Speaker:
Andrea Montanari 

Session 2

Date and Time: 1.00 PM Jan 31st, CSL B02

Abstract:
The classical statistical estimation problem requires to reconstruct an unknown vector of parameters from a set of observations, when the two are connected through a stochastic relationship. Over the last ten years, a whole new generation of statistical estimation problems has emerged, posing fascinating new challenges to the existing theory. Prominent examples include the image reconstruction problems arising in magnetic resonance imaging (MRI), exploration seismology, radar imaging or hyperspectral imaging. While these problems are characterized by a sharp increase in the amount of available data and computational resources, both are outpaced by the increase in dimensionality of the unknown vector of parameters. I will describe a class of reconstruction algorithms that emphasize statistical efficiency under bounded computational resources. These algorithms are known as approximate message passing (AMP) algorithms and are inspired by ideas in graphical models. [Based on joint work with D. Donoho, I. Johnstone, A. Maleki]