Extracting Bounded-Level Modules from Deductive RDF Triplestores

Authors: Marie-Christine Rousset, Federico Ulliana

AAAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results show that the resulting framework is effective in extracting expressive modules from RDF datasets with formal guarantees, whilst controlling their succinctness.
Researcher Affiliation Academia Marie-Christine Rousset Univ. Grenoble Alpes CNRS, LIG, 38000, Grenoble, France IUF (Institut Universitaire de France) marie-christine.rousset@imag.fr Federico Ulliana Universit e de Montpellier II CNRS, LIRMM, 34000, Montpellier, France Grahp IK Team, INRIA Sophia-Antipolis ulliana@lirmm.fr
Pseudocode Yes Algorithm 1: MRE(NTo Unfold, RTo Apply, Σ)
Open Source Code No The paper states 'Our approach has been implemented on top of an RDF engine and experimentally tested. Proofs and experiment details are reported in (Rousset and Ulliana 2014).' This refers to a technical report for details, not a public code repository for their implementation.
Open Datasets Yes We considered the following three Semantic Web datasets. My CF 0.5M triples 11 domain-specific rules GO 1M triples 15 domain-specific rules Yago2 14M triples 6 RDFS rules. ... For the GO ontology (www.geneontology.org)... My CF ontology (Palombi et al. 2014)... YAGO (Suchanek, Kasneci, and Weikum 2007)
Dataset Splits No The paper mentions sampling classes and properties and testing resources (e.g., 'We tested 100 Yago resources for each group'). However, it does not specify explicit train/validation/test dataset splits or percentages for these datasets, nor does it mention a cross-validation setup.
Hardware Specification No The paper mentions software used ('Jena 2.11.2 TDB') but does not specify any hardware details like GPU/CPU models, memory, or cloud computing instances used for experiments.
Software Dependencies Yes We implemented bounded-level module extraction on top of Jena 2.11.2 TDB
Experiment Setup Yes We considered 2500 My CF ontology classes combined with 20 subsets of its properties, of size 1-4. For the GO ontology... we sampled 350 classes and 12 property sets (size 1-4). ... We tested 100 Yago resources for each group. Finally, we made k ranging over {1, 2, 3, 5, 10}.