Detecting low-complexity unobserved causes

Dominik Janzing, Eleni Sgouritsa, Oliver Stegle, Jonas Peters, Bernhard Schoelkopf
Proceedings of the 27th Conference on Uncertainty in Artificial Intelligence, PMLR R9:431-439, 2011.

Abstract

We describe a method that infers whether statistical dependences between two observed variables X and Y are due to a "direct" causal link or only due to a connecting causal path that contains an unobserved variable of low complexity, e.g., a binary variable. This problem is motivated by statistical genetics. Given a genetic marker that is correlated with a phenotype of interest, we want to detect whether this marker is causal or it only correlates with a causal one. Our method is based on the analysis of the location of the conditional distributions P(Y|x) in the simplex of all distributions of Y. We report encouraging results on semi-empirical data.

Cite this Paper

BibTeX
@InProceedings{pmlr-vR9-janzing11a, title = {Detecting low-complexity unobserved causes}, author = {Janzing, Dominik and Sgouritsa, Eleni and Stegle, Oliver and Peters, Jonas and Schoelkopf, Bernhard}, booktitle = {Proceedings of the 27th Conference on Uncertainty in Artificial Intelligence}, pages = {431--439}, year = {2011}, editor = {Cozman, Fabio and Pfeffer, Avi}, volume = {R9}, series = {Proceedings of Machine Learning Research}, month = {14--17 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/r9/main/assets/janzing11a/janzing11a.pdf}, url = {https://proceedings.mlr.press/r9/janzing11a.html}, abstract = {We describe a method that infers whether statistical dependences between two observed variables X and Y are due to a "direct" causal link or only due to a connecting causal path that contains an unobserved variable of low complexity, e.g., a binary variable. This problem is motivated by statistical genetics. Given a genetic marker that is correlated with a phenotype of interest, we want to detect whether this marker is causal or it only correlates with a causal one. Our method is based on the analysis of the location of the conditional distributions P(Y|x) in the simplex of all distributions of Y. We report encouraging results on semi-empirical data.}, note = {Reissued by PMLR on 04 October 2026.} }
Endnote
%0 Conference Paper %T Detecting low-complexity unobserved causes %A Dominik Janzing %A Eleni Sgouritsa %A Oliver Stegle %A Jonas Peters %A Bernhard Schoelkopf %B Proceedings of the 27th Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2011 %E Fabio Cozman %E Avi Pfeffer %F pmlr-vR9-janzing11a %I PMLR %P 431--439 %U https://proceedings.mlr.press/r9/janzing11a.html %V R9 %X We describe a method that infers whether statistical dependences between two observed variables X and Y are due to a "direct" causal link or only due to a connecting causal path that contains an unobserved variable of low complexity, e.g., a binary variable. This problem is motivated by statistical genetics. Given a genetic marker that is correlated with a phenotype of interest, we want to detect whether this marker is causal or it only correlates with a causal one. Our method is based on the analysis of the location of the conditional distributions P(Y|x) in the simplex of all distributions of Y. We report encouraging results on semi-empirical data. %Z Reissued by PMLR on 04 October 2026.
APA
Janzing, D., Sgouritsa, E., Stegle, O., Peters, J. & Schoelkopf, B.. (2011). Detecting low-complexity unobserved causes. Proceedings of the 27th Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research R9:431-439 Available from https://proceedings.mlr.press/r9/janzing11a.html. Reissued by PMLR on 04 October 2026.

Related Material