Audio-based Mobile Health Diagnosis

Machine learning (ML) for disease tracking

This project aims to explore the potential of audio signals for disease detection and tracking, including respiratory diseases, emotion disorders, and mental health. By analyzing audio characteristics, this project aims to push the boundaries of audio signals for remote health monitoring and continuous forecasting of disease progression ahead of time for timely intervention.

Emotion Intelligence in Conversational AI

Exploring and enhancing the understanding of human emotions in human-computer conversational interactions

In addition to advancing cognitive intelligence in conversational AI, this project aims to explore and enhance these systems with emotional intelligence capable of understanding and responding to human emotions. Specifically, the project will focus on the complex, inherently ambiguous, and nuanced nature of human emotions, alongside the development of parallel AI systems for genuine emotional understanding. By integrating advanced speech emotion recognition techniques with natural language processing, we endeavor to create AI assistants that are more empathetic and contextually aware.

Time Series Modelling and Deep Learning

Representation learning for real-world health time series

Time series data, including ECGs and EEGs, and acoustic signals, play a crucial role in health monitoring by providing essential insights through their continuous and dynamic characteristics. This research project seeks to investigate the capabilities and limitations of self-supervised representation learning in the analysis of real-world health time series data, particularly in the presence of distributional shifts such as data missingness and motion artifacts. By developing robust and reliable self-supervised learning models, this project aspires to enhance the accuracy and dependability of health monitoring systems for practical, real-world applications.

Earable Sensing

Multimodal sensing for continuous monitoring of vital signs

This project aims to explore the potential of earable devices for continuous monitoring of vital signs. By leveraging multimodal sensing technologies integrated into earable devices, we seek to develop non-invasive methods for tracking various physiological parameters. The research focuses on overcoming challenges related to signal quality, power efficiency, and user comfort, while ensuring accurate and reliable health monitoring in everyday settings.