Deconfounding via Profiled Transfer Learning

TL;DR

This paper introduces ProTrans, a transfer learning framework that mitigates unmeasured confounding bias in causal effect estimation by leveraging shared confounding patterns across datasets, without needing auxiliary variables.

Contribution

ProTrans is a novel profiled transfer learning method that characterizes shared confounding structures and enhances causal inference accuracy without auxiliary confounder information.

Findings

ProTrans effectively reduces confounding bias in simulations and real data.

Theoretical guarantees show confounding-free model transfer and minimax optimal estimation rates.

Source selection improves robustness against noninformative datasets.

Abstract

Unmeasured confounders are a major source of bias in regression-based effect estimation and causal inference. In this paper, we advocate a new profiled transfer learning framework, ProTrans, to address confounding effects in the target dataset, when additional source datasets that possess similar confounding structures are available. We introduce the concept of profiled residuals to characterize the shared confounding patterns between source and target datasets. By incorporating these profiled residuals into the target debiasing step, we effectively mitigates the latent confounding effects. We also propose a source selection strategy to enhance robustness of ProTrans against noninformative sources. As a byproduct, ProTrans can also be utilized to estimate treatment effects when potential confounders exist, without the use of auxiliary features such as instrumental or proxy variables, which are often challenging to select in practice. Theoretically, we prove that the resulting estimated model shift from sources to target is confounding-free without any assumptions imposed on the true confounding structure, and that the target parameter estimation achieves the minimax optimal rate under mild conditions. Simulated and real-world experiments validate the effectiveness of ProTrans and support the theoretical findings.