Kernel-based Conditional Independence Test and Application in Causal Discovery

Abstract

Conditional independence testing is an important problem, especially in Bayesian network learning and causal discovery. Due to the curse of dimensionality, testing for conditional independence of continuous variables is particularly challenging. We propose a Kernel-based Conditional Independence test (KCI-test), by constructing an appropriate test statistic and deriving its asymptotic distribution under the null hypothesis of conditional independence. The proposed method is computationally efficient and easy to implement. Experimental results show that it outperforms other methods, especially when the conditioning set is large or the sample size is not very large, in which case other methods encounter difficulties.

References in corpus (1)

• On the Identifiability of the Post-Nonlinear Causal Model

Cited by in corpus (96)

• Equivalence of distance-based and RKHS-based statistics in hypothesis testing
• Kernel Mean Embedding of Distributions: A Review and Beyond
• Learning Decentralized Controllers for Robot Swarms with Graph Neural Networks
• Causal Discovery with Continuous Additive Noise Models
• Invariant Models for Causal Transfer Learning
• The Hardness of Conditional Independence Testing and the Generalised Covariance Measure
• Causality for Machine Learning
• Causal Discovery with Reinforcement Learning
• Causal interpretation rules for encoding and decoding models in neuroimaging
• The Randomized Dependence Coefficient
• Large-Scale Kernel Methods for Independence Testing
• Quasi-Monte Carlo Feature Maps for Shift-Invariant Kernels
• Learning Neural Causal Models from Unknown Interventions
• Conditional independence testing based on a nearest-neighbor estimator of conditional mutual information
• Discovering contemporaneous and lagged causal relations in autocorrelated nonlinear time series datasets
• Interpreting Black Box Models via Hypothesis Testing
• Causal Inference via Kernel Deviance Measures
• PPFS: Predictive Permutation Feature Selection
• Characterizing and Learning Equivalence Classes of Causal DAGs under Interventions
• Causality-based Feature Selection: Methods and Evaluations
• Algorithmic Decision Making with Conditional Fairness
• Structural Agnostic Modeling: Adversarial Learning of Causal Graphs
• Generalized Optimal Matching Methods for Causal Inference
• Domain Adaptation as a Problem of Inference on Graphical Models
• A Kernel Test for Three-Variable Interactions
• Nonlinear Invariant Risk Minimization: A Causal Approach
• Forward-Backward Selection with Early Dropping
• A Wild Bootstrap for Degenerate Kernel Tests
• Learning Sparse Nonparametric DAGs
• Learning and Testing Causal Models with Interventions
• Causal Generative Neural Networks
• Causal Generative Domain Adaptation Networks
• On conditional parity as a notion of non-discrimination in machine learning
• Information Flow in Computational Systems
• Non-Parametric Causality Detection: An Application to Social Media and Financial Data
• Differentiable Causal Discovery from Interventional Data
• A Distribution Free Conditional Independence Test with Applications to Causal Discovery
• Estimating Transfer Entropy via Copula Entropy
• Causal Discovery from Heterogeneous/Nonstationary Data with Independent Changes
• SLANTS: Sequential Adaptive Nonlinear Modeling of Vector Time Series
• Conditional Sum-Product Networks: Imposing Structure on Deep Probabilistic Architectures
• Causal Inference in the Presence of Interference in Sponsored Search Advertising
• Necessary and sufficient conditions for causal feature selection in time series with latent common causes
• Exponential Concentration of a Density Functional Estimator
• Kernel Partial Correlation Coefficient -- a Measure of Conditional Dependence
• Identification of Causal Diffusion Effects Under Structural Stationarity
• Fast Conditional Independence Test for Vector Variables with Large Sample Sizes
• Does the Markov Decision Process Fit the Data: Testing for the Markov Property in Sequential Decision Making
• hyppo: A Multivariate Hypothesis Testing Python Package
• Hypothesis testing using pairwise distances and associated kernels (with Appendix)
• Causality-based tests to detect the influence of confounders on mobile health diagnostic applications: a comparison with restricted permutations
• Towards Federated Bayesian Network Structure Learning with Continuous Optimization
• Higher Order Kernel Mean Embeddings to Capture Filtrations of Stochastic Processes
• Causal Discovery for Manufacturing Domains
• SADA: A General Framework to Support Robust Causation Discovery with Theoretical Guarantee
• Model-based causal feature selection for general response types
• Discovery and Visualization of Nonstationary Causal Models
• AdaRL: What, Where, and How to Adapt in Transfer Reinforcement Learning
• copent: Estimating Copula Entropy and Transfer Entropy in R
• Efficient Intervention Design for Causal Discovery with Latents
• A low variance consistent test of relative dependency
• On Distance and Kernel Measures of Conditional Independence
• The FEDHC Bayesian network learning algorithm
• Double Generative Adversarial Networks for Conditional Independence Testing
• A Critical View of the Structural Causal Model
• A Kernel Test for Three-Variable Interactions with Random Processes
• FRITL: A Hybrid Method for Causal Discovery in the Presence of Latent Confounders
• Robust Distance Covariance
• A Junction Tree Framework for Undirected Graphical Model Selection
• Nonparametric causal structure learning in high dimensions
• On Azadkia-Chatterjee's conditional dependence coefficient
• Merging joint distributions via causal model classes with low VC dimension
• A Bayesian nonparametric test for conditional independence
• Nonparametric Causal Feature Selection for Spatiotemporal Risk Mapping of Malaria Incidence in Madagascar
• independence: Fast Rank Tests
• Predictive Independence Testing, Predictive Conditional Independence Testing, and Predictive Graphical Modelling
• Hierarchically Compositional Kernels for Scalable Nonparametric Learning
• Learning from Pairwise Marginal Independencies
• Consistency of permutation tests for HSIC and dHSIC
• Learning local directed acyclic graphs based on multivariate time series data
• A New Measure of Conditional Dependence
• Accelerating Recursive Partition-Based Causal Structure Learning
• Obtaining Causal Information by Merging Datasets with MAXENT
• Massively-Parallel Feature Selection for Big Data
• A Bayesian Nonparametric Conditional Two-sample Test with an Application to Local Causal Discovery
• Robust regression for optimal individualized treatment rules
• On the Sample Complexity of Causal Discovery and the Value of Domain Expertise
• Causal evidence for the primordiality of colours in trans-Neptunian objects
• Distance-based regression analysis for measuring associations
• Fast Non-Parametric Tests of Relative Dependency and Similarity
• Conditional Randomization Rank Test
• A U-statistic Approach to Hypothesis Testing for Structure Discovery in Undirected Graphical Models
• Why did the distribution change?
• Inference on the structure of gene regulatory networks
• Learning binary undirected graph in low dimensional regime
• On a Nonparametric Notion of Residual and its Applications