Foundations of structural causal models with cycles and latent variables

DOI10.1214/21-AOS2064zbMATH Open1486.62176arXiv1611.06221OpenAlexW3214040448MaRDI QIDQ2054537FDOQ2054537

Authors: Stephan Bongers, Patrick Forré, Jonas Peters, Joris Mooij 

Publication date: 3 December 2021

Published in: The Annals of Statistics (Search for Journal in Brave)

Abstract: Structural causal models (SCMs), also known as (nonparametric) structural equation models (SEMs), are widely used for causal modeling purposes. In particular, acyclic SCMs, also known as recursive SEMs, form a well-studied subclass of SCMs that generalize causal Bayesian networks to allow for latent confounders. In this paper, we investigate SCMs in a more general setting, allowing for the presence of both latent confounders and cycles. We show that in the presence of cycles, many of the convenient properties of acyclic SCMs do not hold in general: they do not always have a solution; they do not always induce unique observational, interventional and counterfactual distributions; a marginalization does not always exist, and if it exists the marginal model does not always respect the latent projection; they do not always satisfy a Markov property; and their graphs are not always consistent with their causal semantics. We prove that for SCMs in general each of these properties does hold under certain solvability conditions. Our work generalizes results for SCMs with cycles that were only known for certain special cases so far. We introduce the class of simple SCMs that extends the class of acyclic SCMs to the cyclic setting, while preserving many of the convenient properties of acyclic SCMs. With this paper we aim to provide the foundations for a general theory of statistical causal modeling with SCMs.

Full work available at URL: https://arxiv.org/abs/1611.06221

Recommendations

• The structural theory of causation
• Identification in causal models with hidden variables
• scientific article; zbMATH DE number 3878159
• scientific article; zbMATH DE number 6276218
• Causal Graphical Models with Latent Variables: Learning and Inference

zbMATH Keywords

counterfactualscyclessolvabilityinterventionsMarkov propertiescausal graphmarginalizationstructural causal models

Mathematics Subject Classification ID

Probabilistic graphical models (62H22) Graphical methods in statistics (62A09) Reasoning under uncertainty in the context of artificial intelligence (68T37) Knowledge representation (68T30)

Cites Work

• Causation, prediction, and search
• Title not available (Why is that?)
• Learning stable and predictive structures in kinetic systems
• Title not available (Why is that?)
• Title not available (Why is that?)
• Causality. Models, reasoning, and inference
• Elements of causal inference. Foundations and learning algorithms
• Influence Diagrams for Causal Modelling and Inference
• Estimating high-dimensional intervention effects from observational data
• Half-trek criterion for generic identifiability of linear structural equation models
• Ancestral graph Markov models.
• The book of why. The new science of cause and effect
• Causality and model abstraction
• Title not available (Why is that?)
• Complete identification methods for the causal hierarchy
• Title not available (Why is that?)
• Markov Properties for Acyclic Directed Mixed Graphs
• The Statistical Implications of a System of Simultaneous Equations
• On the Validity of the Markov Interpretation of Path Diagrams of Gaussian Structural Equations Systems with Correlated Errors
• Markov properties of nonrecursive causal models
• On the completeness of orientation rules for causal discovery in the presence of latent confounders and selection bias
• Title not available (Why is that?)
• Independence properties of directed markov fields
• CAM: causal additive models, high-dimensional order search and penalized regression
• Title not available (Why is that?)
• Title not available (Why is that?)
• Distinguishing cause from effect using observational data: methods and benchmarks
• Margins of discrete Bayesian networks
• Graphs for margins of Bayesian networks
• Markov equivalence of marginalized local independence graphs

Cited In (11)

• Discovering agents
• A survey on causal discovery: theory and practice
• Bias of the additive hazard model in the presence of causal effect heterogeneity
• The built-in selection bias of hazard ratios formalized using structural causal models
• Markov conditions and factorization in logical credal networks
• The \(\sigma\)-intervention based on information account of causality and its corresponding causal calculus
• From statistical to causal learning
• A developmental path model and causal analysis of latent dichotomous variables
• Causal structure learning: a combinatorial perspective
• Reasoning about causality in games
• Title not available (Why is that?)

Uses Software

• TETRAD
• JCI
• CausalKinetiX

This page was built for publication: Foundations of structural causal models with cycles and latent variables

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2054537)