Goal-Driven Query Answering for Existential Rules With Equality

Authors: Michael Benedikt, Boris Motik, Efthymia Tsamoura

AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental 8 Empirical Evaluation We evaluated our technique using CHASEBENCH (Benedikt et al. 2017), a recent benchmark offering a mix of scenarios that simulate data exchange and ontology reasoning applications. We selected the scenarios summarized in Table 1, each comprising a set of existential rules, a base instance, and several queries.
Researcher Affiliation Academia Michael Benedikt, Boris Motik University of Oxford Efthymia Tsamoura Alan Turing Institute & University of Oxford
Pseudocode Yes Algorithm 1 Compute the answers to query Q over a finite set of existential rules Σ and a base instance B... Algorithm 2 relevance(P, B)... Algorithm 3 magic(P)
Open Source Code Yes Our system and the test data are available online.2... 2http://github.com/tsamoura/chase Goal
Open Datasets Yes We evaluated our technique using CHASEBENCH (Benedikt et al. 2017), a recent benchmark... LUBM-100 and LUBM-1K are derived from the well-known Semantic Web LUBM (Guo, Pan, and Heflin 2011) benchmark...
Dataset Splits No The paper describes the datasets used but does not provide specific details on how they were split into training, validation, or testing sets.
Hardware Specification No To compute the chase of the final program (line 7 of Algorithm 1), we used the RAM-based RDFox system written in C++. ... We implemented our technique in Java on top of the CHASEBENCH (Benedikt et al. 2017) library. ... on our test machine" The paper mentions the system used (RDFox, Java) but does not specify any hardware details like CPU, GPU, or RAM specifications.
Software Dependencies No To compute the chase of the final program (line 7 of Algorithm 1), we used the RAM-based RDFox system written in C++. ... We implemented our technique in Java on top of the CHASEBENCH (Benedikt et al. 2017) library." The paper mentions programming languages and systems used but does not provide specific version numbers for them or any libraries.
Experiment Setup Yes In each test run, we computed the program P6 from Algorithm 1 (skipping the relevance analysis and/or magic sets, as required for the test type), computed chase(P6, B), and output the certain answers of Q as shown in line 9 of Algorithm 1. We recorded the wall-clock time of each run (without the loading times) and the number of facts derived by the chase; the latter provides an implementation-independent measure of the work needed to answer a query. ... We used just one thread while computing the chase.