A Tractable Approach to ABox Abduction over Description Logic Ontologies

Authors: Jianfeng Du, Kewen Wang, Yi-Dong Shen

AAAI 2014 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We develop a tractable method (in data complexity) for computing all representative explanations in a consistent ontology. Experimental results demonstrate that the method is efficient and scalable for ontologies with large ABoxes.
Researcher Affiliation Academia Jianfeng Du Guangdong University of Foreign Studies, Guangzhou 510006, China jfdu@gdufs.edu.cn Kewen Wang Griffith University, Brisbane, QLD 4111, Australia k.wang@griffith.edu.au Yi-Dong Shen State Key Laboratory of Computer Science, Institute of Software, Chinese Academy of Sciences, Beijing 100190, China ydshen@ios.ac.cn
Pseudocode No The paper presents theoretical lemmas and theorems, and illustrative examples of the method, but it does not include any structured pseudocode or algorithm blocks.
Open Source Code No The paper mentions that the proposed method was 'implemented in Java' and refers to a 'Prolog-based method' with a footnote link to 'http://dataminingcenter.net/abduction/', but it does not explicitly state that the source code for the proposed methodology is open-source or provide a direct link to its repository.
Open Datasets Yes Seven benchmark ontologies with large ABoxes were used. The first two are Semintec (about financial services) and Vicodi (about European history). The remaining ontologies are LUBMn (n = 1, 5, 10, 50, 100) from the Lehigh University Benchmark (Guo, Pan, and Heflin 2005)
Dataset Splits No The paper mentions generating 'observations' or 'BCQs' for testing but does not specify any training, validation, or test splits of the datasets themselves.
Hardware Specification Yes All experiments were conducted on a laptop with Intel Dual-Core 2.20GHz CPU and 4GB RAM, running Windows 7, where the maximum Java heap size was set to 1GB.
Software Dependencies No The proposed method was implemented in Java, using the Requiem (P erez-Urbina, Motik, and Horrocks 2010) API for query rewriting and the My SQL engine to store and access ABoxes. No specific version numbers for Java, Requiem, or MySQL are provided.
Experiment Setup Yes We set a one-hour time limit to both methods for handling one observation.