Answering queries from context-sensitive probabilistic knowledge bases

Ngo, Liem Viet; Haddawy, Peter

doi:10.1016/s0304-3975(96)00128-4

Cited by 117 publications

(95 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…If all these assertions are consistent and believed with certainty, a logical calculus can be used to combine them. If not, a probabilistic calculus may be used (e.g., knowledge-based model construction [25]). However, our focus here is not on deriving beliefs for new statements given an initial set of statements.…”

Section: Modelmentioning

confidence: 99%

Trust Management for the Semantic Web

Richardson

Agrawal

Domingos

2003

Lecture Notes in Computer Science

702

385

View full text Add to dashboard Cite

Abstract. Though research on the Semantic Web has progressed at a steady pace, its promise has yet to be realized. One major difficulty is that, by its very nature, the Semantic Web is a large, uncensored system to which anyone may contribute. This raises the question of how much credence to give each source. We cannot expect each user to know the trustworthiness of each source, nor would we want to assign top-down or global credibility values due to the subjective nature of trust. We tackle this problem by employing a web of trust, in which each user maintains trusts in a small number of other users. We then compose these trusts into trust values for all other users. The result of our computation is not an agglomerate "trustworthiness" of each user. Instead, each user receives a personalized set of trusts, which may vary widely from person to person. We define properties for combination functions which merge such trusts, and define a class of functions for which merging may be done locally while maintaining these properties. We give examples of specific functions and apply them to data from Epinions and our BibServ bibliography server. Experiments confirm that the methods are robust to noise, and do not put unreasonable expectations on users. We hope that these methods will help move the Semantic Web closer to fulfilling its promise.

show abstract

Section: Modelmentioning

confidence: 99%

Trust Management for the Semantic Web

Richardson

Agrawal

Domingos

2003

Lecture Notes in Computer Science

702

385

View full text Add to dashboard Cite

show abstract

“…Especially for large pre-existing knowledge bases, such automation facilitates building models of information analysis and, through the use of pre-existing logical knowledge, makes our specification more complete and objective. Because we are using Bayesian networks for probabilistic reasoning, the actual inference would be done only on the subgraph consisting of the query and evidence nodes and their ancestors [71], even if the model is very large. In particular, the complementary knowledge is always added as child nodes in resulting Bayesian networks.…”

Section: Resultsmentioning

confidence: 99%

On the combination of logical and probabilistic models for information analysis

et al. 2011

View full text Add to dashboard Cite

Formal logical tools are able to provide some amount of reasoning support for information analysis, but are unable to represent uncertainty. Bayesian network tools represent probabilistic and causal information, but in the worst case scale as poorly as some formal logical systems and require specialized expertise to use effectively. We describe a framework for systems that incorporate the advantages of both Bayesian and logical systems. We define a formalism for the conversion of automatically generated natural deduction proof trees into Bayesian networks. We then demonstrate that the merging of such networks with domain-specific causal models forms a consistent Bayesian network with correct values for the formulas derived in the proof. In particular, we show that hard evidential updates in which the premises of a proof are found to be true force the conclusions of the proof to be true with probability one, regardless of any dependencies and prior probability values assumed for the causal model. We provide several examples that demonstrate the generality of the natural deduction system by using inference schemes not supportable directly in Horn clause logic. We compare our approach to other ones, including some that use non-standard logics.

show abstract

“…To illustrate the connection between graphical and logical formalisms for directed graphical models, consider the conditional probability parameters of a Bayes net (BN), which are of the form P (child_value|parent_values) = p. These can be translated into Horn clauses of the form child_value ← parent_values; p where the head is the assignment of a value to a child node, and the body specifies an assignment of values to the parent nodes. Ngo and Haddawy refer to such clauses as p-sentences (Ngo and Haddawy 1997). In this view, the qualitative component of a Bayes net is a set of Horn clauses, and the quantitative component is a set of conditional probabilities, one for the head of each clause.…”

Section: Introduction: Moralization For Relational Datamentioning

confidence: 99%

Learning compact Markov logic networks with decision trees

Khosravi

Gao

2012

Mach Learn

View full text Add to dashboard Cite

Statistical-relational learning combines logical syntax with probabilistic methods. Markov Logic Networks (MLNs) are a prominent model class that generalizes both firstorder logic and undirected graphical models (Markov networks). The qualitative component of an MLN is a set of clauses and the quantitative component is a set of clause weights. Generative MLNs model the joint distribution of relationships and attributes. A state-ofthe-art structure learning method is the moralization approach: learn a set of directed Horn clauses, then convert them to conjunctions to obtain MLN clauses. The directed clauses are learned using Bayes net methods. The moralization approach takes advantage of the high-quality inference algorithms for MLNs and their ability to handle cyclic dependencies. A weakness of moralization is that it leads to an unnecessarily large number of clauses. In this paper we show that using decision trees to represent conditional probabilities in the Bayes net is an effective remedy that leads to much more compact MLN structures. In experiments on benchmark datasets, the decision trees reduce the number of clauses in the moralized MLN by a factor of 5-25, depending on the dataset. The accuracy of predictions is competitive with the models obtained by standard moralization, and in many cases superior.

show abstract

Answering queries from context-sensitive probabilistic knowledge bases

Cited by 117 publications

References 11 publications

Trust Management for the Semantic Web

Trust Management for the Semantic Web

On the combination of logical and probabilistic models for information analysis

Learning compact Markov logic networks with decision trees

Contact Info

Product

Resources

About