December 4, 2012

Mathematics Colloquium today

By Reta McDermott

At 2:30 p.m. today, Mark Iwen from Duke University will speak about "Compressed Sensing: Extensions and Applications" in 122 Cardwell Hall.

Abstract: This talk will give a brief introduction to the theory of compressed sensing developed over the past several years, and then discuss several extensions and applications to related problems. The first extension we consider involves the development of practical recovery algorithms for compressed sensing of signals which are accurately represented via manifold models as opposed to more standard linear dictionary models.

Next, we will consider sketching applications involving massive data sets, and demonstrate the ability of compressed sensing theory to develop fast approximation algorithms for streamed data. Time permitting, we will conclude with an application of sparse recovery techniques to the solution of special types of inverse problems involving the recovery of words from noisy data. In all of these cases the emphasis will be on the development of algorithms with rigorous theoretical accuracy, stability and run time guarantees.