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

1. 1Deploy a hybrid centralized-federated architecture to leverage both centralized and federated learning
2. 2Utilize a Transformer-based architecture to recognize human activities from sensor data
3. 3Implement edge devices such as smartphones and wearables to generate private data from wearable and inertial sensors
4. 4Apply the FED-HARGPT approach to facilitate discreet monitoring of human activities, including resting, sleeping, and walking

AI engineers and researchers on a team can benefit from this approach as it enables discreet monitoring of human activities, while data scientists can apply the findings to improve HAR technologies

💡 A hybrid centralized-federated approach can improve the accuracy and privacy of human activity recognition

🤖 FED-HARGPT: A hybrid approach for human activity recognition using Transformers #HAR #AI