Our 1st talk of 2026, on Causal Inference, will be held on Wednesday the 14th of January at 1pm (GMT). This will be a hybrid event with the speaker online using MS Teams (
link) and the audience in Room 02/008 of the Peter Froggatt Centre (on the main quadrangle) of QUB (campus
map) .
Dr. Lucas Kook of
Vienna University of Economics and Business (WU Wien), Austria, who specializes in statistical methods like conditional independence testing for causal inference, machine learning, and various types of data structures, will give the talk entitled.
In statistics, the concept of causality has been formalized in numerous ways, including potential outcomes, causal graphical models, and structural causal models. What fundamentally distinguishes such causal models from purely statistical models is their ability to describe how a system behaves under interventions. In many settings, interventions are infeasible, forcing scientists to rely on purely observational data to draw causal conclusions. Doing so, however, requires strong and typically untestable assumptions. This makes it imperative to be able to critically assess both the causal model itself and any conclusions drawn from it. In this talk, I will first provide an introduction to graphical approaches to causal inference. I will then focus on methods for criticizing, falsifying, and probing the sensitivity of causal models. All welcome! See previous talks
here. Read previous write-ups
here. Follow RSSNI on
Linkedin