Query Abduction for ELH Ontologies
Authors: Mahsa Chitsaz, Zhe Wang, Kewen Wang
AAAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We implemented a prototypical system using the highly optimized Prolog engine XSB1. We evaluated our algorithm over some realistic ontologies with reasonably large data sets such as extended university benchmark ontology LSTW(n) that has approximately 105n ABox assertions. Our experimental results show that the algorithm is capable of handling query abduction problems for LSTW(n) where n is up to 100. |
| Researcher Affiliation | Academia | Mahsa Chitsaz, Zhe Wang, Kewen Wang School of Information and Communication Technology, Griffith University, Australia mahsa.chitsaz@griffithuni.edu.au, {zhe.wang,k.wang}@griffith.edu.au |
| Pseudocode | No | The paper does not include a clearly labeled pseudocode or algorithm block. |
| Open Source Code | No | The paper mentions implementing a prototypical system using XSB Prolog and provides a link to XSB (http://xsb.sourceforge.net/), which is a third-party engine, not the source code for the authors' specific implementation (ABEL) described in the paper. |
| Open Datasets | No | The paper states that it evaluated the system on 'LSTW(n), an extended version of the university ontology LUBM', and mentions using 'a query generator to generate 13 atomic and 27 conjunctive queries'. While LUBM is a known benchmark, the paper does not provide concrete access information (e.g., URL, DOI, specific citation for LSTW(n) or the generated queries) for the exact dataset and queries used in their experiments. |
| Dataset Splits | No | The paper describes the datasets used (LSTW(n)) and their sizes but does not specify any training, validation, or test dataset splits or cross-validation methodology. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., CPU/GPU models, memory) used for running the experiments. |
| Software Dependencies | No | The paper states using 'the highly optimized Prolog engine XSB' but does not provide specific version numbers for XSB or any other software dependencies. |
| Experiment Setup | Yes | We implemented our new procedure of computing QAP solutions using the highly optimized Prolog engine XSB. To compute all minimal solutions to a QAP DK, Q ( a), Σ, , we encode the QAP into Prolog rules, and use the list structure to store the solutions generated during the resolution. We have also employed the database feature of XSB to allow the initial ABox to be stored in a database, which significantly improves the efficiency. The abducibles are all the concept names in the TBox, and the domain consists of all individuals in the ABox. The sizes of ABoxes range from 100 thousands (LSTW(1)) to 10 millions assertions (LSTW(100)), the sizes of domain are between 17 thousands (LSTW(1)) and 1.7 million (LSTW(100)) and the abducible size is 132. Overall, we ran the first experiment for 200 test cases. All times are in seconds, and the time limit is one hour. |