“Spectral Methods for Regularization in Learning Theory”
Speaker: Alessandro Verri , Universita' di Genova
Date: February 1 2006
Time: 4:00PM to 5:25PM
Location: 46-3189
Host: Prof. Tomaso Poggio, M.I.T., McGovern Institute, BCS and CSAIL
Contact: Mary Pat Fitzgerald, 253-0551, marypat@csail.mit.edu
Abstract:
In this talk we show that a large class of regularization methods designed for solving ill-posed inverse problems gives rise to consistent learning algorithms. The intuition behind our approach is that by looking at regularization from a filter function perspective, filtering out undesired components of the target function ensures stability with respect to the random sampling, thereby inducing good generalization properties. We present a formal derivation of the methods under study by recalling that learning can be written as the inversion of a linear embedding equation given a stochastic discretization. Consistency, as well as finite sample bounds, is derived for both regression and classification.
See other events that are part of Brains & Machines Seminar Series 2006
See other events happening in February 2006
