A Local Method for Identifying Causal Relations under Markov Equivalence

2021/02/28 by Zhuangyan Fang, Yue Liu, Zhi Geng +2
Computer Science · Mathematics · #Advanced Graph Neural Networks #Algorithm #Artificial intelligence #Bayesian Modeling and Causal Inference #Causal model #Causal structure #Causality (physics) #Cognitive Science and Mapping #Computer science #Conditional independence #Directed acyclic graph #Graphical model #Independence (probability theory) #Machine learning #Markov chain #Markov model #Mathematics #Statistics #Theoretical computer science #Variable (mathematics) #cs.LG #stat.ML

paper · pdf · doi:10.1016/j.artint.2022.103669

openalex publication_date 2022/01/30 · arxiv created 2022/03/05 · arxiv updated 2022/03/08 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/05

Abstract

Causality is important for designing interpretable and robust methods in artificial intelligence research. We propose a local approach to identify whether a variable is a cause of a given target under the framework of causal graphical models of directed acyclic graphs (DAGs). In general, the causal relation between two variables may not be identifiable from observational data as many causal DAGs encoding different causal relations are Markov equivalent. In this paper, we first introduce a sufficient and necessary graphical condition to check the existence of a causal path from a variable to a target in every Markov equivalent DAG. Next, we provide local criteria for identifying whether a variable is a cause/non-cause of a target based only on the local structure instead of the entire graph. Finally, we propose a local learning algorithm for this causal query via learning the local structure of the variable and some additional statistical independence tests related to the target. Simulation studies show that our local algorithm is efficient and effective, compared with other state-of-art methods.

Citations

• Proceedings of the Eleventh Conference on Uncertainty in Artificial Intelligence (1995)
• Proceedings of the Fifteenth Conference on Uncertainty in Artificial Intelligence (1999)
• Towards Characterizing Markov Equivalence Classes for Directed Acyclic Graphs with Latent Variables
• Causal Inference and Causal Explanation with Background Knowledge
• Proceedings of the Twenty-First Conference on Uncertainty in Artificial Intelligence (2005)
• An Algorithm for Fast Recovery of Sparse Causal Graphs
• Estimating high-dimensional intervention effects from observational data
• A Coefficient of Agreement for Nominal Scales
• Identifiability of Gaussian structural equation models with equal error variances
• A Logical Characterization of Constraint-Based Causal Discovery
• On the completeness of orientation rules for causal discovery in the presence of latent confounders and selection bias
• On the Identifiability of the Post-Nonlinear Causal Model
• Comparing the Areas under Two or More Correlated Receiver Operating Characteristic Curves: A Nonparametric Approach
• On sampling graphical Markov models
• Interpreting and using CPDAGs with background knowledge
• Counterfactual Fairness
• DAGs with NO TEARS: Continuous Optimization for Structure Learning
• Identifying causal effects in maximally oriented partially directed acyclic graphs
• Domain Adaptation as a Problem of Inference on Graphical Models
• Minimal enumeration of all possible total effects in a Markov equivalence class
• Proceedings of the Twenty-Second Conference on Uncertainty in Artificial Intelligence (2006)