|Mittagsseminar Talk Information|
Date and Time: Thursday, November 10, 2011, 12:15 pm
Duration: 30 minutes
Location: CAB G51
Speaker: Martin Jaggi
The study of sparse solutions to some optimization problems (e.g. underdetermined linear systems) has lead to many interesting new applications in the last few years. This approach is known as compressed sensing, or group testing. The main observation is that very often, this specific class of (hard) combinatorial problems can be solved very efficiently by solving an (easy) convex relaxation of the problem instead.
Here we will discuss a recent generalization of the measure of "sparsity", allowing to incorporate the structure of the underlying solutions, as proposed by
Guillaume Obozinski, Laurent Jacob and Jean-Philippe Vert, arxiv.org/abs/1110.0413.
Automatic MiSe System Software Version 1.4803M | admin login