Obtaining Causal Information by Merging Datasets with MAXENT

Sergio H. Garrido Mejia, Elke Kirschbaum, Dominik Janzing
Proceedings of The 25th International Conference on Artificial Intelligence and Statistics, PMLR 151:581-603, 2022.

Abstract

The investigation of the question "which treatment has a causal effect on a target variable?" is of particular relevance in a large number of scientific disciplines. This challenging task becomes even more difficult if not all treatment variables were or even can not be observed jointly with the target variable. In this paper, we discuss how causal knowledge can be obtained without having observed all variables jointly, but by merging the statistical information from different datasets. We show how the maximum entropy principle can be used to identify edges among random variables when assuming causal sufficiency and an extended version of faithfulness, and when only subsets of the variables have been observed jointly.

Cite this Paper

BibTeX
@InProceedings{pmlr-v151-garrido-mejia22a, title = { Obtaining Causal Information by Merging Datasets with MAXENT }, author = {Garrido Mejia, Sergio H. and Kirschbaum, Elke and Janzing, Dominik}, booktitle = {Proceedings of The 25th International Conference on Artificial Intelligence and Statistics}, pages = {581--603}, year = {2022}, editor = {Camps-Valls, Gustau and Ruiz, Francisco J. R. and Valera, Isabel}, volume = {151}, series = {Proceedings of Machine Learning Research}, month = {28--30 Mar}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v151/garrido-mejia22a/garrido-mejia22a.pdf}, url = {https://proceedings.mlr.press/v151/garrido-mejia22a.html}, abstract = { The investigation of the question "which treatment has a causal effect on a target variable?" is of particular relevance in a large number of scientific disciplines. This challenging task becomes even more difficult if not all treatment variables were or even can not be observed jointly with the target variable. In this paper, we discuss how causal knowledge can be obtained without having observed all variables jointly, but by merging the statistical information from different datasets. We show how the maximum entropy principle can be used to identify edges among random variables when assuming causal sufficiency and an extended version of faithfulness, and when only subsets of the variables have been observed jointly. } }
Endnote
%0 Conference Paper %T Obtaining Causal Information by Merging Datasets with MAXENT %A Sergio H. Garrido Mejia %A Elke Kirschbaum %A Dominik Janzing %B Proceedings of The 25th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2022 %E Gustau Camps-Valls %E Francisco J. R. Ruiz %E Isabel Valera %F pmlr-v151-garrido-mejia22a %I PMLR %P 581--603 %U https://proceedings.mlr.press/v151/garrido-mejia22a.html %V 151 %X The investigation of the question "which treatment has a causal effect on a target variable?" is of particular relevance in a large number of scientific disciplines. This challenging task becomes even more difficult if not all treatment variables were or even can not be observed jointly with the target variable. In this paper, we discuss how causal knowledge can be obtained without having observed all variables jointly, but by merging the statistical information from different datasets. We show how the maximum entropy principle can be used to identify edges among random variables when assuming causal sufficiency and an extended version of faithfulness, and when only subsets of the variables have been observed jointly.
APA
Garrido Mejia, S.H., Kirschbaum, E. & Janzing, D.. (2022). Obtaining Causal Information by Merging Datasets with MAXENT . Proceedings of The 25th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 151:581-603 Available from https://proceedings.mlr.press/v151/garrido-mejia22a.html.

Related Material