Faster Convergence With Less Communication: Broadcast-Based Subgraph Sampling for Decentralized Learning Over Wireless Networks

Decentralized stochastic gradient descent (D-SGD) is a widely adopted optimization algorithm for decentralized training of machine learning models across networked agents. A crucial part of D-SGD is the consensus-based model averaging, which heavily relies on information exchange and fusion among th...

Celý popis

Uloženo v:

Podrobná bibliografie
Vydáno v:	IEEE open journal of the Communications Society Ročník 6; s. 1497 - 1511
Hlavní autoři:	Perez Herrera, Daniel, Chen, Zheng, Larsson, Erik G.
Médium:	Journal Article
Jazyk:	angličtina
Vydáno:	New York IEEE 2025 The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Témata:	Algorithms broadcast Broadcasting Channels Collision avoidance Communication Convergence Costs D-SGD Decentralized machine learning Graph theory Laplace equations Machine learning Network topology node scheduling Nodes Optimization Sampling methods Scheduling Stochastic processes Symmetric matrices Topology Training Vectors Wireless networks broadcast Costs Symmetric matrices Laplace equations Stochastic processes wireless networks Vectors Topology node scheduling Optimization Convergence Training Network topology D-SGD Decentralized machine learning
ISSN:	2644-125X, 2644-125X
On-line přístup:	Získat plný text
Tagy:	Přidat tag Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!

Abstract	Decentralized stochastic gradient descent (D-SGD) is a widely adopted optimization algorithm for decentralized training of machine learning models across networked agents. A crucial part of D-SGD is the consensus-based model averaging, which heavily relies on information exchange and fusion among the nodes. For consensus averaging over wireless networks, due to the broadcast nature of wireless channels, simultaneous transmissions from multiple nodes may cause packet collisions if they share a common receiver. Therefore, communication coordination is necessary to determine when and how a node can transmit (or receive) information to (or from) its neighbors. In this work, we propose BASS , a broadcast-based subgraph sampling method designed to accelerate the convergence of D-SGD while considering the actual communication cost per iteration. BASS creates a set of mixing matrix candidates that represent sparser subgraphs of the base topology. In each consensus iteration, one mixing matrix is randomly sampled, leading to a specific scheduling decision that activates multiple collision-free subsets of nodes. The sampling occurs in a probabilistic manner, and the elements of the mixing matrices, along with their sampling probabilities, are jointly optimized. Simulation results demonstrate that BASS achieves faster convergence and requires fewer transmission slots than existing link-based scheduling methods and the full communication scenario. In conclusion, the inherent broadcasting nature of wireless channels offers intrinsic advantages in accelerating the convergence of decentralized optimization and learning.
AbstractList	Decentralized stochastic gradient descent (D-SGD) is a widely adopted optimization algorithm for decentralized training of machine learning models across networked agents. A crucial part of D-SGD is the consensus-based model averaging, which heavily relies on information exchange and fusion among the nodes. For consensus averaging over wireless networks, due to the broadcast nature of wireless channels, simultaneous transmissions from multiple nodes may cause packet collisions if they share a common receiver. Therefore, communication coordination is necessary to determine when and how a node can transmit (or receive) information to (or from) its neighbors. In this work, we propose BASS , a broadcast-based subgraph sampling method designed to accelerate the convergence of D-SGD while considering the actual communication cost per iteration. BASS creates a set of mixing matrix candidates that represent sparser subgraphs of the base topology. In each consensus iteration, one mixing matrix is randomly sampled, leading to a specific scheduling decision that activates multiple collision-free subsets of nodes. The sampling occurs in a probabilistic manner, and the elements of the mixing matrices, along with their sampling probabilities, are jointly optimized. Simulation results demonstrate that BASS achieves faster convergence and requires fewer transmission slots than existing link-based scheduling methods and the full communication scenario. In conclusion, the inherent broadcasting nature of wireless channels offers intrinsic advantages in accelerating the convergence of decentralized optimization and learning.
Author	Perez Herrera, Daniel Larsson, Erik G. Chen, Zheng
Author_xml	– sequence: 1 givenname: Daniel orcidid: 0000-0002-6097-7935 surname: Perez Herrera fullname: Perez Herrera, Daniel organization: Department of Electrical Engineering, Linköping University, Linköping, Sweden – sequence: 2 givenname: Zheng orcidid: 0000-0001-5621-2860 surname: Chen fullname: Chen, Zheng email: zheng.chen@liu.se organization: Department of Electrical Engineering, Linköping University, Linköping, Sweden – sequence: 3 givenname: Erik G. orcidid: 0000-0002-7599-4367 surname: Larsson fullname: Larsson, Erik G. organization: Department of Electrical Engineering, Linköping University, Linköping, Sweden
BackLink	https://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-212401$$DView record from Swedish Publication Index (Linköpings universitet)
BookMark	eNpVkc1u1DAUhS1UJMrQJ4BFJNaZ-vovCbt2SqFoYBbD385ynJuph0wc7ISqbHh1kkmFysrW9Xc-Xes8Jyetb5GQl0CXALQ433xYbT5ul4wyueRSUOD8CTllSogUmPx-8uj-jJzFuKd0RAGAi1Py59rEHkOy8u0vDDtsLSbfXH-brDHGcXo4DK2zpne-fZNcBm8qOwbSSxOxSrZDuQumu0225tA1rt0ltQ_JFVps-2Aa93tk1mhCOz1tRv-oDthM5k_Y3_nwI74gT2vTRDx7OBfky_Xbz6v36Xrz7mZ1sU4tL4o-FViiUrxkTPEailyWWaGAM2o4FRUImxWVKnOwVEqoa1UJK0pJS5UrZXMEviA3s7fyZq-74A4m3GtvnD4OfNhpE3pnG9RQKVNACZIpKZisc14JUY17gFU58HJ0pbMr3mE3lP_ZrtzXi6OtcYNmwKY6FuT1zHfB_xww9nrvh9CO39UcMgGZ5EeKz5QNPsaA9T8vUD0Vreei9VS0fih6TL2aUw4RHyXyrKA55X8BxaemZQ
CODEN	IOJCAZ
Cites_doi	10.1109/LSP.2023.3240647 10.1109/TWC.2022.3172147 10.1109/SECONWorkshops56311.2022.9926391 10.1109/TAC.2008.2009515 10.1109/FOCS46700.2020.00089 10.1090/conm/352 10.1109/LCSYS.2024.3487796 10.1017/CBO9780511804441 10.1109/TAC.2010.2079650 10.1137/060676866 10.1016/j.sysconle.2004.02.022 10.1109/TSP.2024.3363887 10.1109/CDC.2018.8619228 10.1109/SPAWC53906.2023.10304514 10.1109/JPROC.2020.3024266 10.1109/ACC.2016.7526803 10.1109/ICC51166.2024.10622904 10.1109/TSP.2022.3212536 10.1109/JSAC.2021.3118400 10.1109/JSTSP.2022.3152445 10.1137/060678324 10.4171/022-3/63 10.1109/JSAC.2023.3242735 10.1109/TWC.2023.3271854 10.1109/TSIPN.2022.3205549 10.1109/TAC.2014.2364096 10.1017/9781316216002 10.1109/CDC49753.2023.10383226 10.1109/JPROC.2018.2817461 10.1007/978-1-4612-0619-4 10.1109/JPROC.2006.887293 10.1109/JSAIT.2021.3103920 10.1109/TCCN.2021.3074908
ContentType	Journal Article
Copyright	Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2025
Copyright_xml	– notice: Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2025
DBID	97E ESBDL RIA RIE AAYXX CITATION 7SP 8FD L7M ABXSW ADTPV AOWAS D8T DG8 ZZAVC DOA
DOI	10.1109/OJCOMS.2025.3540133
DatabaseName	IEEE All-Society Periodicals Package (ASPP) 2005–Present IEEE Xplore Open Access Journals IEEE All-Society Periodicals Package (ASPP) 1998–Present IEEE Electronic Library (IEL) CrossRef Electronics & Communications Abstracts Technology Research Database Advanced Technologies Database with Aerospace SWEPUB Linköpings universitet full text SwePub SwePub Articles SWEPUB Freely available online SWEPUB Linköpings universitet SwePub Articles full text DOAJ Directory of Open Access Journals
DatabaseTitle	CrossRef Technology Research Database Advanced Technologies Database with Aerospace Electronics & Communications Abstracts
DatabaseTitleList	Technology Research Database
Database_xml	– sequence: 1 dbid: DOA name: DOAJ Directory of Open Access Journals url: https://www.doaj.org/ sourceTypes: Open Website – sequence: 2 dbid: RIE name: IEEE Xplore Electronic Library (IEL) url: https://ieeexplore.ieee.org/ sourceTypes: Publisher
DeliveryMethod	fulltext_linktorsrc
Discipline	Engineering
EISSN	2644-125X
EndPage	1511
ExternalDocumentID	oai_doaj_org_article_1d6a91b15265425f83d44dc391c6813b oai_DiVA_org_liu_212401 10_1109_OJCOMS_2025_3540133 10879080
Genre	orig-research
GrantInformation_xml	– fundername: Swedish Research Council (VR) – fundername: Zenith, ELLIIT, Knut and Alice Wallenberg Foundation
GroupedDBID	0R~ 97E AAJGR ABAZT ABVLG ALMA_UNASSIGNED_HOLDINGS BEFXN BFFAM BGNUA BKEBE BPEOZ EBS ESBDL GROUPED_DOAJ JAVBF M~E OCL OK1 RIA RIE AAYXX CITATION 7SP 8FD L7M ABXSW ADTPV AOWAS D8T DG8 ZZAVC
ID	FETCH-LOGICAL-c399t-4ebe663b2263f1985b7961320a304d14c79d6b81c0551ff6d4c4b50b6866c8e13
IEDL.DBID	RIE
ISICitedReferencesCount	0
ISICitedReferencesURI	http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001438165700004&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN	2644-125X
IngestDate	Fri Oct 03 12:53:55 EDT 2025 Tue Nov 04 16:32:11 EST 2025 Mon Jun 30 12:15:16 EDT 2025 Sat Nov 29 07:59:38 EST 2025 Wed Apr 23 05:41:09 EDT 2025
IsDoiOpenAccess	true
IsOpenAccess	true
IsPeerReviewed	true
IsScholarly	true
Keywords	broadcast Costs Symmetric matrices Laplace equations Stochastic processes wireless networks Vectors Topology node scheduling Optimization Convergence Training Network topology D-SGD Decentralized machine learning
Language	English
License	https://creativecommons.org/licenses/by-nc-nd/4.0
LinkModel	DirectLink
MergedId	FETCHMERGED-LOGICAL-c399t-4ebe663b2263f1985b7961320a304d14c79d6b81c0551ff6d4c4b50b6866c8e13
Notes	ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14
ORCID	0000-0002-7599-4367 0000-0002-6097-7935 0000-0001-5621-2860
OpenAccessLink	https://ieeexplore.ieee.org/document/10879080
PQID	3174175301
PQPubID	5075783
PageCount	15
ParticipantIDs	doaj_primary_oai_doaj_org_article_1d6a91b15265425f83d44dc391c6813b swepub_primary_oai_DiVA_org_liu_212401 crossref_primary_10_1109_OJCOMS_2025_3540133 ieee_primary_10879080 proquest_journals_3174175301
PublicationCentury	2000
PublicationDate	20250000 2025-00-00 20250101 2025 2025-01-01
PublicationDateYYYYMMDD	2025-01-01
PublicationDate_xml	– year: 2025 text: 20250000
PublicationDecade	2020
PublicationPlace	New York
PublicationPlace_xml	– name: New York
PublicationTitle	IEEE open journal of the Communications Society
PublicationTitleAbbrev	OJCOMS
PublicationYear	2025
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 ref34 ref37 ref36 ref31 ref30 ref11 ref33 ref10 ref32 ref1 McMahan (ref2) Scaman (ref12); 31 ref17 ref39 ref38 ref19 ref18 Lian (ref4); 30 Alistarh (ref26); 30 Krizhevsky (ref47) 2009 Lu (ref24) ref46 Wu (ref21) 2023 ref23 Neglia (ref14) Tang (ref25); 31 ref48 ref20 ref42 Koloskova (ref16) ref41 ref22 ref44 ref43 Vogels (ref15); 35 ref28 ref27 ref29 ref7 Swenson (ref3) 2020 ref6 ref5 Koloskova (ref9) LeCun (ref45) 2010 ref40 Wang (ref8) 2021; 22
References_xml	– ident: ref43 doi: 10.1109/LSP.2023.3240647 – volume: 22 start-page: 9709 issue: 1 year: 2021 ident: ref8 article-title: Cooperative SGD: A unified framework for the design and analysis of local-update SGD algorithms publication-title: J. Mach. Learn. Res. – ident: ref19 doi: 10.1109/TWC.2022.3172147 – year: 2009 ident: ref47 article-title: Learning multiple layers of features from tiny images – ident: ref20 doi: 10.1109/SECONWorkshops56311.2022.9926391 – ident: ref11 doi: 10.1109/TAC.2008.2009515 – ident: ref38 doi: 10.1109/FOCS46700.2020.00089 – volume: 30 start-page: 1 volume-title: Proc. 31st Conf. Neural Inf. Process. Syst. ident: ref4 article-title: Can decentralized algorithms outperform centralized algorithms? A case study for decentralized parallel stochastic gradient descent – volume-title: MNIST handwritten digit database year: 2010 ident: ref45 – ident: ref37 doi: 10.1090/conm/352 – volume: 30 start-page: 1707 volume-title: Proc. 31st Conf. Neural Inf. Process. Syst. ident: ref26 article-title: QSGD: Communication-efficient SGD via gradient quantization and encoding – volume: 31 start-page: 1 volume-title: Proc. 32nd Conf. Neural Inf. Process. Syst. ident: ref25 article-title: Communication compression for decentralized training – ident: ref41 doi: 10.1109/LCSYS.2024.3487796 – ident: ref48 doi: 10.1017/CBO9780511804441 – ident: ref32 doi: 10.1109/TAC.2010.2079650 – ident: ref46 doi: 10.1137/060676866 – start-page: 1273 volume-title: Proc. 20th Int. Conf. Artif. Intell. Statist. ident: ref2 article-title: Communication-efficient learning of deep networks from decentralized data – ident: ref6 doi: 10.1016/j.sysconle.2004.02.022 – ident: ref40 doi: 10.1109/TSP.2024.3363887 – year: 2020 ident: ref3 article-title: Distributed stochastic gradient descent: Nonconvexity, nonsmoothness, and convergence to local minima publication-title: arXiv:2003.02818 – ident: ref10 doi: 10.1109/CDC.2018.8619228 – ident: ref35 doi: 10.1109/SPAWC53906.2023.10304514 – ident: ref33 doi: 10.1109/JPROC.2020.3024266 – ident: ref42 doi: 10.1109/ACC.2016.7526803 – volume: 31 start-page: 1 volume-title: Proc. 32nd Conf. Neural Inf. Process. Syst. ident: ref12 article-title: Optimal algorithms for non-smooth distributed optimization in networks – volume: 35 start-page: 15039 volume-title: Proc. 36th Conf. Neural Inf. Process. Syst. ident: ref15 article-title: Beyond spectral gap: The role of the topology in decentralized learning – ident: ref1 doi: 10.1109/ICC51166.2024.10622904 – ident: ref27 doi: 10.1109/TSP.2022.3212536 – ident: ref34 doi: 10.1109/JSAC.2021.3118400 – ident: ref18 doi: 10.1109/JSTSP.2022.3152445 – ident: ref7 doi: 10.1137/060678324 – start-page: 2348 volume-title: Proc. Int. Conf. Artif. Intell. Statist. ident: ref14 article-title: Decentralized gradient methods: Does topology matter? – ident: ref39 doi: 10.4171/022-3/63 – start-page: 5381 volume-title: Proc. Int. Conf. Mach. Learn. ident: ref9 article-title: A unified theory of decentralized SGD with changing topology and local updates – ident: ref28 doi: 10.1109/JSAC.2023.3242735 – start-page: 6415 volume-title: Proc. Int. Conf. Mach. Learn. ident: ref24 article-title: Moniqua: Modulo quantized communication in decentralized SGD – ident: ref29 doi: 10.1109/TWC.2023.3271854 – ident: ref23 doi: 10.1109/TSIPN.2022.3205549 – ident: ref31 doi: 10.1109/TAC.2014.2364096 – ident: ref44 doi: 10.1017/9781316216002 – ident: ref17 doi: 10.1109/CDC49753.2023.10383226 – start-page: 3478 volume-title: Proc. Int. Conf. Mach. Learn. ident: ref16 article-title: Decentralized stochastic optimization and gossip algorithms with compressed communication – year: 2023 ident: ref21 article-title: Asynchronous distributed optimization with delay-free parameters publication-title: arXiv:2312.06508 – ident: ref13 doi: 10.1109/JPROC.2018.2817461 – ident: ref36 doi: 10.1007/978-1-4612-0619-4 – ident: ref5 doi: 10.1109/JPROC.2006.887293 – ident: ref30 doi: 10.1109/JSAIT.2021.3103920 – ident: ref22 doi: 10.1109/TCCN.2021.3074908
SSID	ssj0002511134
Score	2.2924466
Snippet	Decentralized stochastic gradient descent (D-SGD) is a widely adopted optimization algorithm for decentralized training of machine learning models across...
SourceID	doaj swepub proquest crossref ieee
SourceType	Open Website Open Access Repository Aggregation Database Index Database Publisher
StartPage	1497
SubjectTerms	Algorithms broadcast Broadcasting Channels Collision avoidance Communication Convergence Costs D-SGD Decentralized machine learning Graph theory Laplace equations Machine learning Network topology node scheduling Nodes Optimization Sampling methods Scheduling Stochastic processes Symmetric matrices Topology Training Vectors Wireless networks
SummonAdditionalLinks	– databaseName: DOAJ Directory of Open Access Journals dbid: DOA link: http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwrV1Lb9QwELZQxQEOiEcRgVL5gDgRmokdx-bWblkhBLtI5dGbFT9SIlUL6m459NK_zozjVtkTF66RYzv-Jp5vrPE3jL0KJvYgHZTBQFfKrmlKI5u6jFp6F1qBHsalYhPtYqFPT82XSakvygkb5YHHhTuAoDoDDkjGHe2r1yJIGbww4JUG4Wj3rVozCaZoDybiDEJmmSGozMHy42z5-QQDwrp5S2cdIMSWK0qK_bnEyjbbnCqIJq8zf8geZLrID8dpPmJ34uoxuz8REXzCrucdiR3wGeWPp6uUkf8YNj_5J9zE-NYFkHcco-4ueHyhPEL3FTjuG0mymp90lFq-OuNIYvlxzDmbwxW2yRKsZ3yJ_XNKlz2nnhdjAvl6l32bv_86-1DmsgolrprZlBJxQ57hkHiJHoxuXGsU3aTuRCUDSN-aoJwGXyGb6nsVpJeuqZzSSnkdQTxlO6tfq_iM8QCxim2N0bmqZMQ-G-dro0xdB3DKNQV7c7PC9veonmFT1FEZOwJiCRCbASnYEaFw25Skr9MDNAibDcL-yyAKtksYTsbTrUFeXLC9G1Bt_k_XFtmTJK3SCgr2egR6a_Tj4fthGv18uLTo5XGaz__HJF-we_Th41nOHtvZXFzGl-yu_7MZ1hf7yZb_Ag8L9YA priority: 102 providerName: Directory of Open Access Journals
Title	Faster Convergence With Less Communication: Broadcast-Based Subgraph Sampling for Decentralized Learning Over Wireless Networks
URI	https://ieeexplore.ieee.org/document/10879080 https://www.proquest.com/docview/3174175301 https://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-212401 https://doaj.org/article/1d6a91b15265425f83d44dc391c6813b
Volume	6
WOSCitedRecordID	wos001438165700004&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: PRVAON databaseName: DOAJ Directory of Open Access Journals customDbUrl: eissn: 2644-125X dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0002511134 issn: 2644-125X databaseCode: DOA dateStart: 20200101 isFulltext: true titleUrlDefault: https://www.doaj.org/ providerName: Directory of Open Access Journals – providerCode: PRVHPJ databaseName: ROAD: Directory of Open Access Scholarly Resources customDbUrl: eissn: 2644-125X dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0002511134 issn: 2644-125X databaseCode: M~E dateStart: 20200101 isFulltext: true titleUrlDefault: https://road.issn.org providerName: ISSN International Centre
link	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1Lb9QwEB7RigMcoEBRA23lA-JESiZ2nJhbu-2qQnQXqTx6s-JHSqRqi7q7HDjAX2fsuNXugUMvURQlfujz45vJzGeAN075DoXB3Clsc9FWVa5EVea-Eda4mtMOY-JhE_Vk0lxcqM8pWT3mwnjvY_CZPwi38V--u7bL4CqjGd7UiijOBmzUtRySte4cKoErIxdJWQgL9X76cTQ9OycbsKwOgnsDOV_bfaJIfzpVZZ1groqGxo1m_PSeTdyCJ4lRssNhCDyDB372HB6v6Ay-gL_jNughsFEIMY_Zlp597xc_2Cda59hajsgHRoZ56yx9kB_RDucYLS1R1ZqdtyH6fHbJiOeyY5_COvvf9E5Sab1kUyqfhYjaq1DyZIgxn2_D1_HJl9Fpnk5eyC0RlkUuCFqiIoa4Ge9QNZWplQzJ1i0vhENha-WkadAWRLi6TjphhakKIxspbeORv4TN2fXM7wBz6Atfl2TAy0J4KrMytlRSlaVDI02VwbtbRPTPQWBDR8OkUHoAUAcAdQIwg6OA2t2rQR07PiAUdJpsGp1sFRoM0v-0JnUNd0I46hpa2SA3GWwH5FbqG0DLYPd2EOg0leeaCJYIcqYFZvB2GBhrtR_33w5j7Vf9UhMRoGa--k_5r-FR6MvgwdmFzcXN0u_BQ_tr0c9v9qM7gK5nf07249D-B9ZC9mA
linkProvider	IEEE
linkToHtml	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1Nb9QwEB1BiwQcKB9FDS3gA-JEWjt2nLi3dsuqwHYXqQV6s-KPtJGqLerucuDCX2fsuNXugQO3KHLGtp7teXZmngHeOeVbJgzLnWJNLpqyzJUoi9zXwhpXcfQwJl42UY3H9fm5-pqS1WMujPc-Bp_53fAY_-W7a7sIR2U4w-tKIcW5D-ulEAXt07XujlQCW2ZcJG0hRtXe5PNgcnKKu8Ci3A0HHIzzFf8TZfrTvSqrFHNZNjS6muHGfzbyKTxJnJIc9IPgGdzz0-fweElp8AX8GTZBEYEMQpB5zLf05Ec3vyQjXOnISpbIPsGteeMsfpAfoo9zBBeXqGtNTpsQfz69IMh0yZFPgZ3dbyyTdFovyATtkxBTexUsj_so89kmfBt-PBsc5-nuhdwiZZnnAsFFMmKQnfGWqbo0lZIh3brhVDgmbKWcNDWzFClX20onrDAlNbKW0tae8ZewNr2e-i0gjnnqqwK38JIKjzZLYwslVVE4ZqQpM_hwi4j-2Uts6Lg1oUr3AOoAoE4AZnAYULsrGvSx4wtEQafpppmTjWKGBfF_XJXamjshHHaNWVkzbjLYDMgt1deDlsHO7SDQaTLPNFIsEQRNKcvgfT8wVmo_6r4fxNqvuoVGKoDNfPUP-2_h4fHZyUiPPo2_bMOj0K_-PGcH1uY3C_8aHthf82528yYO7b8qcPeB
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=Faster+Convergence+With+Less+Communication%3A+Broadcast-Based+Subgraph+Sampling+for+Decentralized+Learning+Over+Wireless+Networks&rft.jtitle=IEEE+open+journal+of+the+Communications+Society&rft.au=Perez+Herrera%2C+Daniel&rft.au=Chen%2C+Zheng&rft.au=Larsson%2C+Erik+G.&rft.date=2025&rft.pub=IEEE&rft.eissn=2644-125X&rft.volume=6&rft.spage=1497&rft.epage=1511&rft_id=info:doi/10.1109%2FOJCOMS.2025.3540133&rft.externalDocID=10879080
thumbnail_l	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=2644-125X&client=summon
thumbnail_m	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=2644-125X&client=summon
thumbnail_s	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=2644-125X&client=summon