Identifiability of Causal Graphs using Functional Models

Abstract

This work addresses the following question: Under what assumptions on the data generating process can one infer the causal graph from the joint distribution? The approach taken by conditional independence-based causal discovery methods is based on two assumptions: the Markov condition and faithfulness. It has been shown that under these assumptions the causal graph can be identified up to Markov equivalence (some arrows remain undirected) using methods like the PC algorithm. In this work we propose an alternative by defining Identifiable Functional Model Classes (IFMOCs). As our main theorem we prove that if the data generating process belongs to an IFMOC, one can identify the complete causal graph. To the best of our knowledge this is the first identifiability result of this kind that is not limited to linear functional relationships. We discuss how the IFMOC assumption and the Markov and faithfulness assumptions relate to each other and explain why we believe that the IFMOC assumption can be tested more easily on given data. We further provide a practical algorithm that recovers the causal graph from finitely many data; experiments on simulated data support the theoretical findings.

References in corpus (2)

• Causal Inference and Causal Explanation with Background Knowledge
• On the Identifiability of the Post-Nonlinear Causal Model

Cited by in corpus (20)

• Identifiability of Gaussian structural equation models with equal error variances
• Quantifying causal influences
• Causal Discovery with Continuous Additive Noise Models
• Causality for Machine Learning
• Two Optimal Strategies for Active Learning of Causal Models from Interventional Data
• Jointly interventional and observational data: estimation of interventional Markov equivalence classes of directed acyclic graphs
• Learning Neural Causal Models from Unknown Interventions
• Inferring large graphs using l1-penalized likelihood
• Consistency of Causal Inference under the Additive Noise Model
• Score-based Causal Learning in Additive Noise Models
• Cyclic Causal Discovery from Continuous Equilibrium Data
• Causal Discovery from Incomplete Data: A Deep Learning Approach
• Network Inference Using Steady State Data and Goldbeter-Koshland Kinetics
• On the Equivalence Between Temporal and Static Graph Representations for Observational Predictions
• Causal Influence Detection for Improving Efficiency in Reinforcement Learning
• Active Invariant Causal Prediction: Experiment Selection through Stability
• Distinguishing Cause from Effect Using Quantiles: Bivariate Quantile Causal Discovery
• Amortized learning of neural causal representations
• Contrastive Mixture of Posteriors for Counterfactual Inference, Data Integration and Fairness
• Discovery of Causal Additive Models in the Presence of Unobserved Variables