Teaching & AI
As a university lecturer in medical information science, I am particularly committed to making my teaching engaging and practical. My students should be able to find and evaluate trustworthy medical information. Transparency and replicability play a crucial role in evidence synthesis - after all, medical treatments depend on them. To meet these goals, it is also essential to understand how AI systems for information retrieval work. My conviction: Only when you understand what the machine does can you properly contextualize its output.
Here you can find my human-authored educational materials for medical information science, including human-designed diagrams and graphics on AI retrieval systems. All content is protected by copyright.
Teaching Medical Information Science
To efficiently plan my teaching, I created a linked mind map covering the different phases (in German).
- Conceptualization of teaching
- Content creation (to come)
- Methods of teaching (to come)
- Student supervision (to come)
- Assessement (to come)
- Quality assurance (to come)
AI Systems for Medical Information Retrieval
- Lexical search (the foundation)
- Tokenziation & embedding (first steps in AI-powered information retrieval)
- Semantic search & vector search
- Classic RAG (Retrieval Augmented Generation)
- Agentic RAG (simplified)
- MCP (Model Context Protocol) vs API (Application Programming Interface)
- Machine Learning / Active Learning
- Evidence Cycle
- Distorted Evidence Cycle
- Error Perpetuation in Evidence Synthesis