FRITL: A Hybrid Method for Causal Discovery in the Presence of Latent Confounders

We consider the problem of estimating a particular type of linear non-Gaussian model. Without resorting to the overcomplete Independent Component Analysis (ICA), we show that under some mild assumptions, the model is uniquely identified by a hybrid method. Our method leverages the advantages of constraint-based methods and independent noisebased methods to handle both confounded and unconfounded situations. The first step of our method uses the FCI procedure, which allows confounders and is able to produce asymptotically correct results. The results, unfortunately, usually determine very few unconfounded direct causal relations, because whenever it is possible to have a confounder, it will indicate it. The second step of our procedure finds the unconfounded causal edges between observed variables among only those adjacent pairs informed by the FCI results. By making use of the so-called Triad condition, the third step is able to find confounders and their causal relations with other variables. Afterward, we apply ICA on a notably smaller set of graphs to identify remaining causal relationships if needed. Extensive experiments on simulated data and real-world data validate the correctness and effectiveness of the proposed method.

[1]  Richard Scheines,et al.  Causation, Prediction, and Search, Second Edition , 2000, Adaptive computation and machine learning.

[2]  Marianne Winslett,et al.  Understanding Social Causalities Behind Human Action Sequences , 2017, IEEE Transactions on Neural Networks and Learning Systems.

[3]  Visa Koivunen,et al.  Identifiability, separability, and uniqueness of linear ICA models , 2004, IEEE Signal Processing Letters.

[4]  Peter Spirtes,et al.  A Hybrid Causal Search Algorithm for Latent Variable Models , 2016, Probabilistic Graphical Models.

[5]  Diego Colombo,et al.  Order-independent constraint-based causal structure learning , 2012, J. Mach. Learn. Res..

[6]  E. Lukács,et al.  A Property of the Normal Distribution , 1954 .

[7]  Aapo Hyvärinen,et al.  ParceLiNGAM: A Causal Ordering Method Robust Against Latent Confounders , 2013, Neural Computation.

[8]  Thomas S. Richardson,et al.  Learning high-dimensional directed acyclic graphs with latent and selection variables , 2011, 1104.5617.

[9]  G. Darmois,et al.  Analyse générale des liaisons stochastiques: etude particulière de l'analyse factorielle linéaire , 1953 .

[10]  Juliane Freud,et al.  Characterization Problems In Mathematical Statistics , 2016 .

[11]  K. Sachs,et al.  Causal Protein-Signaling Networks Derived from Multiparameter Single-Cell Data , 2005, Science.

[12]  Jiji Zhang,et al.  Adjacency-Faithfulness and Conservative Causal Inference , 2006, UAI.

[13]  T. Heskes,et al.  Learning Sparse Causal Models is not NP-hard , 2013, UAI.

[14]  Clark Glymour,et al.  Estimating feedforward and feedback effective connections from fMRI time series: Assessments of statistical methods , 2019, Network Neuroscience.

[15]  Jiji Zhang,et al.  On the completeness of orientation rules for causal discovery in the presence of latent confounders and selection bias , 2008, Artif. Intell..

[16]  Aapo Hyvärinen,et al.  DirectLiNGAM: A Direct Method for Learning a Linear Non-Gaussian Structural Equation Model , 2011, J. Mach. Learn. Res..

[17]  Patrik O. Hoyer,et al.  Discovering Unconfounded Causal Relationships Using Linear Non-Gaussian Models , 2010, JSAI-isAI Workshops.

[18]  Bernhard Schölkopf,et al.  Kernel-based Conditional Independence Test and Application in Causal Discovery , 2011, UAI.

[19]  Terrence J. Sejnowski,et al.  Learning Overcomplete Representations , 2000, Neural Computation.

[20]  Ruichu Cai,et al.  Triad Constraints for Learning Causal Structure of Latent Variables , 2019, NeurIPS.

[21]  Russell A. Poldrack,et al.  Six problems for causal inference from fMRI , 2010, NeuroImage.

[22]  Judea Pearl,et al.  Probabilistic reasoning in intelligent systems - networks of plausible inference , 1991, Morgan Kaufmann series in representation and reasoning.

[23]  Patrik O. Hoyer,et al.  Estimation of causal effects using linear non-Gaussian causal models with hidden variables , 2008, Int. J. Approx. Reason..