<p>The integration of AI, IoT, and edge–cloud computing is accelerating smart industrial system improvements, particularly in healthcare and finance. This paper presents TrustFed, a secure and privacy-preserving federated AI platform, to address IIoT data privacy, security, and scalability issues. TrustFed uses Intel SGX–based trusted execution, Federated Deep Learning (FDL), Differential Privacy (DP), PCA-driven feature reduction, and encryption-based secure aggregation for decentralised model training confidentiality and robustness. Two privacy-aware face recognition and brain tumour classification use cases verify the system, showing better accuracy, reduced communication overhead, and robustness to inference and poisoning assaults. TrustFed improves data privacy and performance, adding scientific value to secure AI adoption in large-scale smart industrial environments.</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

Trustfed a scalable privacy preserving federated AI framework for industrial IoT healthcare and finance

  • Dileep Kumar Murala,
  • K. Madhura,
  • Veera Ankalu Vuyyuru,
  • K. Vara Prasada Rao,
  • Eric Hitimana

摘要

The integration of AI, IoT, and edge–cloud computing is accelerating smart industrial system improvements, particularly in healthcare and finance. This paper presents TrustFed, a secure and privacy-preserving federated AI platform, to address IIoT data privacy, security, and scalability issues. TrustFed uses Intel SGX–based trusted execution, Federated Deep Learning (FDL), Differential Privacy (DP), PCA-driven feature reduction, and encryption-based secure aggregation for decentralised model training confidentiality and robustness. Two privacy-aware face recognition and brain tumour classification use cases verify the system, showing better accuracy, reduced communication overhead, and robustness to inference and poisoning assaults. TrustFed improves data privacy and performance, adding scientific value to secure AI adoption in large-scale smart industrial environments.