Prediction-powered Generalization of Causal Inferences

Ilker Demirel, Ahmed Alaa, Anthony Philippakis, David Sontag
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:10385-10408, 2024.

Abstract

Causal inferences from a randomized controlled trial (RCT) may not pertain to a target population where some effect modifiers have a different distribution. Prior work studies generalizing the results of a trial to a target population with no outcome but covariate data available. We show how the limited size of trials makes generalization a statistically infeasible task, as it requires estimating complex nuisance functions. We develop generalization algorithms that supplement the trial data with a prediction model learned from an additional observational study (OS), without making any assumptions on the OS. We theoretically and empirically show that our methods facilitate better generalization when the OS is "high-quality", and remain robust when it is not, and e.g., have unmeasured confounding.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-demirel24a, title = {Prediction-powered Generalization of Causal Inferences}, author = {Demirel, Ilker and Alaa, Ahmed and Philippakis, Anthony and Sontag, David}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {10385--10408}, year = {2024}, editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix}, volume = {235}, series = {Proceedings of Machine Learning Research}, month = {21--27 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v235/main/assets/demirel24a/demirel24a.pdf}, url = {https://proceedings.mlr.press/v235/demirel24a.html}, abstract = {Causal inferences from a randomized controlled trial (RCT) may not pertain to a target population where some effect modifiers have a different distribution. Prior work studies generalizing the results of a trial to a target population with no outcome but covariate data available. We show how the limited size of trials makes generalization a statistically infeasible task, as it requires estimating complex nuisance functions. We develop generalization algorithms that supplement the trial data with a prediction model learned from an additional observational study (OS), without making any assumptions on the OS. We theoretically and empirically show that our methods facilitate better generalization when the OS is "high-quality", and remain robust when it is not, and e.g., have unmeasured confounding.} }
Endnote
%0 Conference Paper %T Prediction-powered Generalization of Causal Inferences %A Ilker Demirel %A Ahmed Alaa %A Anthony Philippakis %A David Sontag %B Proceedings of the 41st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Ruslan Salakhutdinov %E Zico Kolter %E Katherine Heller %E Adrian Weller %E Nuria Oliver %E Jonathan Scarlett %E Felix Berkenkamp %F pmlr-v235-demirel24a %I PMLR %P 10385--10408 %U https://proceedings.mlr.press/v235/demirel24a.html %V 235 %X Causal inferences from a randomized controlled trial (RCT) may not pertain to a target population where some effect modifiers have a different distribution. Prior work studies generalizing the results of a trial to a target population with no outcome but covariate data available. We show how the limited size of trials makes generalization a statistically infeasible task, as it requires estimating complex nuisance functions. We develop generalization algorithms that supplement the trial data with a prediction model learned from an additional observational study (OS), without making any assumptions on the OS. We theoretically and empirically show that our methods facilitate better generalization when the OS is "high-quality", and remain robust when it is not, and e.g., have unmeasured confounding.
APA
Demirel, I., Alaa, A., Philippakis, A. & Sontag, D.. (2024). Prediction-powered Generalization of Causal Inferences. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:10385-10408 Available from https://proceedings.mlr.press/v235/demirel24a.html.

Related Material