Deep Learning in Economics

The seminar Deep Learning in Economics provides an introduction to the foundations and applications of deep learning in quantitative economics. At the beginning of the course, key concepts in linear algebra, Python programming, and economic modeling are systematically reviewed to ensure a common methodological starting point. Subsequently, the structure and functioning of neural networks as well as central training and optimization procedures are covered. The methodological content is reinforced through practical exercises in Python, in which models are implemented, trained, and evaluated. In the final stage of the seminar, students independently apply the learned methods to an economic research question and critically reflect on the results as well as on the limitations of the approach compared to classical methods.

Learning Goals

After successfully completing the seminar, students are able to conceptually understand deep learning methods and systematically integrate them into economic research questions. They can independently implement, train, and evaluate neural networks in Python, analyze economic datasets using deep-learning models, and critically interpret the results within an economic context. Furthermore, students are capable of conducting their own empirical projects in the field of Deep Learning & Economics, documenting them in a methodologically sound manner, and presenting their findings in an appropriate academic format.