Associate Professor in Machine Learning

Causal Inference and Counterfactuals: An Illustration

less than 1 minute read

Published:

Causal inference tackles the problem of dealing with causal statements. A rigorous statistical formalism to assess causality has been proposed by Pearl.

Pearl’s framework is based on the definition of structural equation models, it introduces the notion of intervention and it connects the (statistical, direction-agnostis, equation-based) observational regime with the (causal, direction-aware, graph-based) interventional regime.

In this notebook, we re-consider an illustrative toy example proposed by Ference Huszar on his blog to explain causal inference in the Pearlian formalism. We re-run the basic simulations for causal inference and we extend the original example in order to illustrate counterfactuals, as well. For a deeper disccussion of the setup, the models and the results, we refer to the original blogpost.

Notebook here

You May Also Enjoy

Reviewing Consistencies

less than 1 minute read

Published:

In previous posts we have discussed theoretical and conceptual properties of abstractions between causal models, while at the same time implementing code for checking our statements and for running simulations.

Scenarios of Confounding

less than 1 minute read

Published:

Causality and causal inference deal with expressing and reasoning about relationships of cause and effects, and structural causal model provide a rigorous formalism to assess causality.