This intensive, hands-on course introduces PhD and advanced Master's students in Biomedical Informatics to Agent-Based Modeling (ABM) as a methodology for studying complex, dynamic, and heterogeneous systems in health and biomedicine. The course places particular emphasis on integrating ABMs with foundation models. Through a combination of lectures, lab sessions, and group projects, students explore the theoretical underpinnings of ABMs, implementation strategies using NetLogo, and the application of ABMs to real-world biomedical problems, with a running case study of medication adherence for Type 2 diabetes in an underserved urban neighborhood. The course is divided into three parts:
· Part I — Introduction to ABMs (Sessions 1–2): foundations, design, NetLogo implementation, calibration, and validation.
· Part II — Foundation Models to Enhance ABMs (Session 3): using LLMs and other foundation models for parameterization, synthetic population generation, and output interpretation.
· Part III — Agentic AI (Sessions 4–5): building LLM-driven agents, evaluating agentic ABMs, and presenting final projects.
By the end of the course, students will be able to design, implement, validate, and interpret ABMs in biomedical contexts and articulate when (and when not) to use foundation-model integrations.
- Profesor: JOHN HEISLER HOLMES