An intuitive approach to utilizing unlabeled data in kernel-based classification algorithms is to simply treat unknown labels as additional optimization variables. For marginbased...
Vikas Sindhwani, S. Sathiya Keerthi, Olivier Chape...
Abstract— The role of gradient estimation in global optimization is investigated. The concept of a regional gradient is introduced as a tool for analyzing and comparing different...
Abstract. We study the implementation on grid systems of an efficient algorithm for demanding global optimization problems. Specifically, we consider problems arising in the geneti...
In this paper we propose a Particle Filter-based propagation approach for the segmentation of vascular structures in 3D volumes. Because of pathologies and inhomogeneities, many de...
We present an approach for merging message streams from producers distributed over a network, using a deterministic algorithm that is independent of any nondeterminism of the syst...