Estimating Causal Effects from Learned Causal Networks

TL;DR

This paper introduces a novel approach to causal inference by learning causal Bayesian networks directly from observational data, which can outperform traditional estimand-based methods especially for complex models.

Contribution

It proposes a model completion learning paradigm for causal effect estimation, bypassing the need for explicit estimand derivation from causal diagrams.

Findings

Model completion can be more effective than estimand approaches for large models.

The method is demonstrated on Bayesian network benchmarks and synthetic causal models.

Learning causal networks directly enables efficient query answering in complex scenarios.

Abstract

The standard approach to answering an identifiable causal-effect query (e.g., $P(Y|do(X)$) when given a causal diagram and observational data is to first generate an estimand, or probabilistic expression over the observable variables, which is then evaluated using the observational data. In this paper, we propose an alternative paradigm for answering causal-effect queries over discrete observable variables. We propose to instead learn the causal Bayesian network and its confounding latent variables directly from the observational data. Then, efficient probabilistic graphical model (PGM) algorithms can be applied to the learned model to answer queries. Perhaps surprisingly, we show that this \emph{model completion} learning approach can be more effective than estimand approaches, particularly for larger models in which the estimand expressions become computationally difficult. We illustrate our method's potential using a benchmark collection of Bayesian networks and synthetically generated causal models.