Adversarial Debiasing for Parameter Recovery

Luke Sanford, Megan Ayers, Matthew Gordon, Eliana Stone
Proceedings of The 29th International Conference on Artificial Intelligence and Statistics, PMLR 300:3880-3888, 2026.

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 outcomes of interest. However, prediction errors from machine learning models can lead to bias in downstream estimation tasks, including regression. 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 generate predictions suitable for unbiased downstream estimation. Here, we focus on a 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. Our approach consistently matches or exceeds the performance of existing methods.

Cite this Paper

BibTeX
@InProceedings{pmlr-v300-sanford26a, title = { Adversarial Debiasing for Parameter Recovery }, author = {Sanford, Luke and Ayers, Megan and Gordon, Matthew and Stone, Eliana}, booktitle = {Proceedings of The 29th International Conference on Artificial Intelligence and Statistics}, pages = {3880--3888}, year = {2026}, editor = {Khan, Emtiyaz and Li, Yingzhen and Solin, Arno and Ramdas, Aaditya}, volume = {300}, series = {Proceedings of Machine Learning Research}, month = {02--05 May}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v300/main/assets/sanford26a/sanford26a.pdf}, url = {https://proceedings.mlr.press/v300/sanford26a.html}, 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 outcomes of interest. However, prediction errors from machine learning models can lead to bias in downstream estimation tasks, including regression. 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 generate predictions suitable for unbiased downstream estimation. Here, we focus on a 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. Our approach consistently matches or exceeds the performance of existing methods. } }
Endnote
%0 Conference Paper %T Adversarial Debiasing for Parameter Recovery %A Luke Sanford %A Megan Ayers %A Matthew Gordon %A Eliana Stone %B Proceedings of The 29th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2026 %E Emtiyaz Khan %E Yingzhen Li %E Arno Solin %E Aaditya Ramdas %F pmlr-v300-sanford26a %I PMLR %P 3880--3888 %U https://proceedings.mlr.press/v300/sanford26a.html %V 300 %X 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 outcomes of interest. However, prediction errors from machine learning models can lead to bias in downstream estimation tasks, including regression. 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 generate predictions suitable for unbiased downstream estimation. Here, we focus on a 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. Our approach consistently matches or exceeds the performance of existing methods.
APA
Sanford, L., Ayers, M., Gordon, M. & Stone, E.. (2026). Adversarial Debiasing for Parameter Recovery . Proceedings of The 29th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 300:3880-3888 Available from https://proceedings.mlr.press/v300/sanford26a.html.

Related Material