SiN3: Scalable inferencing with SPARQL CONSTRUCT queries

Dörthe ArndtDörthe Arndt, William Van WoenselWilliam Van Woensel, Dominik TomaszukDominik Tomaszuk

Dörthe Arndt, William Van Woensel, Dominik Tomaszuk
SiN3: Scalable inferencing with SPARQL CONSTRUCT queries
In Irini Fundulaki, Kouji Kozaki, Daniel Garijo, Jose Manuel Gomez-Perez, eds., ISWC-Posters-Demos-Industry 2023, volume 3632 of CEUR Workshop Proceedings, November 2023. ceur-ws.org

Details
Bibtex

KurzfassungAbstract
SPARQL is a widely used and well-established standard in the Semantic Web to efficiently query explicitly stated RDF data. However, the authors (and others before us) suggest that this ignores the full potential of the language: when used recursively, CONSTRUCT queries can serve as fully-fledged rules allowing for reasoning over the data. For this purpose, we present SiN3, a tool that translates SPARQL queries to Notation3 (N3) rules; an N3-compliant reasoner (such as eye) can then recursively generate inferences from RDF data. We provide a demo that uses CONSTRUCT queries to infer the need for Zika screening, based on the Turtle HL7 FHIR notation. We show that a reasoning-based approach, compared to standard SPARQL, makes such use cases more efficient, as rule-based reasoning is optimized for recursively generating inferences; and easier to implement, as only one execution run is necessary to apply interdependent queries, and supporting queries can be used to modularize the code. We compare the performance of SiN3 with SpinRDF, which supports reasoning over the SPIN serialization of SPARQL but is an order of magnitude slower on our test data. SiN3 thus enables efficient reasoning using SPARQL CONSTRUCT queries, opening the door for scalable Semantic Web reasoning using a well-known semantic formalism.
Weitere Informationen unter:Further Information: Link
Projekt:Project: KIMEDS, MEDGE
Forschungsgruppe:Research Group: Computational LogicComputational Logic

@inproceedings{AWT2023,
  author    = {D{\"{o}}rthe Arndt and William Van Woensel and Dominik Tomaszuk},
  title     = {SiN3: Scalable inferencing with {SPARQL} {CONSTRUCT} queries},
  editor    = {Irini Fundulaki and Kouji Kozaki and Daniel Garijo and Jose
               Manuel Gomez-Perez},
  booktitle = {ISWC-Posters-Demos-Industry 2023},
  series    = {CEUR Workshop Proceedings},
  volume    = {3632},
  publisher = {ceur-ws.org},
  year      = {2023},
  month     = {November}
}