This article describes an application of the partially observable Markov (POM) model to the analysis of a large scale commercial web search log. Mathematically, POM is a variant o...
In this paper we describe a statistical method for the integration of an unlimited number of cues within a deformable model framework. We treat each cue as a random variable, each...
Siome Goldenstein, Christian Vogler, Dimitris N. M...
Graphical models have become the basic framework for topic based probabilistic modeling. Especially models with latent variables have proved to be effective in capturing hidden str...
Owing to the stochastic nature of discrete processes such as photon counts in imaging, real-world data measurements often exhibit heteroscedastic behavior. In particular, time ser...
This work describes a forward-looking approach for the solution of dynamic (time-changing) problems using evolutionary algorithms. The main idea of the proposed method is to combi...