A Factorized Representation of Independence of Causal Influence and Lazy Propagation

Theefficiency of algorithmsfor probabilistic inference in Bayesian networks can be improvedby exploiting independenceof causal influence. Thefactorized representation of independenceof causal influence offers a factorized decompositionof certain independenceof causal influence models. Wedescribe howlazy propagation - a junction tree basedinference algorithm easily can be extendedto take advantageof the decomposition offered by the factorized representation. We introduce two extensions to the factorized representation easing the knowledgeacquisition task and reducing the space complexity of the representation exponentially in the state space size of the effect variable of an independenceof causal influence model.Finally, wedescribe howthe factorized representation can be used to solve tasks such as calculating the maximum a posteriori hypothesis, the maximumexpectedutility: and the most probable configuration.

• Type: Journal
• Year: 2000
• Where: IJUFKS
• Authors: Anders L. Madsen, Bruce D'Ambrosio
• Keywords: Causal Influence Models, Factorized Representation, IJUFKS 2000, Independenceof Causal Influence