AI-Powered Virtual Patients for Healthcare
Project Description
This project aims to explore the emerging direction of AI-powered virtual patients, where computational models are developed to simulate patient characteristics, behaviors, clinical states, and responses to different interactions or interventions. The student will work closely with the faculty supervisor and PhD students to investigate potential applications of virtual patients in healthcare.
Supervisor
CHEN, Hao
Quota
10
Course type
UROP1000
UROP1100
UROP2100
UROP3100
UROP3200
UROP4100
Applicant's Roles
Possible research directions include simulation-based medical education and clinical reasoning training, virtual patient cohorts for in-silico clinical trials, patient-specific disease progression and treatment simulation, and personalized treatment optimization and clinical decision support. Depending on the student's background and progress, the project will focus on one or more concrete research questions within these areas.
Applicant's Learning Objectives
Through the project, the student will learn how to conduct literature reviews and formulate research questions, and will gain hands-on experience with modern AI techniques relevant to virtual patient modeling, such as medical multimodal foundation models / vision-language models, generative modeling, world-model-inspired patient simulation, and agentic AI. The student will also learn practical research skills including medical data processing, model implementation and training, experimental design, quantitative evaluation, and analysis of model behaviors and limitations.

For students making strong progress, the project may further involve developing a research prototype, designing new methods or evaluation frameworks, and contributing to the preparation of a research manuscript.
Complexity of the project
Moderate