Numerical Methods for Stochastic Modeling and Inference

3-0-9 units · Graduate

Advanced introduction to numerical methods for treating uncertainty in computational simulation. Draws examples from a range of engineering and science applications, emphasizing systems governed by ordinary and partial differential equations. Uncertainty propagation and assessment: Monte Carlo methods, variance reduction, sensitivity analysis, adjoint methods, polynomial chaos and Karhunen-Loève expansions, and stochastic Galerkin and collocation methods. Interaction of models with observational data, from the perspective of statistical inference: Bayesian parameter estimation, statistical regularization, Markov chain Monte Carlo, sequential data assimilation and filtering, and model selection.