Counterfactual Density Estimation using Kernel Stein Discrepancies

2023/09/28 by Diego Martinez-Taboada, Edward H. Kennedy, Martinez-Taboada, Diego +1 · 2 citations
Mathematics · #Statistical Methods and Bayesian Inference #Statistical Distribution Estimation and Applications

paper · pdf · doi:10.48550/arxiv.2309.16129

Abstract

Causal effects are usually studied in terms of the means of counterfactual distributions, which may be insufficient in many scenarios. Given a class of densities known up to normalizing constants, we propose to model counterfactual distributions by minimizing kernel Stein discrepancies in a doubly robust manner. This enables the estimation of counterfactuals over large classes of distributions while exploiting the desired double robustness. We present a theoretical analysis of the proposed estimator, providing sufficient conditions for consistency and asymptotic normality, as well as an examination of its empirical performance.

Cited by

• Inference on counterfactual distributions using martingale posteriors
• GDR-learners: Orthogonal Learning of Generative Models for Potential Outcomes

Related

• Doubly robust estimation and inference for a log-concave counterfactual density
• Semiparametric counterfactual density estimation
• Causal effects based on distributional distances
• Counterfactual Mean Embeddings
• Inference on counterfactual distributions using martingale posteriors