Lecture Prof. Daniela Calvetti

 

On Tuesday 22nd May at 2.30 pm in the Professors Room I Level of the DMA, Prof. Daniela Calvetti, Case Western Reserve University, will hold a seminar entitled:

"Promoting sparsity with Krylov iterative solvers and hierarchical Bayesian models: the L2 magic."

All interested students are invited to participate.

A brief summary follows.

 

Hierarchical Bayesian models with suitable choices of hyperpriors, combined with Krylov subspace iterative solvers can be very effective at promoting sparsity in the solution of inverse problems while retaining the $\ell_2$ computational efficiency. The use of Krylov iterative solvers makes these methods particular well suited for large scale problems.  In this talk we present some results about the quadratic convergence of these methods for a family of hyperpriors and illustrate their performance with computed examples.