Variational Counterfactual Intervention Planning to Achieve Target Outcomes

Xin Wang, Shengfei Lyu, Chi Luo, Xiren Zhou, Huanhuan Chen
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:63860-63881, 2025.

Abstract

A key challenge in personalized healthcare is identifying optimal intervention sequences to guide temporal systems toward target outcomes, a novel problem we formalize as counterfactual target achievement. In addressing this problem, directly adopting counterfactual estimation methods face compounding errors due to the unobservability of counterfactuals. To overcome this, we propose Variational Counterfactual Intervention Planning (VCIP), which reformulates the problem by modeling the conditional likelihood of achieving target outcomes, implemented through variational inference. By leveraging the g-formula to bridge the gap between interventional and observational log-likelihoods, VCIP enables reliable training from observational data. Experiments on both synthetic and real-world datasets show that VCIP significantly outperforms existing methods in target achievement accuracy.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-wang25bv, title = {Variational Counterfactual Intervention Planning to Achieve Target Outcomes}, author = {Wang, Xin and Lyu, Shengfei and Luo, Chi and Zhou, Xiren and Chen, Huanhuan}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {63860--63881}, 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/wang25bv/wang25bv.pdf}, url = {https://proceedings.mlr.press/v267/wang25bv.html}, abstract = {A key challenge in personalized healthcare is identifying optimal intervention sequences to guide temporal systems toward target outcomes, a novel problem we formalize as counterfactual target achievement. In addressing this problem, directly adopting counterfactual estimation methods face compounding errors due to the unobservability of counterfactuals. To overcome this, we propose Variational Counterfactual Intervention Planning (VCIP), which reformulates the problem by modeling the conditional likelihood of achieving target outcomes, implemented through variational inference. By leveraging the g-formula to bridge the gap between interventional and observational log-likelihoods, VCIP enables reliable training from observational data. Experiments on both synthetic and real-world datasets show that VCIP significantly outperforms existing methods in target achievement accuracy.} }
Endnote
%0 Conference Paper %T Variational Counterfactual Intervention Planning to Achieve Target Outcomes %A Xin Wang %A Shengfei Lyu %A Chi Luo %A Xiren Zhou %A Huanhuan Chen %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-wang25bv %I PMLR %P 63860--63881 %U https://proceedings.mlr.press/v267/wang25bv.html %V 267 %X A key challenge in personalized healthcare is identifying optimal intervention sequences to guide temporal systems toward target outcomes, a novel problem we formalize as counterfactual target achievement. In addressing this problem, directly adopting counterfactual estimation methods face compounding errors due to the unobservability of counterfactuals. To overcome this, we propose Variational Counterfactual Intervention Planning (VCIP), which reformulates the problem by modeling the conditional likelihood of achieving target outcomes, implemented through variational inference. By leveraging the g-formula to bridge the gap between interventional and observational log-likelihoods, VCIP enables reliable training from observational data. Experiments on both synthetic and real-world datasets show that VCIP significantly outperforms existing methods in target achievement accuracy.
APA
Wang, X., Lyu, S., Luo, C., Zhou, X. & Chen, H.. (2025). Variational Counterfactual Intervention Planning to Achieve Target Outcomes. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:63860-63881 Available from https://proceedings.mlr.press/v267/wang25bv.html.

Related Material