We introduce a novel approach to incorporating domain knowledge into Support Vector Machines to improve their example efficiency. Domain knowledge is used in an Explanation Based ...
This paper describes a novel application of support vector machines and multiscale texture and color invariants to a problem in biological oceanography: the identification of 6 sp...
Background: Support Vector Machines (SVMs) ? using a variety of string kernels ? have been successfully applied to biological sequence classification problems. While SVMs achieve ...
Text categorization algorithms usually represent documents as bags of words and consequently have to deal with huge numbers of features. Most previous studies found that the major...
We show how the regularizer of Transductive Support Vector Machines (TSVM) can be trained by stochastic gradient descent for linear models and multi-layer architectures. The resul...
Michael Karlen, Jason Weston, Ayse Erkan, Ronan Co...