An Algorithm for Fast Recovery of Sparse Causal Graphs

By Peter Spirtes, Clark Glymour. 1991

How researchers see this paper

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

Citations in the corpus

Causal Markov condition for submodular information measures (2010, methodology): “The PC algorithm [14, 1] for causal inference takes a set of conditional independences on faithful objects and returns the equivalence class of faithful representations.”
DAGs with NO TEARS: Continuous Optimization for Structure Learning (2018, methodology): “Other approaches that avoid optimizing (4) directly include constraint-based methods (Spirtes and Glymour, 1991; Spirtes et al., 2000), hybrid methods (Tsamardinos et al., 2006; G´amez et al., 2011), and Bayesian methods (Ellis and Wong, 2008; Zhou, 2011; Niinim¨aki et al., 2016).”
Caveats for Causal Reasoning with Equilibrium Models (2001, methodology): “1 Constraint-Based Learning Constraint-Based (CB) learning methods [Verma and Pearl, 1991; Spirtes and Glymour, 1991; Pearl and Verma, 1991; Spirtes et al., 2000; Cheng et al., 2002] (also known as “conditional independence search” methods) build a CI model by systematically checking the data for co”
Detection of Unfaithfulness and Robust Causal Inference (2008, methodology): “The PC algorithm (Spirtes et al. 1991), for example, is provably correct for inferring the Markov equivalence class to which the true causal DAG belongs, from an oracle for conditional independence relations.”
Developing and Validating a Survival Prediction Model for NSCLC Patients Through Distributed Learning Across 3 Countries (2017, unknown): “In addition, we have learned a network structure on the data using an algorithmic approach (48).”
Learning Bayesian networks from data: An information-theory based approach (2002, methodology): “For example, when compared to the PC algorithm (which also uses dependency analysis to learn Bayesian net structures; [48], see Section 7), we found that TPDA typically uses fewer CI tests than PC uses, but these tests are often more complex.”
A survey of Bayesian Network structure learning (2021, methodology): “To counter these issues, the PC algorithm by Spirtes and Glymour (1991) performs the adjacency phase with conditioning sets of increasing size – checking all pairs of nodes 𝐴, 𝐵 at a particular conditioning set size and removing edges 𝐴 − 𝐵 if a Sepset is found before moving to higher…”
BCD Nets: Scalable Variational Approaches for Bayesian Causal Discovery (2021, methodology): “Also included are two methods designed speciﬁcally for Gaussian DAGs, the greedy directed acyclic graph search with equal error variance (GDSEEV) method [38], and the main method (‘Ghoshal’) from [21].”
Directed Cyclic Graphical Representations of Feedback Models (1995, background): “butions considered, there are correct polynomial time algorithms for inferring features of (sparse) directed acyclic graphs from a probability distribution when there are no latent common causes (see Spirtes and Glymour, 1991, Cooper and Herskovitz, 1992). Richardson's characterization of equiv”
Theory Refinement on Bayesian Networks (1991, background): “Some of these algorithms also make the assumption (Geiger et a!., 1990; Spirtes and Glymour, 1990; Verma and Pearl, 1990) that the unknown probability distribution is a DAG-isomorph (Pearl, 1988).”
Distributed learning: Developing a predictive model based on data from multiple hospitals without data leaving the hospital - A real life proof of concept. (2016, methodology): “To explore the effects of using a data-driven approach to determine the network structure, we have used the path condition (PC) algorithm to determine the network structure [35] based on the largest dataset (Maastro).”
DYNOTEARS: Structure Learning from Time-Series Data (2020, methodology): “Constraint-based meth-ods use conditional independence tests to recover the Markov equivalence class of DAGs under the assumption of faithfulness (e.g. Colombo et al., 2012; Spirtes and Glymour, 1991).”
Learning Causal Structures Using Regression Invariance (2017, methodology): “The best known algorithms for causal inference in the observational setup are IC [15] and PC [23] algorithms.”
Detecting crosstalk errors in quantum information processors (2019, methodology): “[54] Peter Spirtes, Clark N. Glymour, and Richard Scheines.”
Inferring causation from time series in Earth system sciences (2019, background): “…promise to substantially advance the state-of-the-art in understanding complex dynamical systems from data also in many other ﬁ elds with similar challenges as in Earth system sciences, if domain scientists and method developers closely work together — and join the ‘ causal revolution ’ 14 .”
A kernel-based causal learning algorithm (2007, methodology): “A refined version of IC is the PC algorithm (after its authors Spirtes and Glymour (1991)).”