no code implementations • NeurIPS 2008 • Ted Sandler, John Blitzer, Partha P. Talukdar, Lyle H. Ungar
Here we present a framework for regularized learning in settings where one has prior knowledge about which features are expected to have similar and dissimilar weights.