Proximal Causal Learning of Conditional Average Treatment Effects

Erik Sverdrup, Yifan Cui
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:33285-33298, 2023.

Abstract

Efficiently and flexibly estimating treatment effect heterogeneity is an important task in a wide variety of settings ranging from medicine to marketing, and there are a considerable number of promising conditional average treatment effect estimators currently available. These, however, typically rely on the assumption that the measured covariates are enough to justify conditional exchangeability. We propose the P-learner, motivated by the R- and DR-learner, a tailored two-stage loss function for learning heterogeneous treatment effects in settings where exchangeability given observed covariates is an implausible assumption, and we wish to rely on proxy variables for causal inference. Our proposed estimator can be implemented by off-the-shelf loss-minimizing machine learning methods, which in the case of kernel regression satisfies an oracle bound on the estimated error as long as the nuisance components are estimated reasonably well.

Cite this Paper

BibTeX
@InProceedings{pmlr-v202-sverdrup23a, title = {Proximal Causal Learning of Conditional Average Treatment Effects}, author = {Sverdrup, Erik and Cui, Yifan}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {33285--33298}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/sverdrup23a/sverdrup23a.pdf}, url = {https://proceedings.mlr.press/v202/sverdrup23a.html}, abstract = {Efficiently and flexibly estimating treatment effect heterogeneity is an important task in a wide variety of settings ranging from medicine to marketing, and there are a considerable number of promising conditional average treatment effect estimators currently available. These, however, typically rely on the assumption that the measured covariates are enough to justify conditional exchangeability. We propose the P-learner, motivated by the R- and DR-learner, a tailored two-stage loss function for learning heterogeneous treatment effects in settings where exchangeability given observed covariates is an implausible assumption, and we wish to rely on proxy variables for causal inference. Our proposed estimator can be implemented by off-the-shelf loss-minimizing machine learning methods, which in the case of kernel regression satisfies an oracle bound on the estimated error as long as the nuisance components are estimated reasonably well.} }
Endnote
%0 Conference Paper %T Proximal Causal Learning of Conditional Average Treatment Effects %A Erik Sverdrup %A Yifan Cui %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-sverdrup23a %I PMLR %P 33285--33298 %U https://proceedings.mlr.press/v202/sverdrup23a.html %V 202 %X Efficiently and flexibly estimating treatment effect heterogeneity is an important task in a wide variety of settings ranging from medicine to marketing, and there are a considerable number of promising conditional average treatment effect estimators currently available. These, however, typically rely on the assumption that the measured covariates are enough to justify conditional exchangeability. We propose the P-learner, motivated by the R- and DR-learner, a tailored two-stage loss function for learning heterogeneous treatment effects in settings where exchangeability given observed covariates is an implausible assumption, and we wish to rely on proxy variables for causal inference. Our proposed estimator can be implemented by off-the-shelf loss-minimizing machine learning methods, which in the case of kernel regression satisfies an oracle bound on the estimated error as long as the nuisance components are estimated reasonably well.
APA
Sverdrup, E. & Cui, Y.. (2023). Proximal Causal Learning of Conditional Average Treatment Effects. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:33285-33298 Available from https://proceedings.mlr.press/v202/sverdrup23a.html.

Related Material