Estimating Causal Direction and Confounding of Two Discrete Variables

TL;DR

This paper introduces a method to determine causal direction and confounding between two discrete variables using only their joint distribution, based on the assumption of independent causal mechanisms, with an acknowledgment of baseline error.

Contribution

It presents a novel classification approach for causal inference between discrete variables that accounts for hidden confounders and does not rely on prior distributions in test data.

Findings

Method effectively classifies causal direction in discrete variables.

Accounts for hidden confounders in causal inference.

Provides open-source code for reproducibility.

Abstract

We propose a method to classify the causal relationship between two discrete variables given only the joint distribution of the variables, acknowledging that the method is subject to an inherent baseline error. We assume that the causal system is acyclicity, but we do allow for hidden common causes. Our algorithm presupposes that the probability distributions $P(C)$ of a cause $C$ is independent from the probability distribution $P(E\mid C)$ of the cause-effect mechanism. While our classifier is trained with a Bayesian assumption of flat hyperpriors, we do not make this assumption about our test data. This work connects to recent developments on the identifiability of causal models over continuous variables under the assumption of "independent mechanisms". Carefully-commented Python notebooks that reproduce all our experiments are available online at http://vision.caltech.edu/~kchalupk/code.html.