A Sound and Complete Algorithm for Learning Causal Models from Relational Data

Marc Maier, Katerina Marazopoulou, David Arbour, David Jensen
Proceedings of the 29th Conference on Uncertainty in Artificial Intelligence, PMLR R11:510-519, 2013.

Abstract

The PC algorithm learns maximally oriented causal Bayesian networks. However, there is no equivalent complete algorithm for learning the structure of relational models, a more ex- pressive generalization of Bayesian networks. Recent developments in the theory and repre- sentation of relational models support lifted reasoning about conditional independence. This enables a powerful constraint for ori- enting bivariate dependencies and forms the basis of a new algorithm for learning struc- ture. We present the relational causal discov- ery (RCD) algorithm that learns causal rela- tional models. We prove that RCD is sound and complete, and we present empirical re- sults that demonstrate effectiveness.

Cite this Paper

BibTeX
@InProceedings{pmlr-vR11-maier13a, title = {A Sound and Complete Algorithm for Learning Causal Models from Relational Data}, author = {Maier, Marc and Marazopoulou, Katerina and Arbour, David and Jensen, David}, booktitle = {Proceedings of the 29th Conference on Uncertainty in Artificial Intelligence}, pages = {510--519}, year = {2013}, editor = {Nicholson, Ann and Smyth, Padhraic}, volume = {R11}, series = {Proceedings of Machine Learning Research}, month = {12--14 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/r11/main/assets/maier13a/maier13a.pdf}, url = {https://proceedings.mlr.press/r11/maier13a.html}, abstract = {The PC algorithm learns maximally oriented causal Bayesian networks. However, there is no equivalent complete algorithm for learning the structure of relational models, a more ex- pressive generalization of Bayesian networks. Recent developments in the theory and repre- sentation of relational models support lifted reasoning about conditional independence. This enables a powerful constraint for ori- enting bivariate dependencies and forms the basis of a new algorithm for learning struc- ture. We present the relational causal discov- ery (RCD) algorithm that learns causal rela- tional models. We prove that RCD is sound and complete, and we present empirical re- sults that demonstrate effectiveness.}, note = {Reissued by PMLR on 04 October 2026.} }
Endnote
%0 Conference Paper %T A Sound and Complete Algorithm for Learning Causal Models from Relational Data %A Marc Maier %A Katerina Marazopoulou %A David Arbour %A David Jensen %B Proceedings of the 29th Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2013 %E Ann Nicholson %E Padhraic Smyth %F pmlr-vR11-maier13a %I PMLR %P 510--519 %U https://proceedings.mlr.press/r11/maier13a.html %V R11 %X The PC algorithm learns maximally oriented causal Bayesian networks. However, there is no equivalent complete algorithm for learning the structure of relational models, a more ex- pressive generalization of Bayesian networks. Recent developments in the theory and repre- sentation of relational models support lifted reasoning about conditional independence. This enables a powerful constraint for ori- enting bivariate dependencies and forms the basis of a new algorithm for learning struc- ture. We present the relational causal discov- ery (RCD) algorithm that learns causal rela- tional models. We prove that RCD is sound and complete, and we present empirical re- sults that demonstrate effectiveness. %Z Reissued by PMLR on 04 October 2026.
APA
Maier, M., Marazopoulou, K., Arbour, D. & Jensen, D.. (2013). A Sound and Complete Algorithm for Learning Causal Models from Relational Data. Proceedings of the 29th Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research R11:510-519 Available from https://proceedings.mlr.press/r11/maier13a.html. Reissued by PMLR on 04 October 2026.

Related Material