Conditioning Graphs: Practical Structures for Inference in Bayesian Networks.

Kevin Grant Department of Computer Scinece University of Saskatchewan

DEPARTMENT SEMINAR DATE: Friday, February 3rd, 2006 TIME: 3:30 pm PLACE: Thorvaldson 271 *** Everyone is welcome ***

Abstract

Programmers employing inference in Bayesian networks typically rely on the inclusion of the model as well as an inference engine into their application. Sophisticated inference engines require non-trivial amounts of space and are also difficult to implement. This limits their use in some applications that would otherwise benefit from probabilistic inference. In this talk, we present a system that minimizes the space requirement of the model. The inference engine is sufficiently simple as to avoid space-limitation and be easily implemented in almost any environment. We show a fast, compact indexing structure that is linear in the size of the network. The additional space required to compute over the model is linear in the number of variables in the network.

About the speaker

Kevin Grant is a PhD student in the Dept. of Computer Science at the University of Saskatchewan. He is supervised by Prof. Michael Horsch. Kevin received his B.Sc. from the University of Saskatchewan in 2001.

Kevin's research interests are in Artificial Intelligence, particularly in Bayesian networks. He is interested in Bayesian computation under constrained resources and architectures. His current work is in conditioning graphs: a model of Bayesian inference that is flexible in its resource requirements, and abstracts out any high-level inference details, making it accessible and implementable to general users (someone unfamiliar with the mechanics of belief networks).

[an error occurred while processing this directive]