On the completeness of orientation rules for causal discovery in the presence of latent confounders and selection bias

By Jiji Zhang. 2008

523 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Information Systems

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Information Systems

Citations in the corpus

Domain Adaptation by Using Causal Inference to Predict Invariant Conditional Distributions (2017, methodology): “ch as [Hyttinen et al., 2014, Triantaﬁllou and Tsamardinos, 2015, Magliacane et al., 2016], are ideally suited to exploit the background knowledge. For other methods, e.g., FCI [Spirtes et al., 2000, Zhang, 2008], incorporating all background knowledge is less straightforward and as far as we know c”
Distinguishing Cause from Effect Using Observational Data: Methods and Benchmarks (2014, background): “For example, in that case, and under a genericity condition called “faithfulness”, conditional independences between subsets of observed variables allow one to draw partial conclusions regarding their causal relationships (Spirtes et al., 2000; Pearl, 2000; Richardson and Spirtes, 2002; Zhang, 2008)”
Order-independent constraint-based causal structure learning (2012, methodology): “We start by discussing several prominent examples of such algorithms, designed for diﬀerent settings.”
Causal Structure Learning (2017, background): “There are also additional orientation rules, which were shown to be complete in Zhang (2008b).”
DAG-GNN: DAG Structure Learning with Graph Neural Networks (2019, methodology): “Popular algorithms include SGS (Spirtes et al., 2000b), PC (Spirtes et al., 2000b), IC (Pearl, 2003), and FCI (Spirtes et al., 1995; Zhang, 2008).”
D’ya Like DAGs? A Survey on Structure Learning and Causal Discovery (2021, unknown): “PC [228] 1993 constraint yes yes yes no CPDAG FCI [229, 275] 1995 constraint no yes yes no POIPG CCD [200] 1996 constraint yes yes both no PAG TPDA [31] 2002 constraint yes yes yes no CPDAG CPC [193] 2006 constraint yes relaxed yes no CPDAG KCL [233] 2007 constraint yes yes yes no CPDAG ION [241] 20”
Foundations of structural causal models with cycles and latent variables (2016, methodology): “Recently, it has been shown that even the well-known Fast Causal Inference (FCI) algorithm [69, 75] is directly applicable to simple SCMs [41] and provides a consistent estimate of the Markov equivalence class (under the faithfulness assumption).”
Causal Transfer Learning (2017, methodology): “For other methods, e.g., FCI [Spirtes et al., 2000, Zhang, 2008], incorporating all background knowledge is less straightforward and as far as we know cannot be done with off-the-shelf implementations.”
Structure Learning in Graphical Modeling (2016, methodology): “The FCI algorithm also uses more complicated orientation rules, which were extended and proved to be complete by Zhang (2008).”
Controlling Selection Bias in Causal Inference (2011, background): “Remarkably, there are special cases in which selection bias can be detected even from observations, as in the form of a non-chordal undirected component (Zhang, 2008).”
A Complete Generalized Adjustment Criterion (2015, background): “Thend∗ is an unshielded possibly directed path fromQ to Z in P (Lemma B.1 from Zhang (2008)).”
A survey of Bayesian Network structure learning (2021, methodology): “Ancestral graphs have three types of edge (Zhang, 2008b): • directed, e.g. 𝐴 ⟶ 𝐵: The mapping between edge types in the ancestral graph and relationships in the underlying DAG is given in Richardson and Spirtes (2002) but is somewhat complicated.”
Comparison of Strategies for Scalable Causal Discovery of Latent Variable Models from Mixed Data (2018, methodology): “Finally, all edges are reoriented to have both circle endpoints, and then orientation rules are applied as given by Zhang [30].”
Learning Sparse Causal Models is not NP-hard (2013, methodology): “…is order O ( N 4 ): for N 2 edge marks it checks for certain paths in G which can be done in O ( N 2 k ) by a ‘reachability’ algorithm, with Nk ∼ nr. of edges, (Zhang, 2008, p.1881) The dominant terms are O ( N 2 k ) possible hierarchies to test for each of O ( N 2 ) possible edges which may…”
Recovering from Selection Bias in Causal and Statistical Inference (2014, background): “1 Remarkably, there are special situations in which selection bias can be detected even from observations, as in the form of a non-chordal undirected component (Zhang 2008). the data-generating model in Fig.”
Distinguishing Cause from Effect Using Observational Data: Methods and Benchmarks (2014, background): “For example, in that case, and under a genericity condition called “faithfulness”, conditional independences between subsets of observed variables allow one to draw partial conclusions regarding their causal relationships (Spirtes et al., 2000; Pearl, 2000; Richardson and Spirtes, 2002; Zhang, 2008)”