Mapping and parallel implementation of Bayesian belief networks

Traditionally, probabilistic reasoning has been plagued by its computational intractability in real sitiiations. Complete joint probabilistic specificat.ions are not possible in most situations and very simplistic probabilistic models which are compiitationally inexpensive do not capture the complexities of a real life problem. With the advent of Bayesian networks, we have seen a renewed interest in probabilistic reasoning. Bayesian networks provide us with an unique way of looking at a probability distribution in terms of a directed graph. Each node in the Bayesian network represents a random variable and the links denote direct dependencies between random variables. The links are quantified with the conditional probabilities. Sot only does the graph structure allow a better visualization of the inherent dependencies among the random variables but the network translates into a distributed computational structure. Bayesian networks are presently being used in computer vision, artificial intelligence, medicine, CAM, troubleshooting and other applications wherein decisions are conditionally dependent on many controlling factors.