Spectral Projected Gradient Descent for Efficient and Large Scale Generalized Multiple Kernel Learning

Abstract

We address the problem of learning the kernel in a Support Vector Machine framework from training data. Learning diverse types of kernels has proved effective in different application areas. For instance, learning the kernel to be a sparse linear combination of given base kernels has proved useful in computer vision while more general kernel… (More)

13 Figures and Tables

Topics

  • Presentations referencing similar topics