Work

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

Year: 2026

Type: article

Abstract: 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.

Source: Discover Internet of Things

DOI: 10.1007/s43926-025-00276-5

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

Institutions: ICFAI Foundation for Higher Education, Manipal Academy of Higher Education, Koneru Lakshmaiah Education Foundation, University of Rwanda, College of Science and Technology - University of Rwanda

Language: English

Topic: Privacy-Preserving Technologies in Data

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

SDG: Industry, innovation and infrastructure

Open Access status: gold

Estimated APC paid: $1,590