Transportability of Causal and Statistical Relations: A Formal Approach

Abstract

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.

Big Data Curation

“…The advancement of automated data curation algorithms will depend on the definition of theoretical models and on the investigation of the principles behind data curation (Buneman et al 2008). Understanding the causal mechanisms behind workflows (Cheney 2010) and the generalization conditions behind data transportability (Pearl and Bareinboim 2011) are examples of theoretical models that can impact data curation, guiding users towards the generation and representation of data that can be reused in broader contexts.…”

Identifiability and transportability in dynamic causal networks

“…Pearl and Bareinboim [22] introduced the sID algorithm, based on do-calculus, to identify a transport formula between two domains, where the effect in a target domain can be estimated from experimental results in a source domain and some observations on the target domain, thus avoiding the need to perform an experiment on the target domain.…”

“…This distinction is genuinely new to DCN, as it appears neither in DBN nor in standard causal graphs, yet the presence or absence of hidden dynamic confounders has crucial impacts on the postintervention evolution of the system over time and on the computational cost of the algorithms. -Finally, we extend from standard Causal Graphs to DCN the results by Pearl and Bareinboim [22] on transportability, namely on whether causal effects obtained from experiments in one domain can be transferred to another domain with similar causal structure. This opens up the way to studying relational knowledge transfer learning [19] of causal information in domains with a time component.…”

Structural Counterfactuals: A Brief Introduction

“…Finally, as the last application, I point to a recent theory of "transportability" (Pearl & Bareinboim, 2011) which provides a formal solution to the century-old problem of "external validity" (Campbell & Stanley, 1966); i.e., under what conditions can experimental findings be transported to another environment, how the results should be calibrated to account for environmental differences, and what measurements need be taken in each of the two environments to license the transport.…”