Main content

De-Causalization of NIN-AND Tree Models

Show simple item record

dc.contributor.advisor Xiang, Yang Loker, Dylan 2018-08-29T17:20:12Z 2018-08-29T17:20:12Z 2018 2018-07-30 2018-08-29
dc.description.abstract Joint probability distributions suffer from combinatorial explosion on the number of variables present. Bayesian networks avoid this issue through encoding conditional independence between variables, making use of a graphical structure alongside tabular representations of probabilistic information for each variable. However, these tabular representations still have exponential growth on the number of incoming connections of the variable in the graph. To address this growth, space-efficient local models have been developed. In this thesis, we make use of the non-impeding noisy-AND tree (NAT) model for expressing local probabilistic information due to its simple causal interactions and expressiveness. We develop a novel approach, which we call de causalization of the NAT model, which exploits causal independence present in the NAT model to improve inference efficiency. We demonstrate the exactness of this approach and evaluate inference efficiency using lazy propagation. en_US
dc.language.iso en en_US
dc.rights Attribution-NoDerivs 2.5 Canada *
dc.rights.uri *
dc.subject Bayesian Networks en_US
dc.subject Uncertain Reasoning en_US
dc.subject Knowledge Representation en_US
dc.subject Non-Impeding Noisy-AND Tree en_US
dc.subject Artificial Intelligence en_US
dc.subject De-causalization en_US
dc.subject Trans-causalization en_US
dc.title De-Causalization of NIN-AND Tree Models en_US
dc.type Thesis en_US Computer Science en_US Master of Science en_US School of Computer Science en_US
dc.rights.license All items in the Atrium are protected by copyright with all rights reserved unless otherwise indicated.

Files in this item

Files Size Format View
Loker_Dylan_201808_Msc.pdf 2.561Mb PDF View/Open

This item appears in the following Collection(s)

Show simple item record

Attribution-NoDerivs 2.5 Canada Except where otherwise noted, this item's license is described as Attribution-NoDerivs 2.5 Canada