Page 81 - AIH-2-4

Artificial Intelligence in Health

ORIGINAL RESEARCH ARTICLE
A hierarchical federated learning-based health

stack for future pandemic preparedness

Rojalini Tripathy 1 , Asmit Balabantaray 2 , Nisarg Shah 2 , Prashant Kumar
Jha 3 , Ajay Kumar Gogineni 1 , Atri Mukhopadhyay 1 , Kisor Kumar Sahu * ,
3,4
and Padmalochan Bera *
1
1 School of Electrical and Computer Sciences, Indian Institute of Technology Bhubaneswar, Odisha,
India
2 School of Mechanical Sciences, Indian Institute of Technology Bhubaneswar, Odisha, India
3 School of Minerals, Metallurgical and Materials Engineering, Indian Institute of Technology
Bhubaneswar, Odisha, India
4 Virtual and Augmented Reality Center of Excellence, Indian Institute of Technology Bhubaneswar,
Odisha, India

Abstract

The COVID-19 pandemic, one of the most disruptive global health crises in recent
history, exposed critical vulnerabilities in existing healthcare infrastructure. Given
*Corresponding authors: the likelihood of future pandemics, it is essential to build a resilient, collaborative,
Kisor Kumar Sahu synergistic, data-driven, and intelligent digital healthcare software. It should be
(kisorsahu@iitbbs.ac.in);
Padmalochan Bera meticulously designed and selectively curated to enhance early detection, rapid
(plb@iitbbs.ac.in) response, and efficient containment of outbreaks. In this article, we propose a federated
learning (FL)-based health stack that prioritizes privacy while fostering collaborative
Citation: Tripathy R,
Balabantaray A, Shah N, intelligence among hospitals or client nodes. Our framework incorporates hierarchical
et al. A hierarchical federated FL, Byzantine-resilient information-theoretic FL (ByITFL), homomorphic encryption,
learning-based health stack for and blockchain-based smart contracts to ensure secure collaboration among
future pandemic preparedness.
Artif Intell Health. 2025;2(4):75-91. healthcare institutions without sharing raw data. Hierarchical FL leverages multilevel
doi: 10.36922/AIH025080013 model aggregation to enhance model convergence, scalability, and resilience. ByITFL
Received: February 21, 2025 strengthens security by incorporating trust mechanisms and information-theoretic
privacy scoring, while blockchain-based smart contracts ensure transparent, verifiable
1st revised: May 22, 2025
coordination among participating nodes. Furthermore, deep vulnerability detection
2nd revised: June 8, 2025 using optimized averaged stochastic gradient descent weight-dropped long short-
Accepted: June 9, 2025 term memory models may further enhance the framework’s security, enabling threat
identification during decentralized data exchanges. Experimental results show that
Published online: June 30, 2025
the proposed hierarchical FL model achieves 94.23% accuracy on the modified
Copyright: © 2025 Author(s). National Institute of Standards and Technology dataset, outperforming federated
This is an Open-Access article averaging (92.66%) under the same environments. In addition, communication
distributed under the terms of the
Creative Commons Attribution analysis proved that the overall transmission is minimized by collecting updates at
License, permitting distribution, and local servers before sending them to central servers. Therefore, it is nearly a future-
reproduction in any medium, which ready technology that can be implemented without many geopolitical issues, even
provided that the original work is
properly cited. in the case of hypersensitive global situations.
Publisher’s Note: AccScience
Publishing remains neutral with Keywords: Global pandemics; Health stack; Federated learning; Medical data privacy;
regard to jurisdictional claims in
published maps and institutional Machine learning
affiliations.

Volume 2 Issue 4 (2025) 75 doi: 10.36922/AIH025080013