Predicting causal effects in large-scale systems from observational data

By Marloes H. Maathuis, Diego Colombo, Markus Kalisch, Peter Buhlmann. Nature Methods, 2010

297 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Molecular Biology, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/ecf8033f122af4630a2d089bef4b448ead324666

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 Molecular Biology

In Statistics and Probability

Citations in the corpus

Learning high-dimensional directed acyclic graphs with latent and selection variables (2011, methodology): “This idea, together with a fast local implementation, forms the basis of the IDA algorithm [10, 11] which estimates bounds on causal eﬀects from observational data that are generated from an unknown causal DAG (IDA stands for Intervention calculus when the DAG is Absent).”
Inferring causal molecular networks: empirical assessment through a community-based effort (2016, methodology): “Our assessment approach shares similarities with other approaches in the literature, including those used by Maathuis et al. 38 , who focused on inferring networks from static observational data, and Olsen et al. 39 , who used a different scoring metric, considering predicted downstream targets in c”
Estimating individual treatment effect: generalization bounds and algorithms (2016, background): “Much recent work in machine learning for causal inference focuses on causal discovery, with the goal of discovering the underlying causal graph or causal direction from data (Hoyer et al., 2009; Maathuis et al., 2010; Triantafillou & Tsamardinos, 2015; Mooij et al., 2016).”
High-Dimensional Statistics with a View Toward Applications in Biology (2014, background): “However, it is still possible to identify lower bounds for |γj | with the so-called IDA (Inference when Dag is Absent) procedure (Maathuis et al., 2009, 2010).”
MAXIMUM LIKELIHOOD FOR SPARSE DIRECTED ACYCLIC GRAPHS (2013, methodology): “Directed acyclic graphs (DAGs) and corresponding directed graphical models are key concepts for causal inference; see, for example, the books by Spirtes, Glymour and Scheines (2000) and Pearl (2000). From an estimation point of view, a first step consists of estimating the Markov equivalence class o”
Controlling false positive selections in high-dimensional regression and causal inference (2013, methodology): “The figure is taken from the study of Maathuis et al.(26) Bühlmann et al.”
Structure Learning in Graphical Modeling (2016, background): “These ideas can be combined with the structure learning methods that we discuss in Section 4 to estimate (bounds on) causal eﬀects from observational data (Maathuis et al., 2009, 2010; Perkovi´c et al., 2015; Nandy et al., 2016b).”
Causal Confusion in Imitation Learning (2019, background): ““Causal discovery” approaches allow causal inference from pre-recorded observations under constraints [54, 17, 29, 15, 30, 31, 26, 14, 34, 57].”
PC algorithm for nonparanormal graphical models (2013, methodology): “PC algorithm can be found in Maathuis et al. (2010), Schmidberger et al. (2011), Le et al. (2013), and Verdugo et al. (2013).”
A Complete Generalized Adjustment Criterion (2015, methodology): “Such an approach is used in the IDA algorithm of Maathuis et al. (2009, 2010), by considering all DAGs represented by aCPDAG and applying back-door adjustment to each of theseDAGs.”
High-dimensional consistency in score-based and hybrid structure learning (2015, background): “3151 and its variants [e.g., Harris and Drton (2013), Colombo and Maathuis (2014)] have been widely applied to high-dimensional datasets [e.g., Maathuis et al. (2010), Schmidberger, Lennert and Mansmann (2011), Stekhoven et al. (2012), Verdugo et al. (2013), Le et al. (2013), Gao and Cui (2015)],…”
$\ell_0$-penalized maximum likelihood for sparse directed acyclic graphs (2012, methodology): “This strategy has been applied to, and to a certain extent validated using high-throughput, and hence high-dimensional, data in biology [Maathuis et al. (2010)].”
Causal Structure Learning (2017, background): “Moreover, the (possible) parents of X j can be used to estimate (bounds on) the total causal eﬀect of X j on any of the other variables (Maathuis et al., 2009, 2010; Stekhoven et al., 2012; Nandy et al., 2017b).”