Large-scale information processing systems are able to extract massive collections of interrelated facts, but unfortunately transforming these candidate facts into useful knowledge is a formidable challenge. In this paper, the authors show how uncertain extractions about entities and their relations can be transformed into a knowledge graph. They demonstrate the power of their method on a synthetic Linked Data corpus derived from the MusicBrainz music community and a real-world set of extractions from the NELL (Never-Ending Language Learner) project. NOTE: Also available as a PDF: http://videolectures.net/site/normal_dl/tag=817827/iswc2013_pujara_graph_identification_01.pdf
URL: http://videolectures.net/iswc2013_pujara_graph_identification/
Keywords: Probabilistic SoftLogic (PSL), Ontology, Named entity extraction, Knowledge graph
Author: Pujara, Jay
Publisher: videolectures.net
Date created: 2013-11-28 05:00:00.000
Language: http://id.loc.gov/vocabulary/iso639-2/eng
Time required: P15M
Educational use: instruction
Educational audience: generalPublic
Interactivity type: expositive
- Cleans a dataset by finding and correcting errors, removing duplicates and unwanted data.
- Knows that the word "ontology" is ambiguous, referring to any RDF vocabulary, but more typically a set of OWL classes and properties designed to support inferencing in a specific domain.
- Uses available resources for named entity recognition, extraction, and reconciliation.