(See past events)
Keri Kornelson, Oklahoma.
Friday, September 21 at 4:00pm in Ritter 231 with refreshments beforehand in the Ritter Hall Lobby.
Norm retrieval via spatiotemporal samples
Abstract: There is a relaxation of the problem of phase retrieval in which the magnitude of a signal is computed from phaseless measurements. We require less information, so can be possible with fewer measurements than phase retrieval. As a ready example, an orthonormal basis yields norm retrieval measurements. In this talk, we introduce the concepts and earlier results about performing phase and norm retrieval in real, finite-dimensional space. We then present recent work with Fatma Bozkurt in which we do norm retrieval with dynamical samples, i.e. samples obtained at selected measurement points but repeated over time.
Qayum Khan, Saint Louis U.
Tuesday, September 25, 4:10pm-5:00pm in Ritter Hall 216
Compact Hausdorff groups are pro–Lie
Abstract: This is a learning talk. We go through the Pontrjagin–Weil proof of von Neumann's 1933 theorem, that any compact Hausdorff group is the projective limit of Lie groups, using the Peter–Weyl theorem from classical harmonic analysis. Graduate students who have taken or are taking General Topology I (point-set topology) are encouraged to attend. Graduate students are also encouraged to give a later talk about their research or about an interesting fact in geometry or topology.
Tim Keller, Saint Louis U.
Thursday, September 27 at 3:00-4:00pm in Ritter 204
Stratified Simple Random Sampling with Multiple Estimation Objectives
Abstract: Non-response is the greatest challenge facing establishment surveys. A major contributing factor for survey non-response is respondent burden. To meet this challenge survey establishments must therefore strive to meet estimation goals with the smallest possible sample size.
A common and basic survey design is the stratified simple random sample, and a common estimate of interest is the population total for a survey item. For this special case, the problem of meeting multiple estimation objectives is formulated as a convex optimization problem, and a numerical method for determining the optimal allocation of a fixed overall sample size is presented.
Elodie Pozzi, Saint Louis U.
Friday, October 5 at 4:00pm in Ritter 231 with refreshments beforehand in the Ritter Hall Lobby.
A 2D inverse problem in magnetism
The Math/Stat Colloquium this semester will meet at 4:00pm in RH 231 with refreshments beforehand at 3:30 in the Ritter Lobby.
Our current schedule:
- Friday, September 7, Haiyan Cai, UMSL (Analysis)
- Friday, September 21, Keri Kornelson, Oklahoma Univ. (Analysis)
- Friday, October 5, Elodie Pozzi, Saint Louis University
- Friday, October 19, Ken Jacobs, Northeastern Univ. (Algebra)
- Friday, November 2, Chris Connell, Indiana Univ. (Topology)
- Wednesday, November 14, Shmuel Wienberger, Univ. of Chicago (Topology)