65th ISI World Statistics Congress

65th ISI World Statistics Congress

Batch-based Bayesian Optimal Experimental Design in Inverse Problems

Author

SM
Sofia Mäkinen

Co-author

Conference

65th ISI World Statistics Congress

Format: IPS Abstract - WSC 2025

Keywords: bayesian, inverse-problem, inverseproblems, optimal design

Abstract

Optimal experimental design (OED) aims to optimize the measurement setting of a problem so that the information obtained from observed variables is maximized. In this presentation, we adopt the Bayesian approach to OED and consider batch-based OED for inverse problems. Batch-based OED aims to optimize several design parameters simultaneously, which is beneficial especially in situations where it is convenient to carry out multiple measurements at once. We present a numerical method for the batch-based OED optimization problem and consider the related theoretical aspects.