End-to-End Balancing for Causal Continuous Treatment-Effect Estimation

Taha Bahadori, Eric Tchetgen Tchetgen, David Heckerman
Proceedings of the 39th International Conference on Machine Learning, PMLR 162:1313-1326, 2022.

Abstract

We study the problem of observational causal inference with continuous treatment. We focus on the challenge of estimating the causal response curve for infrequently-observed treatment values. We design a new algorithm based on the framework of entropy balancing which learns weights that directly maximize causal inference accuracy using end-to-end optimization. Our weights can be customized for different datasets and causal inference algorithms. We propose a new theory for consistency of entropy balancing for continuous treatments. Using synthetic and real-world data, we show that our proposed algorithm outperforms the entropy balancing in terms of causal inference accuracy.

Cite this Paper

BibTeX
@InProceedings{pmlr-v162-bahadori22a, title = {End-to-End Balancing for Causal Continuous Treatment-Effect Estimation}, author = {Bahadori, Taha and Tchetgen, Eric Tchetgen and Heckerman, David}, booktitle = {Proceedings of the 39th International Conference on Machine Learning}, pages = {1313--1326}, year = {2022}, editor = {Chaudhuri, Kamalika and Jegelka, Stefanie and Song, Le and Szepesvari, Csaba and Niu, Gang and Sabato, Sivan}, volume = {162}, series = {Proceedings of Machine Learning Research}, month = {17--23 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v162/bahadori22a/bahadori22a.pdf}, url = {https://proceedings.mlr.press/v162/bahadori22a.html}, abstract = {We study the problem of observational causal inference with continuous treatment. We focus on the challenge of estimating the causal response curve for infrequently-observed treatment values. We design a new algorithm based on the framework of entropy balancing which learns weights that directly maximize causal inference accuracy using end-to-end optimization. Our weights can be customized for different datasets and causal inference algorithms. We propose a new theory for consistency of entropy balancing for continuous treatments. Using synthetic and real-world data, we show that our proposed algorithm outperforms the entropy balancing in terms of causal inference accuracy.} }
Endnote
%0 Conference Paper %T End-to-End Balancing for Causal Continuous Treatment-Effect Estimation %A Taha Bahadori %A Eric Tchetgen Tchetgen %A David Heckerman %B Proceedings of the 39th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2022 %E Kamalika Chaudhuri %E Stefanie Jegelka %E Le Song %E Csaba Szepesvari %E Gang Niu %E Sivan Sabato %F pmlr-v162-bahadori22a %I PMLR %P 1313--1326 %U https://proceedings.mlr.press/v162/bahadori22a.html %V 162 %X We study the problem of observational causal inference with continuous treatment. We focus on the challenge of estimating the causal response curve for infrequently-observed treatment values. We design a new algorithm based on the framework of entropy balancing which learns weights that directly maximize causal inference accuracy using end-to-end optimization. Our weights can be customized for different datasets and causal inference algorithms. We propose a new theory for consistency of entropy balancing for continuous treatments. Using synthetic and real-world data, we show that our proposed algorithm outperforms the entropy balancing in terms of causal inference accuracy.
APA
Bahadori, T., Tchetgen, E.T. & Heckerman, D.. (2022). End-to-End Balancing for Causal Continuous Treatment-Effect Estimation. Proceedings of the 39th International Conference on Machine Learning, in Proceedings of Machine Learning Research 162:1313-1326 Available from https://proceedings.mlr.press/v162/bahadori22a.html.

Related Material