Elements of Causal Learning: Basic Concepts, Theory, Methods, Algorithms and Applications
Abstract
The proliferation of information technology for nontraditional intelligence operations calls for a critical shift in the use of information; actionable decisions must be based causal and interventional evidence, rather than on correlational evidence. This proposal develops the foundations of a new paradigm, which we term causal learning. Reformulating the task of imputing missing potential outcomes as a prediction problem enables us to leverage machine learning to address key challenges of modern causal analyses, and to overcome limitations of purely statistical approaches.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jan 04, 2017
- Source ID
- N000141712131
Entities
People
- Edoardo Airoldi
Organizations
- Office of Naval Research
- President and Fellows of Harvard College
- United States Navy