Scholarlink

<p>This work introduces a hybrid security framework for Consumer IoT environments that integrates Zero Trust Edge computing with Federated Learning and Neuro-Symbolic AI. The approach employs a Random Forest classifier trained across distributed client nodes using federated averaging to preserve data locality and privacy. Symbolic rules are incorporated to enhance the adaptability of intrusion detection based on network traffic behavior. Evaluated using the BoTNeTIoT-L01 dataset, the framework demonstrates high accuracy in identifying botnet attacks while maintaining privacy-preserving constraints. The results indicate the model’s potential as a scalable and resilient solution for CIoT security.</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

Federated learning and Neuro-Symbolic AI for botnet classification in IoT-Enabled consumer devices

  • 期刊论文

  • 发表日期：

    2025年12月8日

查看全文
  • H. V. Balachandra Achar,
  • P. S. Prasad,
  • D. Rajeshwari,
  • N. Raghu,
  • B. M. Hemanth Kumar,
  • T. R. Mahesh

摘要

This work introduces a hybrid security framework for Consumer IoT environments that integrates Zero Trust Edge computing with Federated Learning and Neuro-Symbolic AI. The approach employs a Random Forest classifier trained across distributed client nodes using federated averaging to preserve data locality and privacy. Symbolic rules are incorporated to enhance the adaptability of intrusion detection based on network traffic behavior. Evaluated using the BoTNeTIoT-L01 dataset, the framework demonstrates high accuracy in identifying botnet attacks while maintaining privacy-preserving constraints. The results indicate the model’s potential as a scalable and resilient solution for CIoT security.