[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-82055-en":3,"doc-seo-82055-105":29,"detail-sidebar-cat-0-en-105":90},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},82055,7971461740909,"Levi","https://ap-avatar.wpscdn.com/davatar_155a257f0dc6eb9ab79c44ca47cae57d",8,"Research & Report","Wireless Decentralized Federated Learning via Device Clustering and Inter-Cluster Link Enhancement","Decentralized federated learning enables server-free peer-to-peer model exchanges among edge devices, supporting ad-hoc, flexible learning in large device-to-device networks. Wireless implementations converge slowly because many-to-many over-the-air gradient sharing triggers uncoordinated access, substantial errors from stragglers, and sluggish inter-device consensus, especially under pronounced clustering in large D2D settings. The framework partitions nodes into dense clusters and assigns a limited backhaul budget to selected cluster heads, enabling fast intra-cluster aggregation and rare inter-cluster exchanges with O(1/t) convergence. Experiments on image classification confirm accelerated convergence with only a few targeted backhaul links.","Wireless Decentralized Federated Learning via Device Clustering and Inter-Cluster Link  \nEnhancement  \nWilliam Weijia Zheng†, Hang Liu* , and Ying-Jun Angela Zhang††Department of Information Engineering, The Chinese University of Hong Kong, Hong Kong SAR  \n* State Key Laboratory of Internet of Things for Smart City and Department of  \nElectrical and Computer Engineering, University of Macau, Macao S.A.R. [Email: wjzheng@link.cuhk.edu.hk](Email: wjzheng@link.cuhk.edu.hk), [hangliu@um.edu.mo](hangliu@um.edu.mo), [yjzhang@ie.cuhk.edu.hk](yjzhang@ie.cuhk.edu.hk)  \narXiv :2607 .08797v 1 [ cs .IT] 8 Jul 2026  \nAbstract—Decentralized federated learning (DFL) dispenses with the central server of classical FL by utilizing peer-topeer model exchanges among edge devices. This server-free architecture enables ad-hoc, flexible distributed learning in large device-to-device (D2D) networks. However, wireless DFL converges slowly because peer-to-peer model aggregation incurs high delays and errors. Each DFL training round involves many-to-many gradient sharing over wireless channels, resulting in uncoordinated channel access, large communication errors from stragglers, and slow model consensus, especially in largescale D2D networks with pronounced clustering structures. We address these aggregation bottlenecks by provisioning a few reliable backhaul links at straggling nodes to enhance network connectivity. Building on this idea, our budget-aware, clustercentric DFL framework first partitions the network into densely connected clusters, and then allocates the limited backhaul budget to selected cluster heads. The resulting two-tier protocol executes fast, parallel model aggregation within clusters and infrequent inter-cluster exchanges among the heads, yielding an O(1/t) convergence rate in t iterations. Numerical experiments on image-classification tasks confirm that our approach accelerates convergence compared to state-of-the-art DFL baselines with only a few strategically placed backhaul links.  \nIndex Terms—Decentralized federated learning, device-todevice network, gossip algorithm, over-the-air computation, device clustering.  \nI. INTRODUCTION  \nFederated learning (FL) has emerged as a compelling framework for distributed training of artificial-intelligence (AI) models at the network edge, where both data and computation are distributed over edge devices [1] . Classical FL relies on a central server (e.g., a base station or edge server) to coordinate model aggregation: edge devices train local models and periodically upload their models or gradients to the server for global consensus. Although effective, this server-centric architecture becomes impractical whenever device-to-server connections are unavailable, unreliable, or undesirable, particularly in ad hoc networks [2] . Decentralized FL (DFL) overcomes this limitation by replacing the server with peer-to-peer gossip exchanges [3], thereby aligning with the connectivity patterns and privacy requirements of largescale device-to-device (D2D) networks and the Internet-ofThings (IoT) [4] .  \nA growing body of work confirms that model communication and aggregation, rather than local model computation,  \ndominate the runtime of FL systems [5] . Exchanging highdimensional model parameters across rate-limited links is extremely costly for large networks. The challenge is amplified in DFL as every device must alternately transmit to and receive from its neighbors, turning the “many-to-one”uplink of classical FL into “many-to-many” communication [6] . The absence of a coordinating server further complicates decentralized multiple-access control, leading to collisions, long delays, and ultimately slower convergence.  \nTo mitigate the model aggregation bottlenecks, recent research advocates over-the-air (OTA) computation as a scalable solution. By exploiting the signal-superposition property of the wireless multiple-access channel, OTA computation sums gradients “in the air” by transmit scaling","cbCain5lJFlGlqXv","https://ap.wps.com/l/cbCain5lJFlGlqXv","pdf",1912109,1,6,"English","en",105,"# Introduction\n## Federated learning and decentralized FL background\n## Communication bottlenecks in wireless DFL\n## Over-the-air computation for scalable aggregation\n## Challenges and clustering-based two-tier design","[{\"question\":\"Why does wireless decentralized federated learning converge more slowly than server-based FL?\",\"answer\":\"Wireless DFL performs many-to-many gradient sharing over channels without a coordinator, causing collisions, long delays, and aggregation errors from stragglers, which slows consensus and convergence.\"},{\"question\":\"How does the proposed approach use backhaul links to improve aggregation?\",\"answer\":\"It identifies straggling devices and deploys a small set of extra reliable backhaul links between selected cluster heads, improving connectivity and reducing the impact of poorly connected nodes on aggregation error.\"},{\"question\":\"What role does device clustering play in the two-tier protocol?\",\"answer\":\"The network is partitioned into densely connected clusters, enabling fast parallel model aggregation within clusters and infrequent inter-cluster exchanges among heads, leading to an O(1/t) convergence behavior over iterations.\"}]",1784177856,15,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":85,"head_meta":87,"extra_data":89,"updated_unix":27},"wireless-decentralized-federated-learning-via-device-clustering-and-inter-cluster-link-enhancement","",{"@graph":35,"@context":84},[36,53,67],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/wireless-decentralized-federated-learning-via-device-clustering-and-inter-cluster-link-enhancement/82055/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":61,"encodingFormat":60,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-16",true,{"@type":64,"interactionType":65,"userInteractionCount":4},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70,76,80],{"name":71,"@type":72,"acceptedAnswer":73},"Why does wireless decentralized federated learning converge more slowly than server-based FL?","Question",{"text":74,"@type":75},"Wireless DFL performs many-to-many gradient sharing over channels without a coordinator, causing collisions, long delays, and aggregation errors from stragglers, which slows consensus and convergence.","Answer",{"name":77,"@type":72,"acceptedAnswer":78},"How does the proposed approach use backhaul links to improve aggregation?",{"text":79,"@type":75},"It identifies straggling devices and deploys a small set of extra reliable backhaul links between selected cluster heads, improving connectivity and reducing the impact of poorly connected nodes on aggregation error.",{"name":81,"@type":72,"acceptedAnswer":82},"What role does device clustering play in the two-tier protocol?",{"text":83,"@type":75},"The network is partitioned into densely connected clusters, enabling fast parallel model aggregation within clusters and infrequent inter-cluster exchanges among heads, leading to an O(1/t) convergence behavior over iterations.","https://schema.org",{"og:url":51,"og:type":86,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":88,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":91},[92,96,100,104,109,113,118,121,126,129,133],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":93,"show_sort_weight":94,"slug":95},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":97,"show_sort_weight":98,"slug":99},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":101,"show_sort_weight":102,"slug":103},"Exam",70,"exam",{"id":105,"doc_module":4,"doc_module_name":45,"category_name":106,"show_sort_weight":107,"slug":108},5,"Comic",60,"comic",{"id":21,"doc_module":4,"doc_module_name":45,"category_name":110,"show_sort_weight":111,"slug":112},"Technology",50,"technology",{"id":114,"doc_module":4,"doc_module_name":45,"category_name":115,"show_sort_weight":116,"slug":117},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":119,"slug":120},30,"research-report",{"id":122,"doc_module":4,"doc_module_name":45,"category_name":123,"show_sort_weight":124,"slug":125},9,"Religion & Spirituality",20,"religion-spirituality",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":127,"show_sort_weight":124,"slug":128},"World Cup","world-cup",{"id":130,"doc_module":4,"doc_module_name":45,"category_name":131,"show_sort_weight":130,"slug":132},10,"Lifestyle","lifestyle",{"id":134,"doc_module":4,"doc_module_name":45,"category_name":135,"show_sort_weight":105,"slug":136},19,"General","general"]