Repository for University of Glasgow tutorial material and website
Recent advances in medical multimodal large language models (MLLMs) have enabled integration of diverse data types—clinical reports, images, and knowledge graphs—within unified, interpretable frameworks for healthcare AI. This tutorial covers concepts, real-world applications, and hands-on techniques to adapt these models for use in clinical practice.
| Time | Topic | Speaker | Duration |
|---|---|---|---|
| 10:00am | Era of Large Language Models Impact and limitations of LLMs Open-source models |
Honghan Wu | 30 mins |
| 10:30am | Error Detection & Fact Checking with Knowledge Graphs Knowledge graph integration Enhancing MLLMs |
Honghan Wu | 30 mins |
| 11:00am | Break | 15 mins | |
| 11:15am | Integrating LLMs & Knowledge Graphs for EHRs EHRs, contextual reasoning |
Yusuf Abdulle | 30 mins |
| 11:45am | Unifying Multimodal Clinical & Genomic Data Knowledge graph-driven precision medicine |
Yusuf Abdulle | 30 mins |
| 12:15pm | Lunch Break | 60 mins | |
| 13:15pm | Expanding LLMs to Vision: Multimodal LLMs | Yunsoo Kim | 30 mins |
| 13:45pm | Multimodal LLMs for Radiology Data construction Visual instruction tuning Reasoning enhancement Applications: report generation, diagnostics, error checking |
Jinge Wu | 90 mins |
| 15:15pm | Break | 15 mins | |
| 15:30pm | Multimodal LLMs for Pathology Patch and WSI models Report generation & metrics |
Yunsoo Kim | 30 mins |
- Yunsoo Kim PhD candidate, UCL; former senior NLP researcher, LG Chem. Organizer at HealTac2024, MICCAI2024/2025. Focus: human-expert enhancement in MLLMs.
- Jinge Wu Research fellow, IHI UCL; specializes in radiology MLLMs, published in Nature Biomedical Engineering, Nature Reviews Bioengineering, The Lancet Digital Health.
- Yusuf Abdulle PhD candidate, King's College London; researches EHRs, knowledge graphs, imaging, genomics in ALS/MND, NHS, and startups.
- Prof. Honghan Wu Professor of Health Informatics & AI, University of Glasgow. Co-leads SHW Data Science, leads UCL KnowLab. Expertise: health informatics and LLMs.