[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85429-en":3,"doc-seo-85429-105":29,"detail-sidebar-cat-0-en-105":91},{"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":20,"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},85429,687197100911,"Himbo","https://ap-avatar.wpscdn.com/avatar/a000239b6f1da00475?x-image-process=image/resize,m_fixed,w_180,h_180&k=1782698725881665579",8,"Research & Report","Coding Enforced Robust Secure Aggregation for Federated Learning Under Unreliable Communication","Privacy-preserving federated learning (ppFL) is studied under unreliable communication. Zero-sum privacy noise can protect model privacy while preserving accuracy and easing the privacy–utility trade-off, but unreliable links disrupt the coordination of the noise. This causes unpredictable client participation and aggregation errors that can severely degrade training. A coding-enforced structured secure aggregation scheme, Secure Cooperative Gradient Coding (SecCoGC), is proposed to exploit coding linearity and a target function, yielding an all-or-nothing reconstruction outcome. The work evaluates LDP across layers and introduces fairness as a privacy metric, with a non-convex convergence analysis for binary reconstruction and experiments showing improved performance.","Coding-enforced Robust Secure Aggregation for Federated Learning  \nUnder Unreliable Communication  \nShudi Weng, Graduate Student Member, IEEE, Chao Ren, Member, IEEE, Yizhou Zhao, Member, IEEE, Ming Xiao, Senior Member, IEEE, and Mikael Skoglund, Fellow, IEEE  \narXiv :2507 .07565v 5 [ cs .IT] 12 Jul 2026  \nAbstract—his work studies privacy-preserving federated learning (ppFL) under unreliable communication. Zero-sum privacy noise can protect model privacy without sacrificing model accuracy, effectively overcoming the privacy-utility trade-off. However, unreliable communications can randomly disrupt the coordination of zero-sum noise, resulting in unpredictable client participation and aggregation errors, which can severely impair training effectiveness.his work studies privacypreserving federated learning (ppFL) under unreliable communication. Zero-sum privacy noise can protect model privacy without sacrificing model accuracy, effectively overcoming the privacy-utility trade-off. However, unreliable communications can randomly disrupt the coordination of zero-sum noise, resulting in unpredictable client participation and aggregation errors, which can severely impair training effectiveness.T To address this issue, we exploit the linearity of the coding scheme and the target function to develop a coding-enforced structured secure aggregation method, termed Secure Cooperative Gradient Coding (SecCoGC), which leads to an all-or-nothing outcome under unreliable communication: either exact reconstruction of the global model or a non-meaningful result. We evaluate the local differential privacy (LDP) across all protocol layers in SecCoGC, which accounts for the deterministic network coding scheme, correlations among privacy noise, and random realization of communication networks. Additionally, we present a complete problem formulation and constructions of real-field zero-sum privacy noise, and introduce fairness asa privacy metric among clients. Finally, we provide a distinct non-convex convergence analysis for FL algorithms with binary global model reconstruction. Experimental results demonstrate the superiority of SecCoGC under unreliable communication while maintaining varying levels of privacy preservation, yielding significant performance improvements over benchmarks.  \nIndex Terms—Federated learning, Unreliable communication, Zero-sum noise, Secure aggregation, Straggler mitigation, Gradient coding, Fairness, Local differential privacy.  \nI. Introduction  \nA. Backgrounds  \nFEDERATED LEARNING (FL) enables numerous  \nclients to collaboratively train models based on their local datasets. By only sharing local models (updates), FL enhances user privacy and reduces the volume of data exchanged during training [1] . However, the adversary may still infer sensitive information captured by local models, making FL vulnerable to various model-based attacks, e.g. ,  \nShudi Weng, Ming Xiao, Chao Ren, and Mikael Skoglund are with the Department of Information Science and Engineering (ISE), KTH Royal Institute of Technology, Stockholm, Sweden. Email: {shudiw, mingx, chaor, [skoglund}@kth.se](skoglund}@kth.se).  \nYizhou Zhao is with the College of Electronic and Information Engineering, Southwest University, Chongqing, China. Email:  \n{[onezhou}@swu.edu.cn](onezhou}@swu.edu.cn). Corresponding Author: Shudi Weng.  \ninversion attacks (IA) and membership inference attacks (MIA), etc.  \nTo defend against these and other attacks, a widely adopted strategy in privacy-preserving FL (ppFL) is privacy noise injection to obscure sensitive information, through, e.g., the well-known Gaussian mechanism and Laplacian mechanism. Massive studies investigate the use of independent Gaussian privacy noise with provable convergence guarantee [2] . However, in these settings, the privacy noise is fully preserved in the server aggregation. Strong privacy inevitably leads to significant degradation in model accuracy. This inherent tension between privacy and model ","cbCaiptzRYaILhKQ","https://ap.wps.com/l/cbCaiptzRYaILhKQ","pdf",667059,1,22,"English","en",105,"# Introduction\n## Backgrounds","[{\"question\":\"What problem does the paper address in privacy-preserving federated learning?\",\"answer\":\"It studies ppFL when communication is unreliable, which can disrupt the coordination of privacy noise and lead to unpredictable client participation and aggregation errors that harm training.\"},{\"question\":\"How does Secure Cooperative Gradient Coding (SecCoGC) improve robustness?\",\"answer\":\"SecCoGC enforces secure aggregation using coding linearity and a structured scheme, producing an all-or-nothing outcome under unreliable communication: either exact global model reconstruction or a non-meaningful result.\"},{\"question\":\"How is privacy preservation evaluated in the proposed method?\",\"answer\":\"The paper evaluates local differential privacy (LDP) across all protocol layers, accounting for the deterministic network coding scheme, correlations among privacy noise, and random communication network realizations. It also introduces fairness as a privacy metric among clients.\"}]",1784203416,55,{"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":86,"head_meta":88,"extra_data":90,"updated_unix":27},"coding-enforced-robust-secure-aggregation-for-federated-learning-under-unreliable-communication","",{"@graph":35,"@context":85},[36,53,68],{"@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/coding-enforced-robust-secure-aggregation-for-federated-learning-under-unreliable-communication/85429/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does the paper address in privacy-preserving federated learning?","Question",{"text":75,"@type":76},"It studies ppFL when communication is unreliable, which can disrupt the coordination of privacy noise and lead to unpredictable client participation and aggregation errors that harm training.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does Secure Cooperative Gradient Coding (SecCoGC) improve robustness?",{"text":80,"@type":76},"SecCoGC enforces secure aggregation using coding linearity and a structured scheme, producing an all-or-nothing outcome under unreliable communication: either exact global model reconstruction or a non-meaningful result.",{"name":82,"@type":73,"acceptedAnswer":83},"How is privacy preservation evaluated in the proposed method?",{"text":84,"@type":76},"The paper evaluates local differential privacy (LDP) across all protocol layers, accounting for the deterministic network coding scheme, correlations among privacy noise, and random communication network realizations. It also introduces fairness as a privacy metric among clients.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":45,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":45,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]