JMIR Publications

ABSTRACT

Background:

The rise of wearable sensors marks a pivotal development in the era of affective computing. These sensors, gaining increasing popularity, hold the potential to revolutionize our understanding of human stress. A fundamental aspect within this domain is the ability to discern perceived stress through these unobtrusive devices.

Objective:

This study aims to enhance the performance of emotion recognition utilizing multi-task learning, a technique extensively explored across various machine learning tasks, including affective computing. By leveraging the shared information among related tasks, we seek to augment the accuracy of emotion recognition while confronting the privacy threats inherent in the physiological data captured by these sensors.

Methods:

To address the privacy concerns associated with the sensitive data collected by empathetic sensors, we propose a novel framework that integrates differential privacy and federated learning approaches with multi-task learning. This framework is designed to efficiently identify the user's mental stress while safeguarding their private identity information. Through this approach, we aim to enhance the performance of emotion recognition tasks while preserving user privacy.

Results:

Extensive evaluations of our framework are conducted using two prominent public datasets. The results demonstrate a significant improvement in emotion recognition accuracy, achieving an impressive rate of 90%. Furthermore, our approach effectively mitigates privacy risks, as evidenced by limiting re-identification accuracies to 47%.

Conclusions:

In conclusion, our study presents a promising approach to advancing emotion recognition capabilities while addressing privacy concerns in the context of empathetic sensors. By integrating multi-task learning with differential privacy and federated learning, we have demonstrated the potential to achieve high levels of accuracy in emotion recognition while ensuring the protection of user privacy. This research contributes to the ongoing efforts to harness the power of affective computing in a responsible and ethical manner.