ELLIS Delft Talk: Causality & Machine Learning

Speaker: Jesse Krijthe

Abstract: Whereas much of statistics & machine learning is concerned with learning relationships in a static world, causal inference is concerned with estimating the effects of making changes in the world. How can machine learning help solve this causal inference problem, and how does causality help us solve problems in machine learning? In this talk I will give some examples of how causality and machine learning problems intersect in my own work, by discussing two projects: 1. estimating the effect of treatment duration on disease progression in Parkinson’s disease and 2. domain adaptation through importance weighting. Based on this I will offer some thoughts on future research at the intersection of these two research areas.

Check the website for more info.