Optimization issues in training support vector machines
(invited lecture for ALT 2005)
Author: Chih-Jen Lin
Affiliation: Department of Computer Science and Information Engineering,
National Taiwan University, Taipei, Taiwan
Abstract.
Optimization plays an important role in many machine learning
methods. In this talk we will introduce and discuss how optimization
techniques are applied to support vector machines (SVM), a promising
classification method. Training SVM involves solving a huge quadratic
programming problem (QP). There are many algorithmic issues related to
this special type of QP. We will particularly focus on the
decomposition methods for solving large SVM dual optimization
problems. A brief introduction on the practical use of SVM may also be
included.
©Copyright 2005 Author
|