Health/Wellbeing Prediction using multi-modal human data and machine learning

Using behavioral, physiological, and environmental data collected from wearables, mobile phones, and clinical assessment, we develop robust personalize models to predict future health and wellbeing (mood, stress, mental health symptoms, clinical assessment) with machine learning/ deep learning methods.

Deep representation learning of multimodal data

We train deep networks to automatically extract physiological and biobehavioral features from raw data collected by multimodal wearable/mobile sensors to help understand and improve people’s health and wellbeing.


The project aims to measure craving and provide interventions to patients with opioid use disorders. We combine mobile sensing, non-behavioral sensing, as well as fMRI imaging to characterize craving responses.

Epilepsy Measurement and Prediction

This project focuses on measuring and predicting seizure events using non-invasive multimodal sensor data.

Estimation of Human Circadian Phase

This project aims to develop mathematical/machine learning models to measure circadian rhythm using non-invasive physiological and behavioral sensors.

Design just-in time intervention and delivery timing using machine learning

This project aims to develop systems and machine learning models to measure physical and mental health risks and provide interventions for Alzheimer patient spousal caregivers.

Multi-modal sensory feedback for regulating alertness and relaxation

The project aims to design and validate multi-sensory stimulation to regulate alertness or relaxation.

Health and Wellbeing Measurement, Prediction, and Recommendation for shift workers

We develop and test systems to predict health and wellbeing and provide personalized sleep recommendations for shift workers

Personalized dynamic feedback system for health and wellbeing

Combining sensing, modeling / data analysis and intervention modules, we design personalized dynamic feedback system for stress, mental health and sleep management and other health applications.

Unobtrusive Personalized Work Engagement Analysis and Assistant

We analyze how people work using multimodal data and develop an intelligent assistant to help workers engage in their tasks and improve wellbeing.


The project aims to measure and increase understanding of the interactions between living environment, residents’ daily behavior, and their health and well-being in daily life and related to adverse events using mobile sensing and crowd source assessment.

Fostering Positive Emotions and Psycho-Physio Resilience in Job Seekers and Beyond

This project focuses on understanding, combining, and positively affecting emotional experience and psychological and physiological resilience processes in job seeking.