[Ml-stat-talks] IDeAS Seminar on Thursday 5/24/12

Amit Singer amits at math.Princeton.EDU
Wed May 23 11:35:49 EDT 2012

Seminar:  IDeAS Seminar

Date:  Thursday, May 24, 2012

Time:  3:00 pm

Place:  Joseph Henry Room - Jadwin Hall - First Floor

Speaker: Ewout van den Berg, Stanford University

Title: Exploiting sparsity and low-dimensional structure: Two new


The success of compressed sensing in reconstructing sparse signals

from a small number of linear measurements has had a tremendous

impact on the signal processing community. Signal sparsity is now

routinely used and extensions to different types of low-dimensional

structure, such a low-rank matrices, have been successfully applied

and analyzed.


In this talk I will present two applications whose inherent

low-dimensionality can be leveraged to our advantage. The first

application concerns the design of an efficient high-resolution

silicon photomultiplier chip. By exploiting the spatio-temporal

sparsity in photon arrival, the proposed design achieves a high

level of sensitivity while maintaining a low circuit complexity.


The second application concerns volume reconstruction from

tilt-series data -- one of the central problems in cryo-electron

tomography. In the case of a heterogeneous data set the challenge

is to determine not only the relative orientation of the objects,

but also their class. We pose this problem as a parameterized

low-rank matrix recovery problem, which is then solved using

techniques from convex optimization.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cs.princeton.edu/pipermail/ml-stat-talks/attachments/20120523/3a01097f/attachment.html>

More information about the Ml-stat-talks mailing list