(This is a combined model course for the seminar and lab course "Introduction to Bayesian modeling" -- you can enrol if you participate in either or both)
The Bayesian perspective on probability is a cornerstone of modern applied statistics and probabilistic machine learning. Probabilistic models formulated in this framework allow to explicitly communicate and challenge assumptions, perform consistent reasoning, and quantify the uncertainty of predictions — they are a useful tool in data-driven research as well as decision-making.
The seminar aims to explore the conceptual foundations of building these models and employ them for statistical inference and is meant for students without significant prior exposure to the topic and the focus will be on intuitive understanding and rigorous mathematical introduction of probability theory.
This lab course teaches how such models can be implemented in Python — using commonly used modeling packages like Scipy and Pyro — and fit to data for Bayesian inference, uncertainty quantification and prediction.
- Kursleiter/in: Merlin Schüler