Friday, May 08, 2009

CS: Phase transitions phenomenon in CS video, jobs and a seminar


I just stumbled upon a recent video by Jared Tanner on Phase transitions phenomenon in Compressed Sensing and as usual he is always asking the right questions! The abstract reads:
Compressed Sensing reconstruction algorithms typically exhibit a zeroth-order phase transition phenomenon for large problem sizes, where there is a domain of problem sizes for which successful recovery occurs with overwhelming probability, and there is a domain of problem sizes for which recovery failure occurs with overwhelming probability. The mathematics underlying this phenomenon will be outlined for $\ell1$ regularization and non-negative feasibility point regions. Both instances employ a large deviation analysis of the associated geometric probability event. These results give precise if and only if conditions on the number of samples needed in Compressed Sensing applications. Lower bounds on the phase transitions implied by the Restricted Isometry Property for Gaussian random matrices will also be presented for the following algorithms: $\ell^q$-regularization for $q\in (0,1]$, CoSaMP, Subspace Pursuit, and Iterated Hard Thresholding.
The slides are here. I'll cover the other presentations of that meeting related to CS later.

Gabriel Peyre just told me of a new version for a paper I mentioned earlier and written with Jalal M. Fadili it is here:Total Variation Projection with First Order Schemes

Ramesh Raskar , the head of the Camera Culture Lab at MIT is looking to hire Graduate Students, MEng, PostDocs and UROPs starting Fall 2009. You may want to look here for more information. I added this information on the CS Jobs page.

The seminar organized by Patrick Louis Combettes and Albert Cohen will feature the following speakers:

Mardi 16 juin 2009
À 9h30 :Guillermo Sapiro (University of Minnesota, États-Unis)
À 10h30 (cours de 2h) : Ron DeVore (Texas Agricultural and Mechanical University, États-Unis)


Mardi 26 mai 2009

Simon Foucart (Vanderbilt University, États-Unis, "Sparse recovery via l_q minimization with q ε ]0,1]"
À 11h30 :Gabriel Peyré (Paris-Dauphine)
I still vote for Texas A&M instead of the full name :-)

No comments:

Printfriendly