DEEP MULTI-TASK LEARNING FOR SIMULTANEOUS EMOTION RECOGNITION AND ACTIVITY MONITORING WITH WEARABLE SENSORS

Wearable sensors have become an essential tool in understanding human behavior and mental health by continuously collecting physiological and motion data. Traditionally, emotion recognition and activity monitoring have been treated as separate tasks. However, emotions and activities are closely interrelated in real-world scenarios. In this paper, we propose a Deep Multi-Task Learning (MTL) approach that simultaneously recognizes human emotions and monitors activities using data from wearable sensors. Our model employs a shared feature extraction network with two specialized task-specific branches. We utilize publicly available datasets to evaluate the model and demonstrate that our MTL approach outperforms conventional single-task learning methods in both accuracy and computational efficiency. The proposed method has potential applications in mental health monitoring, fitness tracking, and smart wearable systems.