An Asymmetric Independence Model for Causal Discovery on Path Spaces
Abstract
In this paper, we develop the theory linking directed mixed graphs (DMGs) with the graphical ’E-separation’-criterion to form asymmetric independence models that are closed under marginalization and which graphically describe the conditional independence relations among coordinate processes in stochastic differential equations (SDEs) when testing "which variables enter the governing equations of which other variables." Besides a global Markov property for cyclic SDEs, which naturally extends to latent, cyclic SDEs, we also characterize graphs, which encode the same set of independence relations and show that in the fully observed case, modelled by directed graphs, each class of graphs under this equivalence relation has a maximal element that is graphi- cally characterizable, analogous to the famous ’same skeleton and V-structure’ result for Directed Acyclic Graphs (DAGs) and in addition is recoverable from data. Moreover, we conjecture that same holds true in the partially observed case and verify this empirically for graphs up to 4 nodes.
Cite
Text
Manten et al. "An Asymmetric Independence Model for Causal Discovery on Path Spaces." Proceedings of the Fourth Conference on Causal Learning and Reasoning, 2025.Markdown
[Manten et al. "An Asymmetric Independence Model for Causal Discovery on Path Spaces." Proceedings of the Fourth Conference on Causal Learning and Reasoning, 2025.](https://mlanthology.org/clear/2025/manten2025clear-asymmetric/)BibTeX
@inproceedings{manten2025clear-asymmetric,
title = {{An Asymmetric Independence Model for Causal Discovery on Path Spaces}},
author = {Manten, Georg and Casolo, Cecilia and Mogensen, Søren Wengel and Kilbertus, Niki},
booktitle = {Proceedings of the Fourth Conference on Causal Learning and Reasoning},
year = {2025},
pages = {64-89},
volume = {275},
url = {https://mlanthology.org/clear/2025/manten2025clear-asymmetric/}
}