Theory and Computation for Gaussian Processes

Michael Stein
University of Chicago

I will review work on model estimation and prediction for Gaussian processes in spatial statistics, touching on asymptotic statistical theory and computation for large, irregularly sited spatial datasets.
I will then discuss what relevance this work might have for the use of Gaussian processes for emulation of deterministic computer models, with, where possible, references to many particle systems.

Presentation (PDF File)

Back to Machine Learning for Many-Particle Systems