Learning objectives

1. Learn the paradigm shift and scaling up of medical AI from single-domain, single-site model into generalist, multi-site foundation models.
2. Understand the concept of medical agentic AI: how traditional ML/AI models equipped with LLM can collaborate with each other.
3. Learn how generative AI can facilitate the alignment between human experts and AI system to formulate expert-in-the-loop design.
4. How to train reasoning LLM via reinforcement learning and how such “meta-agent” can facilitate the collaboration among agents.

Speaker biography

Xiang Li is an Assistant Professor of Radiology at Harvard Medical School and Massachusetts General Hospital. He has long been dedicated to the research and application of medical artificial intelligence. He leads multiple research projects focused on medical imaging, medical text analysis, and multimodal data integration, focusing on developing artificial general intelligence solutions tailored for complex clinical scenarios. Prof. Li has made extensive contributions in areas such as medical image and text analysis, AI-assisted disease diagnosis and detection, generative AI, and computational frameworks for medical big data. He has published over 160 papers in top journals and conferences, including Nature Medicine, IEEE TPAMI, NeurIPS, ICLR, and ICML, with more than 12,000 citations and an h-index of 50. He serves on the editorial boards/area chairs of several international journals and conferences, including IEEE TMI, IEEE TAI, NeurIPS, AAAI, and MICCAI. His research has been supported by the U.S. National Institutes of Health (NIH) and focuses on topics such as large language models in medicine, multimodal data fusion, and generative AI for screening support. Professor Li has received numerous honors and awards, including the Google Scholar Program Award, the Thrall Innovation Grants Award, the NVIDIA Global Impact Award, and multiple Best Paper Awards from leading journals and conferences.