An ideal visibility algorithm should a) quickly reject most of the hidden geometry in a model and b) exploit the spatial and perhaps temporal coherence of the images being generat...
In the context of shape and image modeling by manifold learning, we focus on the problem of denoising. A set of shapes or images being known through given samples, we capture its s...
Abstract. In this paper, I present the class of linear context free languages (LCFLs) with a class of non-deterministic one-way two-head (read only) automata, called non-determinis...
Conventional methods for state space exploration are limited to the analysis of small systems because they suffer from excessive memory and computational requirements. We have dev...
William J. Knottenbelt, Peter G. Harrison, Mark Me...
In this paper we present a method for learning classspecific
features for recognition. Recently a greedy layerwise
procedure was proposed to initialize weights of deep
belief ne...
Mohammad Norouzi (Simon Fraser University), Mani R...