2802 search results - page 126 / 561 » Online Learning with Kernels |
143
click to vote
Presentation
Presentation of the exponential families, of the mixtures of such distributions and how to learn it. We then present algorithms to simplify mixture model, using Kullback-Leibler di...
ICML
2008
IEEE
15 years 10 months ago
2008
IEEE
The problem of identifying the minimal gene set required to sustain life is of crucial importance in understanding cellular mechanisms and designing therapeutic drugs. This work d...
ECML
2004
Springer
15 years 3 months ago
2004
Springer
One approach to improve the accuracy of classifications based on generative models is to combine them with successful discriminative algorithms. Fisher kernels were developed to c...
COLT
2008
Springer
14 years 11 months ago
2008
Springer
A kernel over the Boolean domain is said to be reflection-invariant, if its value does not change when we flip the same bit in both arguments. (Many popular kernels have this prop...
JMLR
2008
14 years 10 months ago
2008
We propose a highly efficient framework for penalized likelihood kernel methods applied to multiclass models with a large, structured set of classes. As opposed to many previous a...