Currently, in one of my classes,  Reasoning with Partial Beliefs, I am learning about Bayesian Networks.  What are they? Probabilistic networks in the form of Directed Acylicy Graphs that can be used to model belief networks.  From these representations one can make various statements about the probability of the various events composing the system given evidence.

I find them incredibly fascinating because they offer powerful ways to model and make statements about various systems.  The applications range from medical diagnosis, to problems involving Genetic linkage and the reliablity of circuts.

The techniques are relatively new in that there was some skepticism about these techniques maybe 10 or 15 years ago, but this no longer the case. 

One can view Bayesian Networks as separate from Computer Science but the linkage is that in most cases computer programs process the Bayesian Networks and give us the answers from the model we input.  The graph representation of these networks also easily lends itself to computer algorithms. 

In a follow up to this article I will provide a simple example of a Bayesian Network and show how it can be used to model something.  Stay tuned.