This is a mini course lectured by Prof. Thiemo Fetzer from Warwick University. This course introduces the core principles behind large language models (LLMs), with an emphasis on how they process information, generate language, and can be deployed locally for research use. Drawing analogies to human information processing, the course builds intuition for model behavior while clarifying key differences between statistical prediction and human reasoning. Throughout the course, concrete applications and research architectures illustrate how locally deployed LLMs can be responsibly integrated into applied economics research.

This course interweave the following objectives: 1. Provide a brief introduction to large language models 2. Make connections to how we process information as humans and draw some similarities 3. Interweave concrete applications and architectures for applied economics research 4. Zoom out to provide a running commentary

Grades will be determined by classroom participation (50%) and a final report (50%). The final report is due on June 5. Students are required to apply the methods covered in this course to an empirical application using data from an economic study of their choosing. The report must not exceed five pages. It should clearly articulate the motivation for the analysis, describe how large language models may be used to support the study, and present and discuss the empirical results

https://www.trfetzer.com/ai-for-applied-economics/

https://www.trfetzer.com/ai-for-applied-economics/