Counterfactual Identifiability of Bijective Causal Models

Arash Nasr-Esfahany, Mohammad Alizadeh, Devavrat Shah
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:25733-25754, 2023.

Abstract

We study counterfactual identifiability in causal models with bijective generation mechanisms (BGM), a class that generalizes several widely-used causal models in the literature. We establish their counterfactual identifiability for three common causal structures with unobserved confounding, and propose a practical learning method that casts learning a BGM as structured generative modeling. Learned BGMs enable efficient counterfactual estimation and can be obtained using a variety of deep conditional generative models. We evaluate our techniques in a visual task and demonstrate its application in a real-world video streaming simulation task.

Cite this Paper


BibTeX
@InProceedings{pmlr-v202-nasr-esfahany23a, title = {Counterfactual Identifiability of Bijective Causal Models}, author = {Nasr-Esfahany, Arash and Alizadeh, Mohammad and Shah, Devavrat}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {25733--25754}, 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/nasr-esfahany23a/nasr-esfahany23a.pdf}, url = {https://proceedings.mlr.press/v202/nasr-esfahany23a.html}, abstract = {We study counterfactual identifiability in causal models with bijective generation mechanisms (BGM), a class that generalizes several widely-used causal models in the literature. We establish their counterfactual identifiability for three common causal structures with unobserved confounding, and propose a practical learning method that casts learning a BGM as structured generative modeling. Learned BGMs enable efficient counterfactual estimation and can be obtained using a variety of deep conditional generative models. We evaluate our techniques in a visual task and demonstrate its application in a real-world video streaming simulation task.} }
Endnote
%0 Conference Paper %T Counterfactual Identifiability of Bijective Causal Models %A Arash Nasr-Esfahany %A Mohammad Alizadeh %A Devavrat Shah %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-nasr-esfahany23a %I PMLR %P 25733--25754 %U https://proceedings.mlr.press/v202/nasr-esfahany23a.html %V 202 %X We study counterfactual identifiability in causal models with bijective generation mechanisms (BGM), a class that generalizes several widely-used causal models in the literature. We establish their counterfactual identifiability for three common causal structures with unobserved confounding, and propose a practical learning method that casts learning a BGM as structured generative modeling. Learned BGMs enable efficient counterfactual estimation and can be obtained using a variety of deep conditional generative models. We evaluate our techniques in a visual task and demonstrate its application in a real-world video streaming simulation task.
APA
Nasr-Esfahany, A., Alizadeh, M. & Shah, D.. (2023). Counterfactual Identifiability of Bijective Causal Models. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:25733-25754 Available from https://proceedings.mlr.press/v202/nasr-esfahany23a.html.

Related Material