Alessandro Verri
M.I.T. and Universita' di Genova
In this talk we report some experimental work in which Support Vector Machines (SVM) are used for addressing classification problems in high dimensional spaces. We consider two different applications. In the first, we describe preliminary results on a cancer classification problem based on gene expression monitoring using DNA microarrays. In the second, we discuss trainable systems for object detection and recognition. In both cases we highlight the motivation and the advantages underlying the use of SVMs in high dimensional spaces and illustrate the main open problems.