Blockchain-Enhanced Differential Privacy Federated Learning Framework (B-DPFL) for Healthcare IoT Applications

Xiaoyi Lin; Guangrong Duan

doi:10.63313/JCSFT.9030

Authors

Xiaoyi Lin Foshan First People's Hospital, Foshan, 528000, China Author
Guangrong Duan Foshan First People's Hospital, Foshan, 528000, China Author

DOI:

https://doi.org/10.63313/JCSFT.9030

Keywords:

Federated Learning, Differential Privacy, Blockchain, Healthcare IoT, Privacy-Preserving Machine Learning, Byzantine Fault Tolerance

Abstract

The proliferation of Internet of Medical Things (IoMT) devices has generated massive amounts of sensitive healthcare data, presenting unprecedented opportunities for machine learning applications while raising critical privacy concerns. This paper proposes a novel Blockchain-Enhanced Differential Privacy Federated Learning Framework (B-DPFL) that integrates federated learning, differential privacy mechanisms, and blockchain technology to enable secure, privacy-preserving collaborative learning in healthcare IoT environments. Our framework addresses three fundamental challenges: (1) preserving patient privacy during model training, (2) ensuring data integrity and model transparency, and (3) preventing malicious attacks in decentralized learning systems. Extensive experiments on real-world medical datasets demonstrate that B-DPFL achieves superior performance with 94.3% accuracy while providing strong privacy guarantees (ε = 1.0) and maintaining Byzantine fault tolerance. Compared to conventional federated learning approaches, our framework reduces privacy leakage by 78.5% and improves model convergence speed by 34.2%.

References

[1] Dwork, C., & Roth, A. (2014). The algorithmic foundations of differential privacy. Founda-tions and trends® in theoretical computer science, 9(3–4), 211-407.

[2] Nakamoto, S., & Bit, B. (2007). Bitcoin: A peer-to-peer electronic cash system. 2008.

[3] Castro, M., & Liskov, B. (1999, February). Practical byzantine fault tolerance. In OsDI (Vol. 99, No. 1999, pp. 173-186).

[4] McMahan, B., Moore, E., Ramage, D., Hampson, S., & y Arcas, B. A. (2017, April). Communi-cation-efficient learning of deep networks from decentralized data. In Artificial intelli-gence and statistics (pp. 1273-1282). PMLR.

[5] Abadi, M., Chu, A., Goodfellow, I., McMahan, H. B., Mironov, I., Talwar, K., & Zhang, L. (2016, October). Deep learning with differential privacy. In Proceedings of the 2016 ACM SIGSAC conference on computer and communications security (pp. 308-318).

[6] Rieke, N., Hancox, J., Li, W., Milletari, F., Roth, H. R., Albarqouni, S., ... & Cardoso, M. J. (2020). The future of digital health with federated learning. NPJ digital medicine, 3(1), 119.

[7] Bonawitz, K., Eichner, H., Grieskamp, W., Huba, D., Ingerman, A., Ivanov, V., ... & Roselander, J. (2019). Towards federated learning at scale: System design. Proceedings of machine learning and systems, 1, 374-388.

[8] Shokri, R., Stronati, M., Song, C., & Shmatikov, V. (2017, May). Membership inference at-tacks against machine learning models. In 2017 IEEE symposium on security and privacy (SP) (pp. 3-18).