1109/TIT.1980.1056273 – ident: ref11 doi: 10.1109/TCOMM.2015.2410774 – ident: ref21 doi: 10.1109/JSAC.2022.3180802 |
| SSID | ssj0001626684 |
| Score | 2.2696831 |
| Snippet | With the development of the distributed learning and edge computing, servers must often receive information from multiple terminal devices; thus, the... |
| SourceID | proquest crossref ieee |
| SourceType | Aggregation Database Enrichment Source Index Database Publisher |
| StartPage | 1 |
| SubjectTerms | Adaptation Adaptation models Coding Correlation coefficients Decoding Deep joint source-channel coding Delays domain adaptation Domains Edge computing Encoding Feature extraction Learning multiple-access channel Multivariate analysis multivariate Gaussian Regularization Symbols variational autoencoder |
| Title | Joint Source-Channel Coding for a Multivariate Gaussian over a Gaussian MAC using Variational Domain Adaptation |
| URI | https://ieeexplore.ieee.org/document/10180059 https://www.proquest.com/docview/2899467845 |
| Volume | 9 |
| WOSCitedRecordID | wos001122336700016&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 Electronic Library (IEL) customDbUrl: eissn: 2332-7731 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0001626684 issn: 2332-7731 databaseCode: RIE dateStart: 20150101 isFulltext: true titleUrlDefault: https://ieeexplore.ieee.org/ providerName: IEEE |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV3NS8MwFA86POjBz4nTKTl4Erq1S9Ykx1GdIjgEp-xW0iSVwWzH1u3vNy_t5kQUhB5a-hJKfknf93sIXQtqtwXTxBNGhB4lCfh3E-0FzAdbvEkDxV2zCTYY8NFIPFfJ6i4Xxhjjgs9MC26dL1_nagGmsrarNmXlgW20zVhYJmt9GVSsaB5yWnkuA1-0h1E0aEF78BbpCMq69Bvvcc1UfvyBHVvpH_zzgw7RfiU_4l4J-BHaMtkx2tuoKniC8sd8nBX4xZnlPUgfyMwERzlwKWxlVCyxS7tdWjXZSpr4Xi7mkEqJIZrTvlw_P_UiDHHx7_jNkTqrIb7NP-Q4wz0tp6Ubv45e-3fD6MGr-ip4yq5D4SVcM63tWee8y0xAdRKokBsCHcgE0UqlQhjKO8QQGRpqVcTEKj6a-1KkdlBKTlEtyzNzhrC283SlkIH0E6q1PeyBDKm9QsJUokwD-asVj1VVdBx6X0xip3z4IgaQYgAprkBqoJv1kGlZceMv4jqgskFYAtJAzRWucXUo5zHolpYvcNo9_2XYBdqF2ctwlSaqFbOFuUQ7almM57Mrt98-AZ4_01M |
| linkProvider | IEEE |
| linkToHtml | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1bS8MwFD7oFNQHrxPnNQ8-CZ3tkrXJ46iXqdsQnOJbSZNMBtrKLv5-c9LOC6Ig9KGlJ23JSXru5wM4Fswui0hTTxgReoymGN9NtRdEPvrizSBQ3IFNRL0ef3wUt2WxuquFMca45DNTx1MXy9e5mqKr7NR1m7L6wDwsIHRWWa716VKxynnIWRm7DHxx2o_jXh0Bwuu0IVjUZN-kj4NT-fEPdoLlYu2fn7QOq6UGSVoFyzdgzmSbsPKlr-AW5Nf5MJuQO-eY97CAIDPPJM5RThGrpRJJXOHtmzWUra5JLuV0jMWUBPM57c2P624rJpgZ_0QeHKnzG5Kz_EUOM9LS8rUI5Ffh_uK8H7e9ElnBU3YeJl7KdaS13e2cNyMTMJ0GKuSGIgaZoFqpgRCG8QY1VIaGWSMxtaaP5r4UAztoQLehkuWZ2QGi7XOaUshA-inT2m73QIbMHiGNVKpMDfzZjCeqbDuO6BfPiTM_fJEgkxJkUlIyqQYnH0Nei54bfxFXkStfCAuG1GB_xtek3JbjBK1LKxk4a-7-MuwIltr9bifpXPVu9mAZ31Qkr-xDZTKamgNYVG-T4Xh06NbeO8e51pw |
| 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=Joint+Source-Channel+Coding+for+a+Multivariate+Gaussian+Over+a+Gaussian+MAC+Using+Variational+Domain+Adaptation&rft.jtitle=IEEE+transactions+on+cognitive+communications+and+networking&rft.au=Li%2C+Yishen&rft.au=Chen%2C+Xuechen&rft.au=Deng%2C+Xiaoheng&rft.date=2023-12-01&rft.issn=2332-7731&rft.eissn=2332-7731&rft.volume=9&rft.issue=6&rft.spage=1424&rft.epage=1437&rft_id=info:doi/10.1109%2FTCCN.2023.3294754&rft.externalDBID=n%2Fa&rft.externalDocID=10_1109_TCCN_2023_3294754 |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=2332-7731&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=2332-7731&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=2332-7731&client=summon |