Title: Computational Models in Scientific Inference - Promise and Challenge

• Speaker: David Higdon, Research Staff Member, Los Alamos National Laboratory
• Date: Thursday, February 27, 2014
• Time: 10:00 to 11:00 AM
• Location: Virginia Tech Research Center, 900 N. Glebe Road, Arlington, VA 22203
• Directions: Virginia Tech Research Center is located just a few blocks from the Ballston Metro. Underground parking is available. The entrance to the garage is located off 9th Street. When you arrived at the Virginia Tech Research Center, proceed to the second floor reception desk and ask for Kim Lyman (klyman@vbi.vt.edu, 571-858-3100).
• Sponsor: Social & Decision Analytics Laboratory, Virginia Bioinformatics Institute at Virginia Tech.

Abstract:

Modern computing has made it possible to model complex physical phenomena with an impressive degree of detail and realism. An ever-growing collection of application areas includes: climate, nuclear physics, material science, geophysics, and cosmology. In order to use such models to make predictions in new, untested settings, some physical measurements are required to constrain model parameters and to build error models that connect the model to the true physical system. This talk will describe the statistical principles involved with such efforts, with a focus on the challenge of extrapolation - i.e. how can we attach uncertainty to predictions in such cases? Examples from high-energy physics will be used to motivate statistical approaches described in this talk.