Statistiskt seminarium

Abstrakt finns enbart på engelska: Causal effect estimation from observational data is a fundamental task in empirical sciences. It becomes particularly challenging when unobserved confounders are involved in a system. Front-door adjustment constitutes a classic method that allows identifying the causal effect even in the presence of latent confounding by using observed mediators. This talk presents a recent algorithmic result in this area, namely a linear-time algorithm for finding a front-door adjustment set in a given causal graph. Its run-time is asymptotically optimal and improves on the previous state-of-the-art for this task by a factor that grows cubically in the number of variables. Beyond this result, the presentation explores fundamental algorithmic tools and techniques useful for broader applications in causal inference.