[s20] Seminar Graphical Models & Causality (König, Molnar, Dandl))

Dates & Overview

Kickoff: first week of semester, tbd
Block 1: End of May/Beginning June
Block 2: End of June/Beginning July

Key: will be announced at kickoff event

Target groups: Master Statistics, Master Data Science

Course Description

Classical statistical learning excels at learning associations and making accurate predictions. However, many questions like treatment effect estimation are not associative but causal in nature.
Graphical models provide a general framework for describing statistical relations between random variables and performing inference. More recently, they have become a popular framework for reasoning about causal relations.
In this seminar, we will introduce the basics of inference in graphical models, discuss how they can be employed to perform causal inference and peak into recent research on connections to classical machine learning.