Trustworthy Federated Learning Framework for Privacy-Preserving Artificial Intellgence Healthcare Sector

• Author(s): Rejoice Kelechi Uzodinma ; Francis Chigozie Emmanuel ; Onwuka Ezenwa Julius
• Paper ID: 1709249
• Page: 1214-1222
• Published Date: 23-06-2025
• Published In: Iconic Research And Engineering Journals
• Publisher: IRE Journals
• e-ISSN: 2456-8880
• Volume/Issue: Volume 8 Issue 12 June-2025

Abstract

This study introduces a privacy-preserving federated learning (FL) framework tailored for artificial intelligence (AI) healthcare environment. This Federated learning framework allows collaborative model training throughout decentralized organizations without revealing sensitive patient data. It incorporates aggregation and differential privacy to ensure regulatory compliance with the Health Insurance Portability and Accountability Act (HIPAA), General Data Protection Regulation (GDPR) and Nigeria Data Protection Regulation (NDPR). In addressing client diversity, this framework follows adaptive mechanisms, model compression, and asynchronous updates, which ensures communication efficiency and scalability. The framework is also resilient against poisoning attacks through a security approach. Evaluating this study is based on model accuracy, communication cost, and resistance to adversarial threats. Overall, this study shows that privacy, performance, and scalability can coexist in healthcare artificial intelligence (AI) and can provide a foundation for real-world applications.

Keywords

Federated Learning, Privacy-Preserving, Healthcare AI, Scalability, Machine Learning.

Citations

IRE Journals:
Rejoice Kelechi Uzodinma , Francis Chigozie Emmanuel , Onwuka Ezenwa Julius "Trustworthy Federated Learning Framework for Privacy-Preserving Artificial Intellgence Healthcare Sector" Iconic Research And Engineering Journals Volume 8 Issue 12 2025 Page 1214-1222

IEEE:
Rejoice Kelechi Uzodinma , Francis Chigozie Emmanuel , Onwuka Ezenwa Julius "Trustworthy Federated Learning Framework for Privacy-Preserving Artificial Intellgence Healthcare Sector" Iconic Research And Engineering Journals, 8(12)