Our laboratory is centered on patient-centric medical AI, aiming to build trustworthy and clinically meaningful systems that directly improve patient outcomes. Our research integrates three complementary layers. At the ecosystem level, we design intelligent health systems that support longitudinal health management, personalized interventions, and patient empowerment. At the modeling level, we develop generalist medical decision-making models capable of transferring medical knowledge and clinical reasoning across specialties and care settings. At the foundation level, we advance next-generation medical imaging models based on weakly supervised and self-supervised representation learning to reduce reliance on costly annotations and ensure robustness and transferability in real-world environments. Together, these efforts aim to make high-quality medical AI broadly accessible and genuinely beneficial to every patient.
