Files in this item

FilesDescriptionFormat

application/pdf

application/pdfLifted First-Order Probabilistic Inference.pdf (839kB)
(no description provided)PDF

Description

Title:Lifted First-Order Probabilistic Inference
Author(s):Braz, Rodrigo de Salvo
Subject(s):artificial intelligence
Abstract:There has been a long standing division in AI between logical symbolic and probabilistic reasoning approaches. While probabilistic models can deal well with inherent uncertainty in many real-world domains, they operate on a mostly propositional level. Logic systems, on the other hand, can deal with much richer representations, especially first-order ones. In the last two decades, many probabilistic algorithms accepting first-order specifications have been proposed, but in the inference stage they still operate mostly on a propositional level, where the rich and useful first-order structure is not explicit anymore. In this thesis we present a framework for lifted inference on first-order models, that is, inference where the main operations occur on a first-order level, without the need to propositionalize the model. We clearly define the semantics of first-order probabilistic models, present an algorithm (FOVE) that performs lifted inference, and show detailed proofs of its correctness. Furthermore, we describe how to solve the Most Probable Explanation problem with a variant of FOVE, and present a new anytime probabilistic inference algorithm, ABVE, meant to generalize the ability of logical systems to gradually process a model and stop as soon as an answer is available.
Issue Date:2007-09
Genre:Technical Report
Type:Text
URI:http://hdl.handle.net/2142/11362
Other Identifier(s):UIUCDCS-R-2007-2869
Rights Information:You are granted permission for the non-commercial reproduction, distribution, display, and performance of this technical report in any format, BUT this permission is only for a period of 45 (forty-five) days from the most recent time that you verified that this technical report is still available from the University of Illinois at Urbana-Champaign Computer Science Department under terms that include this permission. All other rights are reserved by the author(s).
Date Available in IDEALS:2009-04-22


This item appears in the following Collection(s)

Item Statistics