What's on in Computing Science?
Date: Tuesday, 20 January, 2009
Sir Alwyn Williams Building, 422 Seminar Room
Learning the Kernel: Theory and Applications
Yiming Ying (University of Bristol)
We will talk about learning the kernel problem in supervised learning. The principal motivations range from the classical model selection problem on tuning the hyper-parameter in SVMs and data integration problems to enhance biological inference in bioinformatics. For the general kernel learning problem, we approach it from regularization theory and statistical generalization analysis. Previous kernel methods on data integration focus on maximizing the margin in SVMs which enjoys the essential idea of block sparse $\ell^1$ -regularization. In contrast, we consider a novel approach based on Kullback-Leibler (KL) divergence to integrate heterogeneous data features. We formulate it as a di?erence of convex (DC) problem which can be solved by a sequence of convex semi-in?nite linear programs. The e?ectiveness of the proposed algorithm is evaluated on a benchmark dataset for protein fold recognition and a yeast protein function prediction problem.
Contact: Dr Rónán Daly (email@example.com)
Add to my calendar