Tuesday, September 26

Invited Lecture: A nonlinear approach to sparse optimization

Time: 8:30 - 9:30
Room: L1, Building L
Chair: Luca Zanni, Universita di Modena e Reggio Emilia

Sparse optimization problems and optimization problems with cardinality constraints have many applications such as portfolio optimization, subset selection, compressed sensing or learning. In the past solution approaches have mostly focused on convex substitutes of the respective problems, e.g. using the l1-norm to induce sparsity. However, recently nonconvex formulations have gained popularity both in learning and in the handling of sparse or cardinality constrained problems. In this talk, we give an overview over these new nonconvex reformulations and related solution approaches.

Alexandra Schwartz
Technische Universität Darmstadt