FED-HARGPT: A Hybrid Centralized-Federated Approach of a Transformer-based Architecture for Human Context Recognition

arXiv cs.AI / 3/27/2026

💬 OpinionIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper proposes a hybrid centralized–federated training approach for Human Activity Recognition (HAR) using a Transformer-based model on mobile wearable/inertial sensor data.
It uses Federated Learning implemented in the Flower framework to train a federated model starting from a centralized baseline, explicitly targeting privacy preservation.
Experiments show the hybrid federated method improves accuracy and robustness compared with alternatives, particularly under non-IID data conditions common in real deployments.
The federated setup achieves performance comparable to centralized models, indicating a feasible trade-off between data privacy and model quality for edge-based HAR.
The work is positioned for real-world monitoring scenarios such as resting, sleeping, and walking, where sensor data is inherently private.

Abstract

The study explores a hybrid centralized-federated approach for Human Activity Recognition (HAR) using a Transformer-based architecture. With the increasing ubiquity of edge devices, such as smartphones and wearables, a significant amount of private data from wearable and inertial sensors is generated, facilitating discreet monitoring of human activities, including resting, sleeping, and walking. This research focuses on deploying HAR technologies using mobile sensor data and leveraging Federated Learning within the Flower framework to evaluate the training of a federated model derived from a centralized baseline. The experimental results demonstrate the effectiveness of the proposed hybrid approach in improving the accuracy and robustness of HAR models while preserving data privacy in a non-IID data scenario. The federated learning setup demonstrated comparable performance to centralized models, highlighting the potential of federated learning to strike a balance between data privacy and model performance in real-world applications.