Undersmoothing Causal Estimators with Generative Trees

TL;DR

This work tackles covariate shift in observational causal inference by addressing model misspecification through undersmoothing via Debiasing Generative Trees (DeGeTs). By partitioning the input space with Generative Trees and modelling leaf-wise distributions with Gaussian Mixture Models, DeGeTs generate balanced synthetic data to augment training for downstream estimators, aiming to improve individualized treatment effect estimation while preserving average effects. Empirically, DeGeTs demonstrate competitive ATE performance and notable gains in ITE accuracy across multiple benchmarks, with increased data complexity reflecting the intended undersmoothing effect and improved robustness compared to reweighting approaches. The approach is estimator-agnostic, data-augmentation based, and shows promise for practical deployment, especially when precise individual-level causal insights are required.

Abstract

Inferring individualised treatment effects from observational data can unlock the potential for targeted interventions. It is, however, hard to infer these effects from observational data. One major problem that can arise is covariate shift where the data (outcome) conditional distribution remains the same but the covariate (input) distribution changes between the training and test set. In an observational data setting, this problem is materialised in control and treated units coming from different distributions. A common solution is to augment learning methods through reweighing schemes (e.g. propensity scores). These are needed due to model misspecification, but might hurt performance in the individual case. In this paper, we explore a novel generative tree based approach that tackles model misspecification directly, helping downstream estimators achieve better robustness. We show empirically that the choice of model class can indeed significantly affect the final performance and that reweighing methods can struggle in individualised effect estimation. Our proposed approach is competitive with reweighing methods on average treatment effects while performing significantly better on individualised treatment effects.

Figures (1)

• Figure 1: An example highlighting model misspecification issue. T and C denote Treated and Control respectively. The difference in ITE error is almost twice as in ATE.