Vera Haugen Kvisgaard is a PhD Candidate at the University of Oslo and part of the Statistics and Data Science group. The aim of her phd-project is to develop methods for computational causal inference within the probabilistic graphical models framework. Our focus will be on Bayesian methods for estimation of causal effects under an unknown causal structure, building upon the Bayesian IDA (BIDA, Pensar (2020)). It is in general not possible to identify a unique causal structure from observed data, and thereby not a unique causal effect.
The BIDA method accounts for the uncertainty about the underlying structure using Bayesian model averaging, averaging the causal effect over different structures. Our objective is to extend the scope of this approach to new settings, and thus increasing its usability for practical applications. The current BIDA algorithm assumes that the underlying model is linear Gaussian, implying a multivariate normal joint distribution of the variables. The first step, which this presentation will focus on, is therefore to generalize the method to categorical variables
Program
11:30 – Doors open and lunch is served
12:00 – "Bayesian estimation of causal effects using directed graphical models" by Vera Haugen Kvisgaard (PhD Candidate, Department of Mathematics)
This event is open for all PhD candidates and postdocs. No registration needed.
About the seminar series
Once a month, dScience will invite you to join us for lunch, soft drinks and professional talks at the Science Library. In addition to these, we will serve lunch to PhD candidates in our lounge in Kristine Bonnevies hus every Thursday. Due to limited space (40 people), this will be first come, first served. See how to find us here (download).
Our lounge can also be booked by PhDs and Postdocs on a regular basis, whether it is for a meeting or just to hang out – we have fresh coffee all day long! Read more about the seminar series here.