Traditional methods of dealing with variability in simulation input data are mainly stochastic. This is most often the best method to use if the factors affecting the variation or...
Background: Most predictive methods currently available for the identification of protein secretion mechanisms have focused on classically secreted proteins. In fact, only two met...
Daniel Restrepo-Montoya, Camilo Pino, Luis F. Ni&n...
An important and unsolved problem today is the automatic quantification of the quality of video flows transmitted over packet networks. In particular, the ability to perform this ...
We present a new approach to bounding the true error rate of a continuous valued classifier based upon PAC-Bayes bounds. The method first constructs a distribution over classifier...
Abstract. Recently, a new method intended to realize conformal mappings has been published. Called Locally Linear Embedding (LLE), this method can map high-dimensional data lying o...