TY - GEN
T1 - ArThUR: A Tool for Markov Logic Network
AU - Bodart, Axel
AU - Evrard, Keyvin
AU - Ortiz Vega, James Jerson
AU - Schobbens, Pierre Yves
PY - 2014
Y1 - 2014
N2 - Logical approaches-and ontologies in particular-offer a welladapted framework for representing knowledge present on the Semantic Web (SW). These ontologies are formulated in Web Ontology Language (OWL2), which are based on expressive Description Logics (DL). DL are a subset of First-Order Logic (FOL) that provides decidable reasoning. Based on DL, it is possible to rely on inference mechanisms to obtain new knowledge from axioms, rules and facts specified in the ontologies. However, these classical inference mechanisms do not deal with : uncertainty probabilities. Several works recently targeted those issues (i.e. Pronto, PR-OWL, BayesOWL, etc.), but none of them combines OWL2 with Markov Logic Networks (MLN) formalism. Several open source software packages for MLN are available (e.g. Alchemy, Tuffy, RockIt, etc.). In this paper, we present ArThUR, a Java framework for reasoning with probabilistic information in the SW. ArThUR incorporate three open source software packages for MLN, which is able to reason with uncertainty information, showing that it can be used in several real-world domains. We also show several experiments of our tool with different ontologies.
AB - Logical approaches-and ontologies in particular-offer a welladapted framework for representing knowledge present on the Semantic Web (SW). These ontologies are formulated in Web Ontology Language (OWL2), which are based on expressive Description Logics (DL). DL are a subset of First-Order Logic (FOL) that provides decidable reasoning. Based on DL, it is possible to rely on inference mechanisms to obtain new knowledge from axioms, rules and facts specified in the ontologies. However, these classical inference mechanisms do not deal with : uncertainty probabilities. Several works recently targeted those issues (i.e. Pronto, PR-OWL, BayesOWL, etc.), but none of them combines OWL2 with Markov Logic Networks (MLN) formalism. Several open source software packages for MLN are available (e.g. Alchemy, Tuffy, RockIt, etc.). In this paper, we present ArThUR, a Java framework for reasoning with probabilistic information in the SW. ArThUR incorporate three open source software packages for MLN, which is able to reason with uncertainty information, showing that it can be used in several real-world domains. We also show several experiments of our tool with different ontologies.
UR - http://www.scopus.com/inward/record.url?scp=84910107369&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:84910107369
SN - 9783662455494
VL - 8842
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 319
EP - 328
BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PB - Springer Verlag
T2 - International Workshops: OTM Academy,OTM Industry Case Studies Program, C and TC, EI2N, INBAST, ISDE, META4eS, MSC, and OnToContent 2014
Y2 - 27 October 2014 through 31 October 2014
ER -