Scholarlink

We consider autonomous federated learning schemes for scalable Internet of Things (IoT) networks that can enhance distributed intelligence while preserving privacy. First, we evaluate the performance of federated learning in comparison to centralized approaches, considering key Figures of Merit (FoMs), including model accuracy, energy consumption, latency, and vulnerability to adversarial attacks. Through simulations, we demonstrate that federated learning scales well with the number of devices, improving model accuracy while significantly reducing communication overhead and energy consumption across a variety of device types. Additionally, we show the resilience of federated learning against privacy leakage attacks and highlight the effectiveness of advanced secure techniques. These results position federated learning as an emerging paradigm to support intelligent, efficient, and secure IoT systems, providing a foundation for further exploration to meet the growing demands underpinning the IoT.

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

Autonomous Federated Learning Architectures for Scalable IoT Networks: Enhancing Distributed Intelligence with Privacy Preservation

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Arun Kumar Arigela,
  • A. Brahmareddy,
  • Mercy Paul Selvan,
  • T. S. Sreenivas,
  • Nookala Venu,
  • Aftab Ahmed Ansari

摘要

We consider autonomous federated learning schemes for scalable Internet of Things (IoT) networks that can enhance distributed intelligence while preserving privacy. First, we evaluate the performance of federated learning in comparison to centralized approaches, considering key Figures of Merit (FoMs), including model accuracy, energy consumption, latency, and vulnerability to adversarial attacks. Through simulations, we demonstrate that federated learning scales well with the number of devices, improving model accuracy while significantly reducing communication overhead and energy consumption across a variety of device types. Additionally, we show the resilience of federated learning against privacy leakage attacks and highlight the effectiveness of advanced secure techniques. These results position federated learning as an emerging paradigm to support intelligent, efficient, and secure IoT systems, providing a foundation for further exploration to meet the growing demands underpinning the IoT.