Causal Discovery for Linear Non-Gaussian Models with Disjoint Cycles

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Drton, Mathias, Garrote-López, Marina, Nikov, Niko, Robeva, Elina, Wang, Y. Samuel
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866912482744860672
author Drton, Mathias
Garrote-López, Marina
Nikov, Niko
Robeva, Elina
Wang, Y. Samuel
author_facet Drton, Mathias
Garrote-López, Marina
Nikov, Niko
Robeva, Elina
Wang, Y. Samuel
contents The paradigm of linear structural equation modeling readily allows one to incorporate causal feedback loops in the model specification. These appear as directed cycles in the common graphical representation of the models. However, the presence of cycles entails difficulties such as the fact that models need no longer be characterized by conditional independence relations. As a result, learning cyclic causal structures remains a challenging problem. In this paper, we offer new insights on this problem in the context of linear non-Gaussian models. First, we precisely characterize when two directed graphs determine the same linear non-Gaussian model. Next, we take up a setting of cycle-disjoint graphs, for which we are able to show that simple quadratic and cubic polynomial relations among low-order moments of a non-Gaussian distribution allow one to locate source cycles. Complementing this with a strategy of decorrelating cycles and multivariate regression allows one to infer a block-topological order among the directed cycles, which leads to a {consistent and computationally efficient algorithm} for learning causal structures with disjoint cycles.
format Preprint
id arxiv_https___arxiv_org_abs_2507_10767
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Causal Discovery for Linear Non-Gaussian Models with Disjoint Cycles
Drton, Mathias
Garrote-López, Marina
Nikov, Niko
Robeva, Elina
Wang, Y. Samuel
Statistics Theory
Methodology
Machine Learning
62H22, 62D20, 62R01
The paradigm of linear structural equation modeling readily allows one to incorporate causal feedback loops in the model specification. These appear as directed cycles in the common graphical representation of the models. However, the presence of cycles entails difficulties such as the fact that models need no longer be characterized by conditional independence relations. As a result, learning cyclic causal structures remains a challenging problem. In this paper, we offer new insights on this problem in the context of linear non-Gaussian models. First, we precisely characterize when two directed graphs determine the same linear non-Gaussian model. Next, we take up a setting of cycle-disjoint graphs, for which we are able to show that simple quadratic and cubic polynomial relations among low-order moments of a non-Gaussian distribution allow one to locate source cycles. Complementing this with a strategy of decorrelating cycles and multivariate regression allows one to infer a block-topological order among the directed cycles, which leads to a {consistent and computationally efficient algorithm} for learning causal structures with disjoint cycles.
title Causal Discovery for Linear Non-Gaussian Models with Disjoint Cycles
topic Statistics Theory
Methodology
Machine Learning
62H22, 62D20, 62R01
url https://arxiv.org/abs/2507.10767