Treatment Effect Estimation using Invariant Risk Minimization

TL;DR

This paper introduces an invariant risk minimization approach for estimating individual treatment effects from observational data, especially effective when there is limited overlap between treatment groups, improving generalization across domains.

Contribution

It proposes a novel IRM-based estimator that creates artificial domain diversity to better handle support mismatch in treatment effect estimation from observational data.

Findings

IRM-based estimator outperforms classical regression methods in support mismatch scenarios

Creating artificial domains enhances the model's ability to generalize to unseen data regions

The approach effectively reduces bias caused by treatment assignment bias

Abstract

Inferring causal individual treatment effect (ITE) from observational data is a challenging problem whose difficulty is exacerbated by the presence of treatment assignment bias. In this work, we propose a new way to estimate the ITE using the domain generalization framework of invariant risk minimization (IRM). IRM uses data from multiple domains, learns predictors that do not exploit spurious domain-dependent factors, and generalizes better to unseen domains. We propose an IRM-based ITE estimator aimed at tackling treatment assignment bias when there is little support overlap between the control group and the treatment group. We accomplish this by creating diversity: given a single dataset, we split the data into multiple domains artificially. These diverse domains are then exploited by IRM to more effectively generalize regression-based models to data regions that lack support overlap. We show gains over classical regression approaches to ITE estimation in settings when support mismatch is more pronounced.