A Multivariate Causal Discovery based on Post-Nonlinear Model

Kento Uemura, Takuya Takagi, Kambayashi Takayuki, Hiroyuki Yoshida, Shohei Shimizu
Proceedings of the First Conference on Causal Learning and Reasoning, PMLR 177:826-839, 2022.

Abstract

Understanding causal relations of systems is a fundamental problem in science. The study of causal discovery aims to infer the underlying causal structure from uncontrolled observational samples. One major approach is to assume that causal structures follow structural equation models (SEMs), such as the additive noise model (ANM) and the post-nonlinear (PNL) model, and to identify these causal structures by estimating the SEMs. Although the PNL model is the most general SEM for causal discovery, its estimation method has not been well-developed except for the bivariate case. In this paper, we propose a new causal discovery method based on the multivariate PNL model. We extend the bivariate method to estimate multi-cause PNL models and combine it with the iterative sink search scheme used for the ANM. We apply the proposed method to synthetic and real-world causal discovery problems and show its effectiveness.

Cite this Paper


BibTeX
@InProceedings{pmlr-v177-uemura22a, title = {A Multivariate Causal Discovery based on Post-Nonlinear Model}, author = {Uemura, Kento and Takagi, Takuya and Takayuki, Kambayashi and Yoshida, Hiroyuki and Shimizu, Shohei}, booktitle = {Proceedings of the First Conference on Causal Learning and Reasoning}, pages = {826--839}, year = {2022}, editor = {Schölkopf, Bernhard and Uhler, Caroline and Zhang, Kun}, volume = {177}, series = {Proceedings of Machine Learning Research}, month = {11--13 Apr}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v177/uemura22a/uemura22a.pdf}, url = {https://proceedings.mlr.press/v177/uemura22a.html}, abstract = {Understanding causal relations of systems is a fundamental problem in science. The study of causal discovery aims to infer the underlying causal structure from uncontrolled observational samples. One major approach is to assume that causal structures follow structural equation models (SEMs), such as the additive noise model (ANM) and the post-nonlinear (PNL) model, and to identify these causal structures by estimating the SEMs. Although the PNL model is the most general SEM for causal discovery, its estimation method has not been well-developed except for the bivariate case. In this paper, we propose a new causal discovery method based on the multivariate PNL model. We extend the bivariate method to estimate multi-cause PNL models and combine it with the iterative sink search scheme used for the ANM. We apply the proposed method to synthetic and real-world causal discovery problems and show its effectiveness. } }
Endnote
%0 Conference Paper %T A Multivariate Causal Discovery based on Post-Nonlinear Model %A Kento Uemura %A Takuya Takagi %A Kambayashi Takayuki %A Hiroyuki Yoshida %A Shohei Shimizu %B Proceedings of the First Conference on Causal Learning and Reasoning %C Proceedings of Machine Learning Research %D 2022 %E Bernhard Schölkopf %E Caroline Uhler %E Kun Zhang %F pmlr-v177-uemura22a %I PMLR %P 826--839 %U https://proceedings.mlr.press/v177/uemura22a.html %V 177 %X Understanding causal relations of systems is a fundamental problem in science. The study of causal discovery aims to infer the underlying causal structure from uncontrolled observational samples. One major approach is to assume that causal structures follow structural equation models (SEMs), such as the additive noise model (ANM) and the post-nonlinear (PNL) model, and to identify these causal structures by estimating the SEMs. Although the PNL model is the most general SEM for causal discovery, its estimation method has not been well-developed except for the bivariate case. In this paper, we propose a new causal discovery method based on the multivariate PNL model. We extend the bivariate method to estimate multi-cause PNL models and combine it with the iterative sink search scheme used for the ANM. We apply the proposed method to synthetic and real-world causal discovery problems and show its effectiveness.
APA
Uemura, K., Takagi, T., Takayuki, K., Yoshida, H. & Shimizu, S.. (2022). A Multivariate Causal Discovery based on Post-Nonlinear Model. Proceedings of the First Conference on Causal Learning and Reasoning, in Proceedings of Machine Learning Research 177:826-839 Available from https://proceedings.mlr.press/v177/uemura22a.html.

Related Material