Predicting Proliferation: High Reliability Forecasting Models of Nuclear

Abstract

Scholars have spent decades studying and explaining nuclear proliferation. This project will develop a model to predict the behavior of states regarding their pursuit and acquisition of nuclear weapons. An accurate prediction model will allow for action against potential suppliers, interdiction of nuclear trade, intelligence collection on covert nuclear activities, and credible military action against countries of concern. The model will not only be a yes or no predictor; it will assess the probability of a state pursuing or acquiring nuclear weapons under a particular set of conditions. It is hoped that this model will be replicated and adjusted to suit the needs of analysts and scholars. Additionally, the model will provide the foundation for future predictive work. To create the model, researchers will leverage 65 years of data on the characteristics of states that pursue nuclear weapons. They will examine mechanisms that particular countries use in their nuclear pursuits. This model will harness the collective wisdom and data on past nuclear weapons programs. The vetted data will be used to construct a set of predictions of states’ proliferation under a variety of scenarios. In addition, the project will use statistical learning approaches (e.g., neural networks, support vector machines, and methods to combine multiple statistical models). Once built, the model will be tested using separate data. This model, built upon flexible quantitative techniques, will identify an empirically grounded set of triggers or conditions under which countries are most likely to shift from latent nuclear capability to a full-fledged nuclear weapons effort.

Document Details

Document Type
DoD Grant Award
Publication Date
Apr 28, 2016
Source ID
N002441510003

Entities

People

  • Erik Gartzke

Organizations

  • Office of the Secretary of Defense
  • University of California, San Diego

Tags

Fields of Study

  • Physics

Readers

  • Computational Modeling and Simulation
  • Strategic Security Studies

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference