From statistical knowledge bases to degrees of belief

Bacchus, Fahiem; Grove, Adam J.; Halpern, Joseph Y.; Koller, Daphne

doi:10.1016/s0004-3702(96)00003-3

Cited by 159 publications

(170 citation statements)

References 36 publications

(80 reference statements)

Supporting

Mentioning

168

Contrasting

Unclassified

Order By: Relevance

“…Bacchus (1990), Halpern (1990) and coworkers (e.g., Bacchus et al, 1996) studied the problem in detail from a theoretical standpoint. They made a distinction between statistical statements (e.g., "65% of the students in our department are undergraduate") and statements about possible worlds (e.g., "The probability that Anna is an undergraduate is 65%"), and provided methods for computing the latter from the former.…”

Section: Early Workmentioning

confidence: 99%

Markov logic networks

2006

View full text Add to dashboard Cite

We propose a simple approach to combining first-order logic and probabilistic graphical models in a single representation. A Markov logic network (MLN) is a firstorder knowledge base with a weight attached to each formula (or clause). Together with a set of constants representing objects in the domain, it specifies a ground Markov network containing one feature for each possible grounding of a first-order formula in the KB, with the corresponding weight. Inference in MLNs is performed by MCMC over the minimal subset of the ground network required for answering the query. Weights are efficiently learned from relational databases by iteratively optimizing a pseudo-likelihood measure. Optionally, additional clauses are learned using inductive logic programming techniques. Experiments with a real-world database and knowledge base in a university domain illustrate the promise of this approach.

show abstract

Section: Early Workmentioning

confidence: 99%

Markov logic networks

2006

View full text Add to dashboard Cite

show abstract

“…Interestingly, the weighted-worlds approach to degrees of belief suggested in this study provides a natural generalization of the random-worlds approach advocated for the particular case of unweighted knowledge bases (Grove et al 1994;Bacchus et al 1996;Halpern 2003). Indeed, if KB is reduced to a set of hard constraints, then KB corresponds to the number of models of KB, and hence, in this case the degree of belief in Q given KB is just the probability of choosing an interpretation at random that satisfies Q out of all the interpretations that satisfy KB.…”

Section: Fig 2 the Indirect Eg-l2r Algorithmmentioning

confidence: 95%

“…The main assumption behind this paradigm is that, in general, the description is acquired independently of the queries that will be posed. To provide compact and reliable descriptions of complex probability measures, various representation formalisms have been proposed in the literature; some of them extend first-order logical representation languages to probabilistic inference (Poole 1993;Bacchus et al 1996;Muggleton 1996;Ngo and Haddawy 1997;Costa et al 2003;Kersting 2006), while others advocate a dual approach by extending graphical representation languages to relational inference (Jaeger 1997;Friedman et al 1999;Pfeffer 2000;Taskar et al 2002;Richardson and Domingos 2006). For example, if the domain is represented by a knowledge base in first-order logic, the probability measure is induced by assigning equal likelihood to all interpretations that satisfy the knowledge base; the degree of belief of any query is simply the fraction of those interpretations which are consistent with the query (Bacchus et al 1996;Halpern 2003).…”

Section: Introductionmentioning

confidence: 99%

Learning to assign degrees of belief in relational domains

Koriche

2008

Mach Learn

View full text Add to dashboard Cite

A recurrent problem in the development of reasoning agents is how to assign degrees of beliefs to uncertain events in a complex environment. The standard knowledge representation framework imposes a sharp separation between learning and reasoning; the agent starts by acquiring a "model" of its environment, represented into an expressive language, and then uses this model to quantify the likelihood of various queries. Yet, even for simple queries, the problem of evaluating probabilities from a general purpose representation is computationally prohibitive. In contrast, this study embarks on the learning to reason (L2R) framework that aims at eliciting degrees of belief in an inductive manner. The agent is viewed as an anytime reasoner that iteratively improves its performance in light of the knowledge induced from its mistakes. Indeed, by coupling exponentiated gradient strategies in learning and weighted model counting techniques in reasoning, the L2R framework is shown to provide efficient solutions to relational probabilistic reasoning problems that are provably intractable in the classical paradigm.

show abstract

“…In particular, it shows a similar behavior as reference-class reasoning in a number of uncontroversial examples. 1 But it also avoids many drawbacks of reference-class reasoning (which are pointed out in [5,87]): differently from reference-class reasoning, probabilistic lexicographic entailment can handle complex scenarios and even purely probabilistic subjective knowledge as input, and probabilistic lexicographic entailment draws conclusions in a global way from all the available knowledge as a whole. Furthermore, probabilistic lexicographic entailment also has very nice nonmonotonic properties, which are essentially inherited from Lehmann's lexicographic entailment [72].…”

Section: Probabilistic Uncertainty and Description Logicsmentioning

confidence: 99%

“…The fuzzy datatype predicate Young may be defined as Young(x) = ls(x; 10, 30). Then, YoungPerson = Person ∃age.Young (5) denotes young persons.…”

Section: Fuzzy Datatype Theoriesmentioning

confidence: 99%

Managing uncertainty and vagueness in description logics for the Semantic Web

Lukasiewicz

Straccia

2008

Journal of Web Semantics

328

101

View full text Add to dashboard Cite

a b s t r a c tOntologies play a crucial role in the development of the Semantic Web as a means for defining shared terms in web resources. They are formulated in web ontology languages, which are based on expressive description logics. Significant research efforts in the semantic web community are recently directed towards representing and reasoning with uncertainty and vagueness in ontologies for the Semantic Web. In this paper, we give an overview of approaches in this context to managing probabilistic uncertainty, possibilistic uncertainty, and vagueness in expressive description logics for the Semantic Web.

show abstract

From statistical knowledge bases to degrees of belief

Cited by 159 publications

References 36 publications

Markov logic networks

Markov logic networks

Learning to assign degrees of belief in relational domains

Managing uncertainty and vagueness in description logics for the Semantic Web

Contact Info

Product

Resources

About