Causal Discovery with Continuous Additive Noise Models
arXiv:1309.6779
Abstract
We consider the problem of learning causal directed acyclic graphs from an observational joint distribution. One can use these graphs to predict the outcome of interventional experiments, from which data are often not available. We show that if the observational distribution follows a structural equation model with an additive noise structure, the directed acyclic graph becomes identifiable from the distribution under mild conditions. This constitutes an interesting alternative to traditional methods that assume faithfulness and identify only the Markov equivalence class of the graph, thus leaving some edges undirected. We provide practical algorithms for finitely many samples, RESIT (Regression with Subsequent Independence Test) and two methods based on an independence score. We prove that RESIT is correct in the population setting and provide an empirical evaluation.
References in corpus (12)
- Estimating high-dimensional directed acyclic graphs with the PC-algorithm
- Kernel-based Conditional Independence Test and Application in Causal Discovery
- Distinguishing cause from effect using observational data: methods and benchmarks
- A Transformational Characterization of Equivalent Bayesian Network Structures
- CAM: Causal additive models, high-dimensional order search and penalized regression
- Identifiability of Gaussian structural equation models with equal error variances
- Geometry of the faithfulness assumption in causal inference
- Causal Inference on Discrete Data using Additive Noise Models
- On Causal and Anticausal Learning
- Searching for Bayesian Network Structures in the Space of Restricted Acyclic Partially Directed Graphs
- Identifiability of Causal Graphs using Functional Models
- Identifying confounders using additive noise models
Cited by in corpus (25)
- The lesson of causal discovery algorithms for quantum correlations: Causal explanations of Bell-inequality violations require fine-tuning
- CAM: Causal additive models, high-dimensional order search and penalized regression
- A Clarification of the Nuances in the Fairness Metrics Landscape
- Dark, Beyond Deep: A Paradigm Shift to Cognitive AI with Humanlike Common Sense
- Large-Scale Kernel Methods for Independence Testing
- Evolution of statistical analysis in empirical software engineering research: Current state and steps forward
- Causality in cognitive neuroscience: concepts, challenges, and distributional robustness
- Causal inference via algebraic geometry: feasibility tests for functional causal structures with two binary observed variables
- Telling Cause from Effect using MDL-based Local and Global Regression
- Inferring large graphs using l1-penalized likelihood
- Distributed Learning of Generalized Linear Causal Networks
- Estimation of Structural Causal Model via Sparsely Mixing Independent Component Analysis
- Nonlinear Causal Discovery with Confounders
- Causal Effect Identification from Multiple Incomplete Data Sources: A General Search-based Approach
- Computably Continuous Reinforcement-Learning Objectives are PAC-learnable
- A Local Method for Identifying Causal Relations under Markov Equivalence
- Treatment Learning Causal Transformer for Noisy Image Classification
- Graph-attention-based Casual Discovery with Trust Region-navigated Clipping Policy Optimization
- A review of Gaussian Markov models for conditional independence
- REX: Causal discovery based on machine learning and explainability techniques
- Higher-order least squares: assessing partial goodness of fit of linear causal models
- Dagma-DCE: Interpretable, Non-Parametric Differentiable Causal Discovery
- Latent Instrumental Variables as Priors in Causal Inference based on Independence of Cause and Mechanism
- A Bayesian nonparametric test for conditional independence
- Learning interpretable causal networks from very large datasets, application to 400,000 medical records of breast cancer patients