DeepMatch: Balancing Deep Covariate Representations for Causal Inference Using Adversarial Training

TL;DR

DeepMatch introduces an adversarial training approach to improve covariate balance in causal inference from observational data, effectively handling complex relationships and rich covariates with neural networks.

Contribution

It proposes a novel adversarial training method for covariate balancing, addressing limitations of existing approaches in complex, high-dimensional settings.

Findings

Theoretical characterization of DeepMatch's effectiveness.

Empirical validation with neural network architectures on complex data.

Demonstrated ability to handle image confounders in causal analysis.

Abstract

We study optimal covariate balance for causal inferences from observational data when rich covariates and complex relationships necessitate flexible modeling with neural networks. Standard approaches such as propensity weighting and matching/balancing fail in such settings due to miscalibrated propensity nets and inappropriate covariate representations, respectively. We propose a new method based on adversarial training of a weighting and a discriminator network that effectively addresses this methodological gap. This is demonstrated through new theoretical characterizations of the method as well as empirical results using both fully connected architectures to learn complex relationships and convolutional architectures to handle image confounders, showing how this new method can enable strong causal analyses in these challenging settings.