Transportability of Causal and Statistical Relations: A Formal Approach

By Judea Pearl, Elias Bareinboim. 2011

356 citations recorded by the source corpus.

We address the problem of transferring information learned from experiments to a different environment, in which only passive observations can be collected. We introduce a formal representation called ``selection diagrams'' for expressing knowledge about differences and commonalities between environments and, using this representation, we derive procedures for deciding whether effects in the target environment can be inferred from experiments conducted elsewhere. When the answer is affirmative, the procedures identify the set of experiments and observations that need be conducted to license the transport. We further discuss how transportability analysis can guide the transfer of knowledge in non-experimental learning to minimize re-measurement cost and improve prediction power.

Subfields: Causal Inference & Discovery

How researchers see this paper

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

In Causal Inference & Discovery

Citations in the corpus

Counterfactual Normalization: Proactively Addressing Dataset Shift and Improving Reliability Using Causal Mechanisms (2018, background): “Similarly, in causal inference transportability , formalized in Pearl and Bareinboim (2011), transfers causal effect estimates from one environment to another.”
Multi-Source Domain Adaptation: A Causal View (2015, methodology): “They, for instance, help characterize transportability of experimental ﬁndings (Pearl and Bareinboim 2011) or recoverability from selection bias (Bareinboim, Tian, and Pearl 2014).”
Causal Transportability with Limited Experiments (2013, background): “for representing differences and commonalities among domains, environments, and populations [Pearl and Bareinboim, 2011] (henceforth, PB).”
Transfer Learning with Label Noise (2017, background): “Figure 1 represents the causal relations between variables in transfer learning using selection diagram deﬁned in [18]. s and V 2 s change P Y and P X | Y , respectively.”
Semi-supervised Learning in Causal and Anticausal Settings (2013, background): “Pearl and Bareinboim [11] introduce a variable S that labels different domains or datasets and explains how the way in which S is causally linked to variables of interest is relevant for transferring causal or statistical statements across domains.”
Preventing Failures Due to Dataset Shift: Learning Predictive Models That Transport (2018, methodology): “A natural representation is to use selection diagrams (Pearl and Bareinboim, 2011), which consist of two types of nodes: nodes representing variables relevant to the DGP (e.”
Bayesian Workflow (2020, methodology): “Inmachine learning and causal inference, the terms “domain adaptation” and “transportability” have been used to represent the challenges of taking inference from a particular dataset and applying it in a new problem (Blitzer, Dredze, and Pereira, 2007, Pearl and Bareinboim, 2011).”
A Survey of Learning Causality with Data (2018, background): “2020. 1:28 R. Guo et al. of the predictors remain the same, one may allow apriori causal knowledge to guide the learning process and circumvent the discrepancies between the source and target domains [115], a.k.a. causal transportability. The study of transportability seeks to identify conditions un”
TRYGVE HAAVELMO AND THE EMERGENCE OF CAUSAL CALCULUS (2014, unknown): “A principled treatment of the problem of “external validity” (Campbell and Stanley, 1963), including, formal methods of deciding if a causal relation estimated in one population can be transported to another population, in which experimental conditions are different (Pearl and Bareinboim, 2011).”
Adjustment Criteria for Generalizing Experimental Findings (2019, background): “Different conditions were studied in the literature, for instance, in (Pearl and Bareinboim, 2011; Bareinboim et al., 2013; Bareinboim and Pearl, 2014).”
Controlling Selection Bias in Causal Inference (2011, background): “This paper complements recent work on transportability (Pearl and Bareinboim, 2011) which deals with transferring causal information from one environment to another, in which only passive observations can be collected.”
Generalizing causal inferences from randomized trials: counterfactual and graphical identification (2019, methodology): “Bareinboim and Pearl have proposed general methods to assess identifiability and, when possible, to identify the post-intervention distribution in a target population by “transporting” information from other sources [47–49].”
A Survey of Learning Causality with Data: Problems and Methods (2018, background): “43 is violated but the marginal distribution of predictors remain, one may allow apriori causal knowledge to guide the learning process and circumvent the discrepancies between the source and target domains [106], a.”
The Do-Calculus Revisited (2012, background): “This problem, labeled “transportability” in [Pearl and Bareinboim, 2011] can also be reduced to syntactic operation using the do -calculus but here the aim will be to separate the do -operator from a set S of variables, that indicate disparities between the two environment.”
Understanding and misunderstanding randomized controlled trials (2016, unknown): “Pearl and Bareinboim (2011, 2014) and Bareinboim and Pearl (2013, 2014) provide strategies for inferring information about new populations from trial results that are more general than reweighting.”
A Review of Generalizability and Transportability (2021, methodology): “A corresponding set of assumptions under the structural equation model (SEM) framework has also been derived (Pearl and Bareinboim, 2014; Pearl, 2015; Pearl and Bareinboim, 2011; Bareinboim and Pearl, 2014; Bareinboim and Tian, 2015; Bareinboim and Pearl, 2016; Correa, Tian and Bareinboim, 2018).”