Convergence Analysis of Distributed Gradient Descent Algorithms With One and Two Momentum Terms
For the centralized optimization, it is well known that adding one momentum term (also called the heavy-ball method) can obtain a faster convergence rate than the gradient method. However, for the distributed counterpart, there is quite few results about the effect of added momentum terms on the con...
Gespeichert in:
| Veröffentlicht in: | IEEE transactions on cybernetics Jg. 54; H. 3; S. 1511 - 1522 |
|---|---|
| Hauptverfasser: | , , |
| Format: | Journal Article |
| Sprache: | Englisch |
| Veröffentlicht: |
United States
IEEE
01.03.2024
The Institute of Electrical and Electronics Engineers, Inc. (IEEE) |
| Schlagworte: | |
| ISSN: | 2168-2267, 2168-2275, 2168-2275 |
| Online-Zugang: | Volltext |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| Abstract | For the centralized optimization, it is well known that adding one momentum term (also called the heavy-ball method) can obtain a faster convergence rate than the gradient method. However, for the distributed counterpart, there is quite few results about the effect of added momentum terms on the convergence rate. This article is aimed at studying the issue in the distributed setup, where <inline-formula> <tex-math notation="LaTeX">N </tex-math></inline-formula> agents minimize the sum of their individual cost functions using local communication over a network. The cost functions are twice continuously differentiable. We first study the algorithm with one momentum term and develop a distributed heavy-ball (D-HB) method by adding one momentum term on to the distributed gradient algorithm. By borrowing tools from the control theory, we provide a simple convergence proof and an explicit expression of the optimal convergence rate. Furthermore, we consider adding two momentum terms case and propose a distributed double-heavy-ball (D-DHB) method. We show that adding one momentum term allows faster convergence while adding two momentum terms does not perform any superiorities. Finally, simulation examples are given to illustrate our findings. |
|---|---|
| AbstractList | For the centralized optimization, it is well known that adding one momentum term (also called the heavy-ball method) can obtain a faster convergence rate than the gradient method. However, for the distributed counterpart, there is quite few results about the effect of added momentum terms on the convergence rate. This article is aimed at studying the issue in the distributed setup, where N agents minimize the sum of their individual cost functions using local communication over a network. The cost functions are twice continuously differentiable. We first study the algorithm with one momentum term and develop a distributed heavy-ball (D-HB) method by adding one momentum term on to the distributed gradient algorithm. By borrowing tools from the control theory, we provide a simple convergence proof and an explicit expression of the optimal convergence rate. Furthermore, we consider adding two momentum terms case and propose a distributed double-heavy-ball (D-DHB) method. We show that adding one momentum term allows faster convergence while adding two momentum terms does not perform any superiorities. Finally, simulation examples are given to illustrate our findings. For the centralized optimization, it is well known that adding one momentum term (also called the heavy-ball method) can obtain a faster convergence rate than the gradient method. However, for the distributed counterpart, there is quite few results about the effect of added momentum terms on the convergence rate. This article is aimed at studying the issue in the distributed setup, where N agents minimize the sum of their individual cost functions using local communication over a network. The cost functions are twice continuously differentiable. We first study the algorithm with one momentum term and develop a distributed heavy-ball (D-HB) method by adding one momentum term on to the distributed gradient algorithm. By borrowing tools from the control theory, we provide a simple convergence proof and an explicit expression of the optimal convergence rate. Furthermore, we consider adding two momentum terms case and propose a distributed double-heavy-ball (D-DHB) method. We show that adding one momentum term allows faster convergence while adding two momentum terms does not perform any superiorities. Finally, simulation examples are given to illustrate our findings.For the centralized optimization, it is well known that adding one momentum term (also called the heavy-ball method) can obtain a faster convergence rate than the gradient method. However, for the distributed counterpart, there is quite few results about the effect of added momentum terms on the convergence rate. This article is aimed at studying the issue in the distributed setup, where N agents minimize the sum of their individual cost functions using local communication over a network. The cost functions are twice continuously differentiable. We first study the algorithm with one momentum term and develop a distributed heavy-ball (D-HB) method by adding one momentum term on to the distributed gradient algorithm. By borrowing tools from the control theory, we provide a simple convergence proof and an explicit expression of the optimal convergence rate. Furthermore, we consider adding two momentum terms case and propose a distributed double-heavy-ball (D-DHB) method. We show that adding one momentum term allows faster convergence while adding two momentum terms does not perform any superiorities. Finally, simulation examples are given to illustrate our findings. For the centralized optimization, it is well known that adding one momentum term (also called the heavy-ball method) can obtain a faster convergence rate than the gradient method. However, for the distributed counterpart, there is quite few results about the effect of added momentum terms on the convergence rate. This article is aimed at studying the issue in the distributed setup, where <inline-formula> <tex-math notation="LaTeX">N </tex-math></inline-formula> agents minimize the sum of their individual cost functions using local communication over a network. The cost functions are twice continuously differentiable. We first study the algorithm with one momentum term and develop a distributed heavy-ball (D-HB) method by adding one momentum term on to the distributed gradient algorithm. By borrowing tools from the control theory, we provide a simple convergence proof and an explicit expression of the optimal convergence rate. Furthermore, we consider adding two momentum terms case and propose a distributed double-heavy-ball (D-DHB) method. We show that adding one momentum term allows faster convergence while adding two momentum terms does not perform any superiorities. Finally, simulation examples are given to illustrate our findings. For the centralized optimization, it is well known that adding one momentum term (also called the heavy-ball method) can obtain a faster convergence rate than the gradient method. However, for the distributed counterpart, there is quite few results about the effect of added momentum terms on the convergence rate. This article is aimed at studying the issue in the distributed setup, where [Formula Omitted] agents minimize the sum of their individual cost functions using local communication over a network. The cost functions are twice continuously differentiable. We first study the algorithm with one momentum term and develop a distributed heavy-ball (D-HB) method by adding one momentum term on to the distributed gradient algorithm. By borrowing tools from the control theory, we provide a simple convergence proof and an explicit expression of the optimal convergence rate. Furthermore, we consider adding two momentum terms case and propose a distributed double-heavy-ball (D-DHB) method. We show that adding one momentum term allows faster convergence while adding two momentum terms does not perform any superiorities. Finally, simulation examples are given to illustrate our findings. |
| Author | Liu, Bing Yi, Jingwen Chai, Li |
| Author_xml | – sequence: 1 givenname: Bing orcidid: 0000-0002-6667-9250 surname: Liu fullname: Liu, Bing email: liubing17@wust.edu.cn organization: Engineering Research Center of Metallurgical Automation and Measurement Technology, Wuhan University of Science and Technology, Wuhan, China – sequence: 2 givenname: Li orcidid: 0000-0002-4331-0565 surname: Chai fullname: Chai, Li email: chaili@zju.edu.cn organization: College of Control Science and Engineering, Zhejiang University, Hangzhou, China – sequence: 3 givenname: Jingwen orcidid: 0000-0003-0835-5335 surname: Yi fullname: Yi, Jingwen email: yijingwen@wust.edu.cn organization: Engineering Research Center of Metallurgical Automation and Measurement Technology, Wuhan University of Science and Technology, Wuhan, China |
| BackLink | https://www.ncbi.nlm.nih.gov/pubmed/36355726$$D View this record in MEDLINE/PubMed |
| BookMark | eNp9kctq3DAYhUVJadI0D1AKRZBNNzPRxbotp5MmLSRkYyhdGdn-lSrYUirZDXn7ysxkFllUmyOk7xxdznt0FGIAhD5SsqaUmIt6--vrmhHG1pxRLSV_g04YlXrFmBJHh7lUx-gs5wdShi5LRr9Dx1xyIRSTJ6jZxvAX0j2EDvAm2OE5-4yjw5c-T8m38wQ9vk629xAmfAm5W3Qz3Mfkp99jxj-L4LsA2IYe108R38axIPOIa0hj_oDeOjtkONvrKaqvvtXb76ubu-sf283NquOVmVaqq5wWjvNWOaWBWEGJqGTbaUF1ywWtwBjFbcV62VZOWteSVnIH3ErdEX6KvuxiH1P8M0OemtGXqw6DDRDn3DBVMhQ1VBf0_BX6EOdUXl4owyrCBddL4Oc9Nbcj9M1j8qNNz83LzxWA7oAuxZwTuANCSbMU1CwFNUtBzb6g4lGvPJ2f7ORjmJL1w3-dn3ZODwCHk4ypRNnk_wAFG5vJ |
| CODEN | ITCEB8 |
| CitedBy_id | crossref_primary_10_1109_TCSII_2024_3435066 crossref_primary_10_1109_TAC_2025_3546087 |
| Cites_doi | 10.1090/cbms/092 10.1109/TCNS.2015.2399191 10.1109/cdc.1984.272358 10.1561/2200000016 10.1109/TAC.1983.1103183 10.1109/TSP.2011.2182347 10.1016/j.ifacol.2017.08.1513 10.1109/MSP.2020.2975210 10.1109/TAC.2019.2937496 10.1109/TCOMM.2004.831346 10.1109/TAC.2014.2298712 10.1109/CDC.2009.5400289 10.1109/TAC.2019.2942513 10.1109/TAC.1986.1104412 10.1109/TSP.2016.2544743 10.1017/cbo9781139042918 10.1109/TSP.2014.2304432 10.1109/TCYB.2016.2570808 10.1109/TAC.2008.2009515 10.1007/978-1-4419-8853-9 10.1109/TAC.2010.2041686 10.1016/j.arcontrol.2019.05.006 10.1109/TAC.2017.2730481 10.1109/LCSYS.2017.2722406 10.1109/TAC.2017.2677879 10.1109/TAC.2015.2448011 10.1109/TCYB.2018.2870487 10.1109/TSP.2016.2602803 10.1109/TSP.2011.2146776 10.1016/0041-5553(64)90137-5 10.1137/16M1084316 10.1137/14096668X 10.1109/TCYB.2017.2728644 10.1137/130943170 10.1109/icassp.2016.7472612 10.1109/TCYB.2019.2933003 10.1137/0320018 10.1109/TAC.2013.2253218 10.1109/TCYB.2015.2453167 10.1145/984622.984626 10.1109/TSP.2013.2278149 10.1109/ACC.2000.878579 10.1109/TAC.2010.2091295 10.23919/ACC.2018.8430824 10.1109/tcyb.2020.3011819 |
| 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 NPM 7SC 7SP 7TB 8FD F28 FR3 H8D JQ2 L7M L~C L~D 7X8 |
| DOI | 10.1109/TCYB.2022.3218663 |
| DatabaseName | IEEE All-Society Periodicals Package (ASPP) 2005–Present IEEE All-Society Periodicals Package (ASPP) 1998–Present IEEE Electronic Library (IEL) CrossRef PubMed Computer and Information Systems Abstracts Electronics & Communications Abstracts Mechanical & Transportation Engineering Abstracts Technology Research Database ANTE: Abstracts in New Technology & Engineering Engineering Research Database Aerospace Database ProQuest Computer Science Collection Advanced Technologies Database with Aerospace Computer and Information Systems Abstracts Academic Computer and Information Systems Abstracts Professional MEDLINE - Academic |
| DatabaseTitle | CrossRef PubMed Aerospace Database Technology Research Database Computer and Information Systems Abstracts – Academic Mechanical & Transportation Engineering Abstracts Electronics & Communications Abstracts ProQuest Computer Science Collection Computer and Information Systems Abstracts Engineering Research Database Advanced Technologies Database with Aerospace ANTE: Abstracts in New Technology & Engineering Computer and Information Systems Abstracts Professional MEDLINE - Academic |
| DatabaseTitleList | PubMed MEDLINE - Academic Aerospace Database |
| Database_xml | – sequence: 1 dbid: NPM name: PubMed url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed sourceTypes: Index Database – sequence: 2 dbid: RIE name: IEEE Electronic Library (IEL) url: https://ieeexplore.ieee.org/ sourceTypes: Publisher – sequence: 3 dbid: 7X8 name: MEDLINE - Academic url: https://search.proquest.com/medline sourceTypes: Aggregation Database |
| DeliveryMethod | fulltext_linktorsrc |
| Discipline | Sciences (General) |
| EISSN | 2168-2275 |
| EndPage | 1522 |
| ExternalDocumentID | 36355726 10_1109_TCYB_2022_3218663 9945633 |
| Genre | orig-research Journal Article |
| GrantInformation_xml | – fundername: National Natural Science Foundation of China grantid: 61903284; 62173259 funderid: 10.13039/501100001809 |
| GroupedDBID | 0R~ 4.4 6IK 97E AAJGR AARMG AASAJ AAWTH ABAZT ABQJQ ABVLG ACIWK AENEX AGQYO AGSQL AHBIQ AKJIK AKQYR ALMA_UNASSIGNED_HOLDINGS ATWAV BEFXN BFFAM BGNUA BKEBE BPEOZ EBS EJD HZ~ IFIPE IPLJI JAVBF M43 O9- OCL PQQKQ RIA RIE RNS AAYXX CITATION NPM RIG 7SC 7SP 7TB 8FD F28 FR3 H8D JQ2 L7M L~C L~D 7X8 |
| ID | FETCH-LOGICAL-c349t-7c4f85f33b7f78e0a510546bc8518b3514e9973a42d6b4f6afb0b63fe3a68c03 |
| IEDL.DBID | RIE |
| ISICitedReferencesCount | 4 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000881966000001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| ISSN | 2168-2267 2168-2275 |
| IngestDate | Sun Sep 28 06:46:47 EDT 2025 Sun Nov 30 05:15:03 EST 2025 Mon Jul 21 05:14:32 EDT 2025 Tue Nov 18 21:14:32 EST 2025 Sat Nov 29 02:02:38 EST 2025 Wed Aug 27 02:02:17 EDT 2025 |
| IsPeerReviewed | true |
| IsScholarly | true |
| Issue | 3 |
| 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-c349t-7c4f85f33b7f78e0a510546bc8518b3514e9973a42d6b4f6afb0b63fe3a68c03 |
| Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14 content type line 23 |
| ORCID | 0000-0002-6667-9250 0000-0002-4331-0565 0000-0003-0835-5335 |
| PMID | 36355726 |
| PQID | 2924035380 |
| PQPubID | 85422 |
| PageCount | 12 |
| ParticipantIDs | pubmed_primary_36355726 proquest_miscellaneous_2735171918 crossref_citationtrail_10_1109_TCYB_2022_3218663 proquest_journals_2924035380 crossref_primary_10_1109_TCYB_2022_3218663 ieee_primary_9945633 |
| PublicationCentury | 2000 |
| PublicationDate | 2024-03-01 |
| PublicationDateYYYYMMDD | 2024-03-01 |
| PublicationDate_xml | – month: 03 year: 2024 text: 2024-03-01 day: 01 |
| PublicationDecade | 2020 |
| PublicationPlace | United States |
| PublicationPlace_xml | – name: United States – name: Piscataway |
| PublicationTitle | IEEE transactions on cybernetics |
| PublicationTitleAbbrev | TCYB |
| PublicationTitleAlternate | IEEE Trans Cybern |
| 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 | ref13 ref12 ref11 ref10 ref17 ref16 ref19 ref18 Nesterov (ref36) 1983; 27 ref51 ref50 ref46 ref45 ref48 ref47 ref42 Wilson (ref40) 2018 ref41 ref43 ref49 ref8 ref7 ref9 ref4 ref3 ref6 ref5 Hu (ref44) ref34 ref37 ref31 ref30 ref33 ref32 ref2 ref1 ref39 ref38 Polyak (ref35) 1987 ref24 ref23 Yu (ref15) ref26 ref25 ref20 ref22 ref21 ref28 ref27 ref29 Xie (ref14) 2019 |
| References_xml | – ident: ref50 doi: 10.1090/cbms/092 – ident: ref3 doi: 10.1109/TCNS.2015.2399191 – ident: ref4 doi: 10.1109/cdc.1984.272358 – ident: ref27 doi: 10.1561/2200000016 – ident: ref19 doi: 10.1109/TAC.1983.1103183 – ident: ref25 doi: 10.1109/TSP.2011.2182347 – ident: ref45 doi: 10.1016/j.ifacol.2017.08.1513 – ident: ref17 doi: 10.1109/MSP.2020.2975210 – ident: ref43 doi: 10.1109/TAC.2019.2937496 – ident: ref2 doi: 10.1109/TCOMM.2004.831346 – start-page: 7184 volume-title: Proc. 36th Int. Conf. Mach. Learn. ident: ref15 article-title: On the linear speedup analysis of communication efficient momentum SGD for distributed non-convex optimization – ident: ref42 doi: 10.1109/TAC.2014.2298712 – ident: ref20 doi: 10.1109/CDC.2009.5400289 – ident: ref49 doi: 10.1109/TAC.2019.2942513 – ident: ref5 doi: 10.1109/TAC.1986.1104412 – ident: ref29 doi: 10.1109/TSP.2016.2544743 – ident: ref23 doi: 10.1017/cbo9781139042918 – ident: ref28 doi: 10.1109/TSP.2014.2304432 – ident: ref31 doi: 10.1109/TCYB.2016.2570808 – ident: ref6 doi: 10.1109/TAC.2008.2009515 – ident: ref37 doi: 10.1007/978-1-4419-8853-9 – ident: ref7 doi: 10.1109/TAC.2010.2041686 – ident: ref16 doi: 10.1016/j.arcontrol.2019.05.006 – volume-title: arXiv: 1611.02635v4 year: 2018 ident: ref40 article-title: A Lyapunov analysis of momentum methods in optimization – ident: ref13 doi: 10.1109/TAC.2017.2730481 – volume-title: Introduction to Optimization year: 1987 ident: ref35 – ident: ref38 doi: 10.1109/LCSYS.2017.2722406 – ident: ref30 doi: 10.1109/TAC.2017.2677879 – ident: ref32 doi: 10.1109/TAC.2015.2448011 – ident: ref26 doi: 10.1109/TCYB.2018.2870487 – ident: ref33 doi: 10.1109/TSP.2016.2602803 – ident: ref24 doi: 10.1109/TSP.2011.2146776 – ident: ref34 doi: 10.1016/0041-5553(64)90137-5 – ident: ref48 doi: 10.1137/16M1084316 – ident: ref47 doi: 10.1137/14096668X – ident: ref11 doi: 10.1109/TCYB.2017.2728644 – ident: ref10 doi: 10.1137/130943170 – ident: ref41 doi: 10.1109/icassp.2016.7472612 – ident: ref9 doi: 10.1109/TCYB.2019.2933003 – ident: ref18 doi: 10.1137/0320018 – ident: ref21 doi: 10.1109/TAC.2013.2253218 – ident: ref12 doi: 10.1109/TCYB.2015.2453167 – start-page: 1549 volume-title: Proc. 34th Int. Conf. Mach. Learn. ident: ref44 article-title: Dissipativity theory for Nesterov’s accelerated method – ident: ref1 doi: 10.1145/984622.984626 – ident: ref46 doi: 10.1109/TSP.2013.2278149 – ident: ref51 doi: 10.1109/ACC.2000.878579 – volume-title: arXiv:1911.09030 year: 2019 ident: ref14 article-title: Local adaalter: Communication-efficient stochastic gradient descent with adaptive learning rates – volume: 27 start-page: 372 issue: 2 year: 1983 ident: ref36 article-title: A method for solving a convex programming problem with convergence rate O(1/k²) publication-title: Soviet Math. Doklady – ident: ref8 doi: 10.1109/TAC.2010.2091295 – ident: ref39 doi: 10.23919/ACC.2018.8430824 – ident: ref22 doi: 10.1109/tcyb.2020.3011819 |
| SSID | ssj0000816898 |
| Score | 2.3976398 |
| Snippet | For the centralized optimization, it is well known that adding one momentum term (also called the heavy-ball method) can obtain a faster convergence rate than... |
| SourceID | proquest pubmed crossref ieee |
| SourceType | Aggregation Database Index Database Enrichment Source Publisher |
| StartPage | 1511 |
| SubjectTerms | Algorithms Control theory Convergence Convergence rate Cost function distributed optimization Linear programming Momentum momentum term Network topology Newton method Privacy Routh criterion Signal processing algorithms |
| Title | Convergence Analysis of Distributed Gradient Descent Algorithms With One and Two Momentum Terms |
| URI | https://ieeexplore.ieee.org/document/9945633 https://www.ncbi.nlm.nih.gov/pubmed/36355726 https://www.proquest.com/docview/2924035380 https://www.proquest.com/docview/2735171918 |
| Volume | 54 |
| WOSCitedRecordID | wos000881966000001&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: 2168-2275 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0000816898 issn: 2168-2267 databaseCode: RIE dateStart: 20130101 isFulltext: true titleUrlDefault: https://ieeexplore.ieee.org/ providerName: IEEE |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1Nb9QwEB21FQcuQCnQQFsZiQMgtnXjxB_HdkvhAIVDBMspipNxW6lN0CYLfx-P4404ABKnRIqTOJkZz4w9fg_gheaNMc4boKi09AmK8uMgUoGNyEVlj7XOA3_Klw_q4kIvFubzBryZ9sIgYig-w0M6DWv5TVevaKrsyBjv7oXYhE2l5LhXa5pPCQQSgfo29SczH1WouIh5zM1RMf926pPBND0URMIkiT5HkK9VhKrwm0cKFCt_jzaD1zm__3_9fQD3YnTJTkZ12IYNbB_CdrTfnr2MINOvdqCcU7l52HmJbI1MwjrHzghJl0iwsGHvlqEgbGBnI-gTO7m57JbXw9Vtz776A_vUIqvahhU_O_aRsByG1S0r_GDfP4Li_G0xfz-LZAuzWmRmmKk6czp3QljllEZeUeSVSVv7kExbqvdHY5SosrSRNnOycpZbKRyKSuqai8ew1XYt7gJzXCNKjqnTNkPRaJkr6xCdT61savIE-Pp_l3UEIic-jJsyJCTclCStkqRVRmkl8Hq65fuIwvGvxjskiqlhlEICe2uhltFO-zI1hEfoB32ewPPpsrcwWjapWuxWvo3yX698XqsTeDIqw_TstQ49_fM7n8Fd37NsrFnbg61hucJ9uFP_GK775YFX44U-CGr8C2yM6hI |
| linkProvider | IEEE |
| linkToHtml | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1Lb9QwEB61BalcgFIegQJG4gAV27px4sexbClFbBcOEW1PUZyMoVKboE0W_j4eJxtxACROiRQncTIznhl7_H0ALzWvjHHeAEWhpU9QlB8HkQpsRCoKe6B1GvhTvszUfK7Pz83nNXgz7oVBxFB8hnt0Gtbyq6Zc0lTZvjHe3QuxDjfSJIl5v1trnFEJFBKB_Db2JxMfV6hhGfOAm_1sevHWp4NxvCeIhkkSgY4gb6sIV-E3nxRIVv4ebwa_c3zn_3p8F24P8SU77BViC9awvgdbgwW37NUAM_16G_IpFZyHvZfIVtgkrHHsiLB0iQYLK_Z-EUrCOnbUwz6xw6uvzeKy-3bdsjN_YJ9qZEVdsexnw04JzaFbXrPMD_ftfciO32XTk8lAtzApRWK6iSoTp1MnhFVOaeQFxV6JtKUPyrSlin80RokiiStpEycLZ7mVwqEopC65eAAbdVPjI2COa0TJMXbaJigqLVNlHaLzyZWNTRoBX_3vvBygyIkR4yoPKQk3OUkrJ2nlg7Qi2B1v-d7jcPyr8TaJYmw4SCGCnZVQ88FS2zw2hEjoh30ewYvxsrcxWjgpamyWvo3yX698ZqsjeNgrw_jslQ49_vM7n8PmSXY6y2cf5h-fwC3fy6SvYNuBjW6xxKdws_zRXbaLZ0GZfwE6pexx |
| 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=Convergence+Analysis+of+Distributed+Gradient+Descent+Algorithms+With+One+and+Two+Momentum+Terms&rft.jtitle=IEEE+transactions+on+cybernetics&rft.au=Liu%2C+Bing&rft.au=Chai%2C+Li&rft.au=Yi%2C+Jingwen&rft.date=2024-03-01&rft.pub=IEEE&rft.issn=2168-2267&rft.volume=54&rft.issue=3&rft.spage=1511&rft.epage=1522&rft_id=info:doi/10.1109%2FTCYB.2022.3218663&rft_id=info%3Apmid%2F36355726&rft.externalDocID=9945633 |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=2168-2267&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=2168-2267&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=2168-2267&client=summon |