Scholarlink

<p>The advent of beyond fifth generation (B5G) network promises to revolutionize the internet of things (IoT) by enabling seamless, ultra-reliable and low latency connectivity across a vast array of devices. Federated learning emerges as an essential technology in this landscape, allow a distributed collaboratively train a machine learning models over users private datasets without sharing raw data with a central server, thereby it improves data privacy. However the integration of federated learning in B5G-based IoT networks operates centrally and introduces significant security and privacy challenges like model privacy leakage, single point of failure, malicious attacks that can reconstruct users’ private data by exploiting model parameters during model training and transmission process. This survey paper concentrate on the exploration of the current state of security and privacy models which are designed for the protecting FL operation within B5G-based IoT network environment. Therefore we conducted a systematic investigation to identify the prevalent threats and privacy taxonomy against federated leaning in B5G-enabled IoT network. Our research analysis highlights the strength and weakness of these frameworks and identifies key trends and gaps in the research. Consequently, we also proposed a framework that essential for ensuring the robust security and privacy in these advanced networks. The finding of survey provides a comprehensive understanding of the security and privacy landscape for FL in B5G- based IoT networks, offering insights for future research and development in this critical area.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Enhanced Security and Privacy Framework for Federated Learning in Beyond 5G IoT Networks

  • 期刊论文

  • 发表日期：

    2025年9月1日

查看全文
  • Nasir Ahmad Jalali,
  • Chen Hongsong,
  • Farid Ahmad Zahin,
  • Aliyu Ashiru

摘要

The advent of beyond fifth generation (B5G) network promises to revolutionize the internet of things (IoT) by enabling seamless, ultra-reliable and low latency connectivity across a vast array of devices. Federated learning emerges as an essential technology in this landscape, allow a distributed collaboratively train a machine learning models over users private datasets without sharing raw data with a central server, thereby it improves data privacy. However the integration of federated learning in B5G-based IoT networks operates centrally and introduces significant security and privacy challenges like model privacy leakage, single point of failure, malicious attacks that can reconstruct users’ private data by exploiting model parameters during model training and transmission process. This survey paper concentrate on the exploration of the current state of security and privacy models which are designed for the protecting FL operation within B5G-based IoT network environment. Therefore we conducted a systematic investigation to identify the prevalent threats and privacy taxonomy against federated leaning in B5G-enabled IoT network. Our research analysis highlights the strength and weakness of these frameworks and identifies key trends and gaps in the research. Consequently, we also proposed a framework that essential for ensuring the robust security and privacy in these advanced networks. The finding of survey provides a comprehensive understanding of the security and privacy landscape for FL in B5G- based IoT networks, offering insights for future research and development in this critical area.