Knowledge Caching for Federated Learning in Wireless Cellular Networks

This work examines a novel wireless knowledge caching framework where machine learning models (i.e., knowledge) are cached at local small cell base-stations (SBSs) to facilitate both federated training and access of the models by users. We first consider a single-SBS scenario, where the caching deci...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE transactions on wireless communications 2024-08, Vol.23 (8), p.9235-9250
Hauptverfasser: Zheng, Xin-Ying, Lee, Ming-Chun, Hsu, Kai-Chieh, Hong, Y.-W. Peter
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page 9250
container_issue 8
container_start_page 9235
container_title IEEE transactions on wireless communications
container_volume 23
creator Zheng, Xin-Ying
Lee, Ming-Chun
Hsu, Kai-Chieh
Hong, Y.-W. Peter
description This work examines a novel wireless knowledge caching framework where machine learning models (i.e., knowledge) are cached at local small cell base-stations (SBSs) to facilitate both federated training and access of the models by users. We first consider a single-SBS scenario, where the caching decision, user selection, and wireless resource allocation are jointly determined by minimizing a training error bound subject to constraints on the cache capacity, the communication and computation latency, and the energy consumption. The solution is obtained by first computing the minimum achievable training loss for each model, followed by the optimization of the binary caching variables, which reduces to a 0-1 knapsack problem. The proposed framework is then extended to the multiple-SBS scenario where the user association among SBSs is further examined. We adopt a dual-ascent method where Lagrange multipliers are introduced and updated in each iteration to regularize the dependence among user selection and association. Given the Lagrange multipliers, the caching decision, user selection, resource allocation and user association variables are optimized in turn using a block coordinate descent algorithm. Simulation results show that the proposed scheme can achieve a training error bound that is lower than preference-only and random caching policies in both scenarios.
doi_str_mv 10.1109/TWC.2024.3360642
format Article
fullrecord <record><control><sourceid>proquest_RIE</sourceid><recordid>TN_cdi_ieee_primary_10433131</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>10433131</ieee_id><sourcerecordid>3092924491</sourcerecordid><originalsourceid>FETCH-LOGICAL-c245t-cf06cae829158bb4fe4e827b2fff219ad3e2c99b606e1841fd2b6a03730819723</originalsourceid><addsrcrecordid>eNpNkD1PwzAQhi0EEqWwMzBYYk6xz44TjyiiBVHBUtTRcpJzSQlJsVNV_HsctQO64T70vne6h5BbzmacM_2wWhczYCBnQiimJJyRCU_TPAGQ-flYC5VwyNQluQphyxjPVJpOyPy16w8t1hukha0-m25DXe_pHGv0dsCaLtH6bhw3HV03HlsMgRbYtvvWevqGw6H3X-GaXDjbBrw55Sn5mD-tiudk-b54KR6XSQUyHZLKMVVZzEHzNC9L6VDGJivBOQdc21ogVFqX8QPkueSuhlJZJjLBcq4zEFNyf9y78_3PHsNgtv3ed_GkEUyDBik1jyp2VFW-D8GjMzvffFv_azgzIy0TaZmRljnRipa7o6VBxH9yKQSP8QfPZGUc</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>3092924491</pqid></control><display><type>article</type><title>Knowledge Caching for Federated Learning in Wireless Cellular Networks</title><source>IEEE Electronic Library (IEL)</source><creator>Zheng, Xin-Ying ; Lee, Ming-Chun ; Hsu, Kai-Chieh ; Hong, Y.-W. Peter</creator><creatorcontrib>Zheng, Xin-Ying ; Lee, Ming-Chun ; Hsu, Kai-Chieh ; Hong, Y.-W. Peter</creatorcontrib><description>This work examines a novel wireless knowledge caching framework where machine learning models (i.e., knowledge) are cached at local small cell base-stations (SBSs) to facilitate both federated training and access of the models by users. We first consider a single-SBS scenario, where the caching decision, user selection, and wireless resource allocation are jointly determined by minimizing a training error bound subject to constraints on the cache capacity, the communication and computation latency, and the energy consumption. The solution is obtained by first computing the minimum achievable training loss for each model, followed by the optimization of the binary caching variables, which reduces to a 0-1 knapsack problem. The proposed framework is then extended to the multiple-SBS scenario where the user association among SBSs is further examined. We adopt a dual-ascent method where Lagrange multipliers are introduced and updated in each iteration to regularize the dependence among user selection and association. Given the Lagrange multipliers, the caching decision, user selection, resource allocation and user association variables are optimized in turn using a block coordinate descent algorithm. Simulation results show that the proposed scheme can achieve a training error bound that is lower than preference-only and random caching policies in both scenarios.</description><identifier>ISSN: 1536-1276</identifier><identifier>EISSN: 1558-2248</identifier><identifier>DOI: 10.1109/TWC.2024.3360642</identifier><identifier>CODEN: ITWCAX</identifier><language>eng</language><publisher>New York: IEEE</publisher><subject>Algorithms ; Caching ; Cellular communication ; Computational modeling ; Data models ; Energy consumption ; Federated learning ; Knapsack problem ; Lagrange multiplier ; Loss measurement ; Machine learning ; multi-cell multi-user networks ; Network latency ; Random variables ; Resource allocation ; Resource management ; resource optimization ; Servers ; Training ; Wireless caching networks ; Wireless communication ; Wireless communications ; Wireless networks</subject><ispartof>IEEE transactions on wireless communications, 2024-08, Vol.23 (8), p.9235-9250</ispartof><rights>Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2024</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><cites>FETCH-LOGICAL-c245t-cf06cae829158bb4fe4e827b2fff219ad3e2c99b606e1841fd2b6a03730819723</cites><orcidid>0000-0002-3493-3998 ; 0009-0007-3311-381X ; 0000-0001-7043-8276</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/10433131$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>314,780,784,796,27924,27925,54758</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/10433131$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc></links><search><creatorcontrib>Zheng, Xin-Ying</creatorcontrib><creatorcontrib>Lee, Ming-Chun</creatorcontrib><creatorcontrib>Hsu, Kai-Chieh</creatorcontrib><creatorcontrib>Hong, Y.-W. Peter</creatorcontrib><title>Knowledge Caching for Federated Learning in Wireless Cellular Networks</title><title>IEEE transactions on wireless communications</title><addtitle>TWC</addtitle><description>This work examines a novel wireless knowledge caching framework where machine learning models (i.e., knowledge) are cached at local small cell base-stations (SBSs) to facilitate both federated training and access of the models by users. We first consider a single-SBS scenario, where the caching decision, user selection, and wireless resource allocation are jointly determined by minimizing a training error bound subject to constraints on the cache capacity, the communication and computation latency, and the energy consumption. The solution is obtained by first computing the minimum achievable training loss for each model, followed by the optimization of the binary caching variables, which reduces to a 0-1 knapsack problem. The proposed framework is then extended to the multiple-SBS scenario where the user association among SBSs is further examined. We adopt a dual-ascent method where Lagrange multipliers are introduced and updated in each iteration to regularize the dependence among user selection and association. Given the Lagrange multipliers, the caching decision, user selection, resource allocation and user association variables are optimized in turn using a block coordinate descent algorithm. Simulation results show that the proposed scheme can achieve a training error bound that is lower than preference-only and random caching policies in both scenarios.</description><subject>Algorithms</subject><subject>Caching</subject><subject>Cellular communication</subject><subject>Computational modeling</subject><subject>Data models</subject><subject>Energy consumption</subject><subject>Federated learning</subject><subject>Knapsack problem</subject><subject>Lagrange multiplier</subject><subject>Loss measurement</subject><subject>Machine learning</subject><subject>multi-cell multi-user networks</subject><subject>Network latency</subject><subject>Random variables</subject><subject>Resource allocation</subject><subject>Resource management</subject><subject>resource optimization</subject><subject>Servers</subject><subject>Training</subject><subject>Wireless caching networks</subject><subject>Wireless communication</subject><subject>Wireless communications</subject><subject>Wireless networks</subject><issn>1536-1276</issn><issn>1558-2248</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>RIE</sourceid><recordid>eNpNkD1PwzAQhi0EEqWwMzBYYk6xz44TjyiiBVHBUtTRcpJzSQlJsVNV_HsctQO64T70vne6h5BbzmacM_2wWhczYCBnQiimJJyRCU_TPAGQ-flYC5VwyNQluQphyxjPVJpOyPy16w8t1hukha0-m25DXe_pHGv0dsCaLtH6bhw3HV03HlsMgRbYtvvWevqGw6H3X-GaXDjbBrw55Sn5mD-tiudk-b54KR6XSQUyHZLKMVVZzEHzNC9L6VDGJivBOQdc21ogVFqX8QPkueSuhlJZJjLBcq4zEFNyf9y78_3PHsNgtv3ed_GkEUyDBik1jyp2VFW-D8GjMzvffFv_azgzIy0TaZmRljnRipa7o6VBxH9yKQSP8QfPZGUc</recordid><startdate>20240801</startdate><enddate>20240801</enddate><creator>Zheng, Xin-Ying</creator><creator>Lee, Ming-Chun</creator><creator>Hsu, Kai-Chieh</creator><creator>Hong, Y.-W. Peter</creator><general>IEEE</general><general>The Institute of Electrical and Electronics Engineers, Inc. (IEEE)</general><scope>97E</scope><scope>RIA</scope><scope>RIE</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>7SC</scope><scope>7SP</scope><scope>8FD</scope><scope>JQ2</scope><scope>L7M</scope><scope>L~C</scope><scope>L~D</scope><orcidid>https://orcid.org/0000-0002-3493-3998</orcidid><orcidid>https://orcid.org/0009-0007-3311-381X</orcidid><orcidid>https://orcid.org/0000-0001-7043-8276</orcidid></search><sort><creationdate>20240801</creationdate><title>Knowledge Caching for Federated Learning in Wireless Cellular Networks</title><author>Zheng, Xin-Ying ; Lee, Ming-Chun ; Hsu, Kai-Chieh ; Hong, Y.-W. Peter</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c245t-cf06cae829158bb4fe4e827b2fff219ad3e2c99b606e1841fd2b6a03730819723</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Algorithms</topic><topic>Caching</topic><topic>Cellular communication</topic><topic>Computational modeling</topic><topic>Data models</topic><topic>Energy consumption</topic><topic>Federated learning</topic><topic>Knapsack problem</topic><topic>Lagrange multiplier</topic><topic>Loss measurement</topic><topic>Machine learning</topic><topic>multi-cell multi-user networks</topic><topic>Network latency</topic><topic>Random variables</topic><topic>Resource allocation</topic><topic>Resource management</topic><topic>resource optimization</topic><topic>Servers</topic><topic>Training</topic><topic>Wireless caching networks</topic><topic>Wireless communication</topic><topic>Wireless communications</topic><topic>Wireless networks</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Zheng, Xin-Ying</creatorcontrib><creatorcontrib>Lee, Ming-Chun</creatorcontrib><creatorcontrib>Hsu, Kai-Chieh</creatorcontrib><creatorcontrib>Hong, Y.-W. Peter</creatorcontrib><collection>IEEE All-Society Periodicals Package (ASPP) 2005-present</collection><collection>IEEE All-Society Periodicals Package (ASPP) 1998-Present</collection><collection>IEEE Electronic Library (IEL)</collection><collection>CrossRef</collection><collection>Computer and Information Systems Abstracts</collection><collection>Electronics &amp; Communications Abstracts</collection><collection>Technology Research Database</collection><collection>ProQuest Computer Science Collection</collection><collection>Advanced Technologies Database with Aerospace</collection><collection>Computer and Information Systems Abstracts – Academic</collection><collection>Computer and Information Systems Abstracts Professional</collection><jtitle>IEEE transactions on wireless communications</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Zheng, Xin-Ying</au><au>Lee, Ming-Chun</au><au>Hsu, Kai-Chieh</au><au>Hong, Y.-W. Peter</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Knowledge Caching for Federated Learning in Wireless Cellular Networks</atitle><jtitle>IEEE transactions on wireless communications</jtitle><stitle>TWC</stitle><date>2024-08-01</date><risdate>2024</risdate><volume>23</volume><issue>8</issue><spage>9235</spage><epage>9250</epage><pages>9235-9250</pages><issn>1536-1276</issn><eissn>1558-2248</eissn><coden>ITWCAX</coden><abstract>This work examines a novel wireless knowledge caching framework where machine learning models (i.e., knowledge) are cached at local small cell base-stations (SBSs) to facilitate both federated training and access of the models by users. We first consider a single-SBS scenario, where the caching decision, user selection, and wireless resource allocation are jointly determined by minimizing a training error bound subject to constraints on the cache capacity, the communication and computation latency, and the energy consumption. The solution is obtained by first computing the minimum achievable training loss for each model, followed by the optimization of the binary caching variables, which reduces to a 0-1 knapsack problem. The proposed framework is then extended to the multiple-SBS scenario where the user association among SBSs is further examined. We adopt a dual-ascent method where Lagrange multipliers are introduced and updated in each iteration to regularize the dependence among user selection and association. Given the Lagrange multipliers, the caching decision, user selection, resource allocation and user association variables are optimized in turn using a block coordinate descent algorithm. Simulation results show that the proposed scheme can achieve a training error bound that is lower than preference-only and random caching policies in both scenarios.</abstract><cop>New York</cop><pub>IEEE</pub><doi>10.1109/TWC.2024.3360642</doi><tpages>16</tpages><orcidid>https://orcid.org/0000-0002-3493-3998</orcidid><orcidid>https://orcid.org/0009-0007-3311-381X</orcidid><orcidid>https://orcid.org/0000-0001-7043-8276</orcidid></addata></record>
fulltext fulltext_linktorsrc
identifier ISSN: 1536-1276
ispartof IEEE transactions on wireless communications, 2024-08, Vol.23 (8), p.9235-9250
issn 1536-1276
1558-2248
language eng
recordid cdi_ieee_primary_10433131
source IEEE Electronic Library (IEL)
subjects Algorithms
Caching
Cellular communication
Computational modeling
Data models
Energy consumption
Federated learning
Knapsack problem
Lagrange multiplier
Loss measurement
Machine learning
multi-cell multi-user networks
Network latency
Random variables
Resource allocation
Resource management
resource optimization
Servers
Training
Wireless caching networks
Wireless communication
Wireless communications
Wireless networks
title Knowledge Caching for Federated Learning in Wireless Cellular Networks
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-24T02%3A13%3A50IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_RIE&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Knowledge%20Caching%20for%20Federated%20Learning%20in%20Wireless%20Cellular%20Networks&rft.jtitle=IEEE%20transactions%20on%20wireless%20communications&rft.au=Zheng,%20Xin-Ying&rft.date=2024-08-01&rft.volume=23&rft.issue=8&rft.spage=9235&rft.epage=9250&rft.pages=9235-9250&rft.issn=1536-1276&rft.eissn=1558-2248&rft.coden=ITWCAX&rft_id=info:doi/10.1109/TWC.2024.3360642&rft_dat=%3Cproquest_RIE%3E3092924491%3C/proquest_RIE%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=3092924491&rft_id=info:pmid/&rft_ieee_id=10433131&rfr_iscdi=true