: One important problem which may arise in designing a deployment strategy for a wireless sensor network is how to deploy a specific number of sensor nodes throughout an unknown ne...
Graphical models are usually learned without regard to the cost of doing inference with them. As a result, even if a good model is learned, it may perform poorly at prediction, be...
This paper describes a novel view-based learning algorithm for 3D object recognition from 2D images using a network of linear units. The SNoW learning architecture is a sparse netw...
Background: Probability based statistical learning methods such as mutual information and Bayesian networks have emerged as a major category of tools for reverse engineering mecha...
Taking a global analogy with the structure of perceptual biological systems, we present a system composed of two layers of real-valued sigmoidal neurons. The primary layer receives...