SCALE: Self-regulated Clustered federAted LEarning in a Homogeneous Environment
School of Computing(计算机学院);Southern Illinois University Carbondale(南伊利诺伊大学卡本代尔分校);The University of Texas at Arlington(德克萨斯大学阿灵顿分校);University of Illinois Urbana-Champaign(伊利诺伊大学厄巴纳-香槟分校)
联邦学习(FL)已成为一种变革性方法,能够在保护用户隐私的同时实现分布式机器学习,但它面临着通信效率低下和依赖中心化基础设施等挑战,导致延迟和成本增加。本文提出了一种新颖的 FL 方法,通过消除对边缘服务器的依赖来克服这些限制,采用服务器辅助的 Proximity Evaluation,基于数据相似性、性能指数和地理邻近性进行动态集群形成。我们的集成方法通过 Hybrid Decentralized Aggregation Protocol 提高了运行效率和可扩展性,该协议将本地模型训练与对等权重交换以及由动态选举的驱动节点管理的中心化最终聚合相结合,显著减少了全局通信开销。此外,该方法包括 Decentralized Driver Selection、用于减少网络流量的 Check-pointing 以及用于系统鲁棒性的 Health Status Verification Mechanism。使用乳腺癌数据集进行验证,我们的架构不仅展示了通信开销近十倍的减少,还在保持高学习性能的同时,在减少训练延迟和能耗方面显示出显著改善,为未来联邦学习生态系统提供了可扩展、高效且保护隐私的解决方案。
英文摘要:
Federated Learning (FL) has emerged as a transformative approach for enabling distributed machine learning while preserving user privacy, yet it faces challenges like communication inefficiencies and reliance on centralized infrastructures, leading to increased latency and costs. This paper presents a novel FL methodology that overcomes these limitations by eliminating the dependency on edge servers, employing a server-assisted Proximity Evaluation for dynamic cluster formation based on data similarity, performance indices, and geographical proximity. Our integrated approach enhances operational efficiency and scalability through a Hybrid Decentralized Aggregation Protocol, which merges local model training with peer-to-peer weight exchange and a centralized final aggregation managed by a dynamically elected driver node, significantly curtailing global communication overhead. Additionally, the methodology includes Decentralized Driver Selection, Check-pointing to reduce network traffic, and a Health Status Verification Mechanism for system robustness. Validated using the breast cancer dataset, our architecture not only demonstrates a nearly tenfold reduction in communication overhead but also shows remarkable improvements in reducing training latency and energy consumption while maintaining high learning performance, offering a scalable, efficient, and privacy-preserving solution for the future of federated learning ecosystems.
Novel clustered federated learning based on local loss
LMIB, School of Mathematical Sciences, Beihang University(力学系,数学科学学院,北航);College of Science, China Agricultural University(中国农业大学,理学院);School of Mathematical Sciences, Nanjing Normal University(南京师范大学,数学科学学院);School of Mathematical Sciences, Beihang University(力学系,数学科学学院,北航)
DAG-ACFL: Asynchronous Clustered Federated Learning based on DAG-DLT
Information Countermeasure Technique Institute(信息对抗技术研究所);School of Cyberspace Science(网络空间科学学院);Faculty of Computing(计算学部);Harbin Institute of Technology(哈尔滨工业大学)
On the Convergence of Clustered Federated Learning
Australian Artificial Intelligence Institute, University of Technology Sydney(悉尼科技大学澳大利亚人工智能研究所);University of Washington, Seattle(西雅图华盛顿大学);University of Maryland, College Park(马里兰大学帕克分校)
Clustered Federated Learning: Model-Agnostic Distributed Multi-Task Optimization under Privacy Constraints
Fraunhofer Heinrich Hertz Institute(弗劳恩霍夫海因里希·赫兹研究所);Technische Universität Berlin(柏林工业大学);Max Planck Institute for Informatics(马克斯·普朗克信息学研究所);Korea University(高丽大学)
FedSCAM (Federated Sharpness-Aware Minimization with Clustered Aggregation and Modulation): Scam-resistant SAM for Robust Federated Optimization in Heterogeneous Environments
Department of Computer Science, Lahore University of Management Sciences, Lahore, Pakistan(计算机科学系,拉合尔管理科学大学,拉合尔,巴基斯坦)
Fedcompass: Federated Clustered and Periodic Aggregation Framework for Hybrid Classical-Quantum Models
1 School of Artificial Intelligence;Computer Science, Jiangnan University 2 Antai College of Economics \& Management, Shanghai Jiao Tong University 3 School of Computer Science, Shanghai Jiao Tong University 4 Cloud Computing;Distributed Systems Laboratory, University of Melbourne
Single-Round Clustered Federated Learning via Data Collaboration Analysis for Non-IID Data
Graduate School of Science and Technology, University of Tsukuba(茨川大学理学技术研究生院);Center for Artificial Intelligence Research, Tsukuba Institute for Advanced Research, University of Tsukuba(茨川大学先进研究所人工智能研究中心);Institute of Systems and Information Engineering, University of Tsukuba(茨川大学系统与信息工程研究所)
Auction Based Clustered Federated Learning in Mobile Edge Computing System
School of Computer Science and Technology, Harbin Institute of Technology(哈尔滨工业大学计算机科学与技术学院);Cyberspace Security Research Center, Peng Cheng Laboratory(鹏城实验室网络空间安全研究中心);School of Electrical and Data Engineering, University of Technology Sydney(悉尼科技大学电气与数据工程学院);Department of Computer Science and Technology, Fuzhou University(福州大学计算机科学与技术学院);Department of Computer Science, Electrical and Space Engineering, Lulea University of Technology(吕勒奥理工大学计算机、电气与空间工程系)
Enhancing Clustered Federated Learning: Integration of Strategies and Improved Methodologies
School of Science and Engineering, The Chinese University of Hong Kong, Shenzhen(香港中文大学(深圳)理工学院);Shenzhen Institute of Artificial Intelligence and Robotics for Society (AIRS)(深圳市人工智能与机器人研究院);School of Engineering, Westlake University(西湖大学工学院);Research Center for Industries of the Future, Westlake University(西湖大学未来产业研究中心)
Flexible Clustered Federated Learning for Client-Level Data Distribution Shift
Chongqing University(重庆大学);College of Computer Science, Chongqing University(重庆大学计算机学院);School of Microelectronics and Communication Engineering, Chongqing University(重庆大学微电子与通信工程学院)
Clustered Scheduling and Communication Pipelining For Efficient Resource Management Of Wireless Federated Learning
Texas A&M University(德克萨斯农工大学);Texas A&M University at Qatar(德克萨斯农工大学卡塔尔分校);Qatar Foundation Research, Development and Innovation RDI(卡塔尔基金会研究、开发与创新(RDI));Tennessee Tech University(田纳西理工大学)
Automated Byzantine-Resilient Clustered Decentralized Federated Learning for Battery Intelligence in Connected EVs
SAMOVAR, Télécom SudParis(SAMOVAR,法国电信南巴黎学院);Department of Computer Science, German University of Technology in Oman (GUtech)(阿曼技术大学计算机科学系);L3i, La Rochelle University(拉罗什大学L3i)
Distilling A Universal Expert from Clustered Federated Learning
Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education(教育部符号计算与知识工程重点实验室);College of Computer Science and Technology, Jilin University(吉林大学计算机科学与技术学院);Australian Artificial Intelligence Institute, FEIT, University of Technology Sydney(悉尼大学技术学院澳大利亚人工智能研究所);College of Computer Science, Inner Mongolia University(内蒙古大学计算机科学学院)