FoCM 2014 conference

Workshop C3 - Learning Theory

December 20, 17:30 ~ 18:00 - Room B23

Iterative Regularization for Computational Learning

Lorenzo Rosasco

Universita' di Genova , Italy   -

Iterative regularization approaches to ill-posed inverse problems are known to provide a viable alternative to Tikhonov regularization, especially in large scale problems. Supervised learning can be seen as an inverse problem under a suitable stochastic data model. In this context, iterative regularization is particularly suited since statistical and computational aspects are tackled at once, a key property when dealing with large data-sets. In this talk we will discuss old and new results on learning with iterative regularization and connect them with recent results in online learning.

View abstract PDF