Information Methods for Uncertainty Quantification and Performance Guarantees in Predictive Modeling

Abstract

The proposed research is the development of the foundations of Uncertainty Quantification for complex systems. The novelty of the proposal is a broad and general approach to the systematic development of UQ methods using tools from applied and computational probability, information theory, and quantum and statistical mechanics to assess and measure predictive bias for a range of quantities of interest and statistical estimators. Our methods can discriminate and assess multiple sources of model uncertainties, can treat models with multiple scales and those driven by rare events, and provide the first systematic approach for Uncertainty Quantification in quantum systems.

Document Details

Document Type
DoD Grant Award
Publication Date
Jul 24, 2019
Source ID
FA95501810530

Entities

People

  • Paul Dupuis

Organizations

  • Air Force Office of Scientific Research
  • Brown University
  • United States Air Force

Tags

Readers

  • Computational Modeling and Simulation
  • Neural Network Machine Learning.

Technology Areas

  • Quantum Computing