Mathematics 797PM - ST-Predictv Mod & Uncert Quant

In this course we will discuss an array of topics that straddle probabilistic modeling and simulation, uncertainty quantification and statistical learning methods. We focus on developing systematic mathematical and computational tools for building data-driven, predictive models for complex systems and dynamics. In particular, we will discuss material that includes: uncertainty quantification, local and global sensitivity analysis methods for dynamical systems, data assimilation and approximate inference methods, stochastic optimization algorithms and model selection. A significant component of the course focuses on related software such as DAKOTA and others.

Open to Graduate students only. Prerequisites include instructor consent and graduate courses in probability theory (STAT 605), numerical analysis for ODE/PDE (MATH 652) and familiarity with the basic concepts of statistical inference (STAT 608).