Learning Representations of Instruments for Partial Identification of Treatment Effects

Jonas Schweisthal, Dennis Frauen, Maresa Schröder, Konstantin Hess, Niki Kilbertus, Stefan Feuerriegel
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:53703-53727, 2025.

Abstract

Reliable estimation of treatment effects from observational data is important in many disciplines such as medicine. However, estimation is challenging when unconfoundedness as a standard assumption in the causal inference literature is violated. In this work, we leverage arbitrary (potentially high-dimensional) instruments to estimate bounds on the conditional average treatment effect (CATE). Our contributions are three-fold: (1) We propose a novel approach for partial identification through a mapping of instruments to a discrete representation space so that we yield valid bounds on the CATE. This is crucial for reliable decision-making in real-world applications. (2) We derive a two-step procedure that learns tight bounds using a tailored neural partitioning of the latent instrument space. As a result, we avoid instability issues due to numerical approximations or adversarial training. Furthermore, our procedure aims to reduce the estimation variance in finite-sample settings to yield more reliable estimates. (3) We show theoretically that our procedure obtains valid bounds while reducing estimation variance. We further perform extensive experiments to demonstrate the effectiveness across various settings. Overall, our procedure offers a novel path for practitioners to make use of potentially high-dimensional instruments (e.g., as in Mendelian randomization).

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-schweisthal25a, title = {Learning Representations of Instruments for Partial Identification of Treatment Effects}, author = {Schweisthal, Jonas and Frauen, Dennis and Schr\"{o}der, Maresa and Hess, Konstantin and Kilbertus, Niki and Feuerriegel, Stefan}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {53703--53727}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/schweisthal25a/schweisthal25a.pdf}, url = {https://proceedings.mlr.press/v267/schweisthal25a.html}, abstract = {Reliable estimation of treatment effects from observational data is important in many disciplines such as medicine. However, estimation is challenging when unconfoundedness as a standard assumption in the causal inference literature is violated. In this work, we leverage arbitrary (potentially high-dimensional) instruments to estimate bounds on the conditional average treatment effect (CATE). Our contributions are three-fold: (1) We propose a novel approach for partial identification through a mapping of instruments to a discrete representation space so that we yield valid bounds on the CATE. This is crucial for reliable decision-making in real-world applications. (2) We derive a two-step procedure that learns tight bounds using a tailored neural partitioning of the latent instrument space. As a result, we avoid instability issues due to numerical approximations or adversarial training. Furthermore, our procedure aims to reduce the estimation variance in finite-sample settings to yield more reliable estimates. (3) We show theoretically that our procedure obtains valid bounds while reducing estimation variance. We further perform extensive experiments to demonstrate the effectiveness across various settings. Overall, our procedure offers a novel path for practitioners to make use of potentially high-dimensional instruments (e.g., as in Mendelian randomization).} }
Endnote
%0 Conference Paper %T Learning Representations of Instruments for Partial Identification of Treatment Effects %A Jonas Schweisthal %A Dennis Frauen %A Maresa Schröder %A Konstantin Hess %A Niki Kilbertus %A Stefan Feuerriegel %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-schweisthal25a %I PMLR %P 53703--53727 %U https://proceedings.mlr.press/v267/schweisthal25a.html %V 267 %X Reliable estimation of treatment effects from observational data is important in many disciplines such as medicine. However, estimation is challenging when unconfoundedness as a standard assumption in the causal inference literature is violated. In this work, we leverage arbitrary (potentially high-dimensional) instruments to estimate bounds on the conditional average treatment effect (CATE). Our contributions are three-fold: (1) We propose a novel approach for partial identification through a mapping of instruments to a discrete representation space so that we yield valid bounds on the CATE. This is crucial for reliable decision-making in real-world applications. (2) We derive a two-step procedure that learns tight bounds using a tailored neural partitioning of the latent instrument space. As a result, we avoid instability issues due to numerical approximations or adversarial training. Furthermore, our procedure aims to reduce the estimation variance in finite-sample settings to yield more reliable estimates. (3) We show theoretically that our procedure obtains valid bounds while reducing estimation variance. We further perform extensive experiments to demonstrate the effectiveness across various settings. Overall, our procedure offers a novel path for practitioners to make use of potentially high-dimensional instruments (e.g., as in Mendelian randomization).
APA
Schweisthal, J., Frauen, D., Schröder, M., Hess, K., Kilbertus, N. & Feuerriegel, S.. (2025). Learning Representations of Instruments for Partial Identification of Treatment Effects. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:53703-53727 Available from https://proceedings.mlr.press/v267/schweisthal25a.html.

Related Material