Causal Inference by Surrogate Experiments: z-Identifiability

Abstract

We address the problem of estimating the effect of intervening on a set of variables X from experiments on a different set, Z, that is more accessible to manipulation. This problem, which we call z-identifiability reduces to ordinary identifiability when Z = phi and like the latter, can be given syntactic characterization using the do-calculus [Pearl, 1995; 2000]. We provide a graphical necessary and sufficient condition for z- identifiability for arbitrary sets X,Z, and Y (the out- comes). We further develop a complete algorithm for computing the causal effect of X on Y using information provided by experiments on Z. Finally, we use our results to prove completeness of do-calculus relative to z-identifiability, a result that does not follow from completeness relative to ordinary identifiability.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jun 01, 2012
Accession Number
ADA564088

Entities

People

  • Elias Bareinboim
  • Judea Pearl

Organizations

  • University of California, Los Angeles

Tags

Communities of Interest

  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Algorithms
  • Artificial Intelligence
  • Bayesian Networks
  • Calculus
  • Causal Reasoning
  • Computer Science
  • Construction
  • Experimental Data
  • Heart Diseases
  • Identification
  • Intervention
  • Models
  • New York
  • Probability
  • Probability Distributions
  • Reasoning
  • Uncertainty

Fields of Study

  • Computer science

Readers

  • Artificial Intelligence
  • Computational Linguistics
  • Linear Algebra

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference