Ensemble Federated Learning for Classifying IoMT Data Streams
Monika Arya, Hanumat Sastry G · 2022 IEEE 7th International conference for Convergence in Technology (I2CT) · 2022
Recent advances in Internet-of-Medical-Things have transformed smart healthcare enabled by artificial intelligence (AI). However, to train classical AI-based machine learning (ML) and deep learning (DL) models, patients' data from individual devices, sensors, and wearables must be uploaded to central servers. This is infeasible in realistic healthcare scenarios as it may result in significant security and privacy issues because of the sensitive nature of healthcare data. Federated Learning (FL) is an emerging distributed collaborative AI paradigm that is particularly attractive for smart healthcare. FL coordinates multiple clients (e.g., hospitals, wearable devices) to train the model without sharing raw data. This paper proposed an ensemble federated learning approach for training a deep learning model for classifying decentralized data streams in IoMT environments. The individual federated models are trained locally, and the ensemble learning is carried out on a server. The results demonstrate that the ensemble federated model outperforms the base federated models.