Adversarial Debiasing for Unbiased Parameter Recovery

TL;DR

This paper introduces an adversarial machine learning approach to correct bias in regression estimates caused by prediction errors in high-dimensional data, improving the accuracy of parameter recovery in social science applications.

Contribution

It proposes a novel adversarial debiasing method for unbiased parameter estimation when using machine learning predictions as proxies.

Findings

Adversarial debiasing reduces bias in regression coefficients.

Naive machine learning predictions lead to biased estimates.

The proposed method accurately recovers true coefficients in simulations and real data.

Abstract

Advances in machine learning and the increasing availability of high-dimensional data have led to the proliferation of social science research that uses the predictions of machine learning models as proxies for measures of human activity or environmental outcomes. However, prediction errors from machine learning models can lead to bias in the estimates of regression coefficients. In this paper, we show how this bias can arise, propose a test for detecting bias, and demonstrate the use of an adversarial machine learning algorithm in order to de-bias predictions. These methods are applicable to any setting where machine-learned predictions are the dependent variable in a regression. We conduct simulations and empirical exercises using ground truth and satellite data on forest cover in Africa. Using the predictions from a naive machine learning model leads to biased parameter estimates, while the predictions from the adversarial model recover the true coefficients.