Causal learning without DAGs

David Duvenaud, Daniel Eaton, Kevin Murphy, Mark Schmidt
Proceedings of Workshop on Causality: Objectives and Assessment at NIPS 2008, PMLR 6:177-190, 2010.

Abstract

Causal learning methods are often evaluated in terms of their ability to discover a true underlying directed acyclic graph (DAG) structure. However, in general the true structure is unknown and may not be a DAG structure. We therefore consider evaluating causal learning methods in terms of predicting the effects of interventions on unseen test data. Given this task, we show that there exist a variety of approaches to modeling causality, generalizing DAG-based methods. Our experiments on synthetic and biological data indicate that some non-DAG models perform as well or better than DAG-based methods at causal prediction tasks.

Cite this Paper

BibTeX
@InProceedings{pmlr-v6-duvenaud10a, title = {Causal learning without DAGs}, author = {Duvenaud, David and Eaton, Daniel and Murphy, Kevin and Schmidt, Mark}, booktitle = {Proceedings of Workshop on Causality: Objectives and Assessment at NIPS 2008}, pages = {177--190}, year = {2010}, editor = {Guyon, Isabelle and Janzing, Dominik and Schölkopf, Bernhard}, volume = {6}, series = {Proceedings of Machine Learning Research}, address = {Whistler, Canada}, month = {12 Dec}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v6/duvenaud10a/duvenaud10a.pdf}, url = {https://proceedings.mlr.press/v6/duvenaud10a.html}, abstract = {Causal learning methods are often evaluated in terms of their ability to discover a true underlying directed acyclic graph (DAG) structure. However, in general the true structure is unknown and may not be a DAG structure. We therefore consider evaluating causal learning methods in terms of predicting the effects of interventions on unseen test data. Given this task, we show that there exist a variety of approaches to modeling causality, generalizing DAG-based methods. Our experiments on synthetic and biological data indicate that some non-DAG models perform as well or better than DAG-based methods at causal prediction tasks.} }
Endnote
%0 Conference Paper %T Causal learning without DAGs %A David Duvenaud %A Daniel Eaton %A Kevin Murphy %A Mark Schmidt %B Proceedings of Workshop on Causality: Objectives and Assessment at NIPS 2008 %C Proceedings of Machine Learning Research %D 2010 %E Isabelle Guyon %E Dominik Janzing %E Bernhard Schölkopf %F pmlr-v6-duvenaud10a %I PMLR %P 177--190 %U https://proceedings.mlr.press/v6/duvenaud10a.html %V 6 %X Causal learning methods are often evaluated in terms of their ability to discover a true underlying directed acyclic graph (DAG) structure. However, in general the true structure is unknown and may not be a DAG structure. We therefore consider evaluating causal learning methods in terms of predicting the effects of interventions on unseen test data. Given this task, we show that there exist a variety of approaches to modeling causality, generalizing DAG-based methods. Our experiments on synthetic and biological data indicate that some non-DAG models perform as well or better than DAG-based methods at causal prediction tasks.
RIS
TY - CPAPER TI - Causal learning without DAGs AU - David Duvenaud AU - Daniel Eaton AU - Kevin Murphy AU - Mark Schmidt BT - Proceedings of Workshop on Causality: Objectives and Assessment at NIPS 2008 DA - 2010/02/18 ED - Isabelle Guyon ED - Dominik Janzing ED - Bernhard Schölkopf ID - pmlr-v6-duvenaud10a PB - PMLR DP - Proceedings of Machine Learning Research VL - 6 SP - 177 EP - 190 L1 - http://proceedings.mlr.press/v6/duvenaud10a/duvenaud10a.pdf UR - https://proceedings.mlr.press/v6/duvenaud10a.html AB - Causal learning methods are often evaluated in terms of their ability to discover a true underlying directed acyclic graph (DAG) structure. However, in general the true structure is unknown and may not be a DAG structure. We therefore consider evaluating causal learning methods in terms of predicting the effects of interventions on unseen test data. Given this task, we show that there exist a variety of approaches to modeling causality, generalizing DAG-based methods. Our experiments on synthetic and biological data indicate that some non-DAG models perform as well or better than DAG-based methods at causal prediction tasks. ER -
APA
Duvenaud, D., Eaton, D., Murphy, K. & Schmidt, M.. (2010). Causal learning without DAGs. Proceedings of Workshop on Causality: Objectives and Assessment at NIPS 2008, in Proceedings of Machine Learning Research 6:177-190 Available from https://proceedings.mlr.press/v6/duvenaud10a.html.

Related Material