Computational Mathematics and Statistics Seminar by Srinivas Eswar: Randomized Approaches for Optimal Experimental Design

Speaker: Srinivas  Eswar, Argonne National Laboratory

Title: Randomized Approaches for Optimal Experimental Design

Abstract: This work describe connections between optimal experiment design (OED) for PDE-based Bayesian linear inverse problems and the column subset selection problem (CSSP) in matrix approximation.We derive bounds, both lower and upper, for the D-optimality criterion via CSSP for the independent and colored noise cases.Additionally, we describe ways to interpolate "left-out" sensor data using the "selected" sensors along with the errors in the data completion process.We develop and analyse randomized algorithms which achieve these bounds.Finally, we experimentally verify these results on a model advection-diffusion problem.

 

Computational Mathematics and Statistics Seminar

Zoom Link