Decentralized Machine Learning in Healthcare: Federated Learning for IoMT and Beyond
2nd International Symposium on AI-Driven Engineering Systems, ISADES 2026, Hybrid, Mbale, Uganda, 19 - 20 Haziran 2026, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Doi Numarası: 10.1109/isades69945.2026.11608204
- Basıldığı Şehir: Hybrid, Mbale
- Basıldığı Ülke: Uganda
- Anahtar Kelimeler: cardiac diagnostics, decentralized learning, differential privacy, ECG analysis, Federated Learning, healthcare privacy, HIPAA compliance, IoMT, medical AI, MIT-BIH database, privacy-preserving machine learning, secure multi-party computation
- Gazi Üniversitesi Adresli: Evet
Özet
Federated Learning (FL) enables collaborative machine learning across institutions without sharing raw patient data, addressing privacy concerns in healthcare. This paper presents an FL framework for ECG arrhythmia classification using the MIT-BIH database, combining differential privacy with Federated Averaging (FedAvg). Our approach achieves a Macro-F1 of 91.77% and Balanced Accuracy of 90.09%, with per-class F1-scores of N: 99.39%, V: 96.75%, S: 86.24%, and F: 84.71%, comparable to centralized models while preserving data privacy. We quantify communication overhead and provide practical deployment insights for healthcare organizations.