Title :
Dynamic Graphical Models
Author_Institution :
An associate professor in the Department of Electrical Engineering at the University of Washington, Seattle, and is an adjunct professor in the Department of Computer Science and Engineering and the Department of Linguistics.
Abstract :
A graphical model consists of I a graph G = (V,E) and a V set of properties that determine a family of V probability distributions. There are many different types of graphs and properties, each determining a family. It is common to be able to develop algorithms that work for all members of the family by considering only a graph and its properties. Thus, solving difficult problems (such as deriving an approximation to an NP-complete optimization problem) might become worthwhile only because a solution can be applied many times for different problem instances.
Keywords :
graph theory; statistical distributions; NP-complete optimization problem; dynamic graphical models; graph theory; probability distribution; Graphical models; Hidden Markov models; Inference algorithms; Junctions; Markov processes; Signal processing algorithms;
Journal_Title :
Signal Processing Magazine, IEEE
DOI :
10.1109/MSP.2010.938078
