Learning objectives

• Summarize the shift from task-specific medical AI to generative foundation models.
• Explain how generative models can address data scarcity, long-tail distributions, privacy, and domain shift.
• Describe how structured clinical reasoning supports interpretable and auditable medical AI.
• Discuss pathways toward multimodal medical world models for longitudinal prediction and simulation.

Speaker Bio

Youcheng Li is a Ph.D. candidate at the School of Intelligence Science and Technology, Peking University, advised by Prof. Liwei Wang. His research focuses on medical AI, generative foundation models for medical imaging, interpretable diagnostic reasoning, and multimodal AI. He is a co-first author of BUSGen, a generative foundation model for breast ultrasound, and of the BUS-CoT and MammoExpert clinical reasoning datasets. His work has appeared in Nature Biomedical Engineering, Scientific Data, KDD, and MICCAI. He aims to develop AI systems that translate into clinical practice and advance human health.