Tracking Logical Difference in Large-Scale Ontologies: A Forgetting-Based Approach
Authors: Yizheng Zhao, Ghadah Alghamdi, Renate A. Schmidt, Hao Feng, Giorgos Stoilos, Damir Juric, Mohammad Khodadadi3116-3124
AAAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Our evaluation shows that the method can achieve considerably better success rates (>90%) and provides a feasible approach to computing the logical difference in large-scale ontologies, as a case study on different versions of SNOMED CT and NCIt ontologies shows. ... The prototype was evaluated for three settings: forgetting 10% (199), 30% (597) and 50% (995) of concept and role names in the signature of each ontology. We compared the obtained results with those computed by LETHE and FAME 1.0. ... The results obtained from the experiments are shown in Table 1. |
| Researcher Affiliation | Collaboration | Yizheng Zhao The University of Manchester, UK Ghadah Alghamdi The University of Manchester, UK Renate A. Schmidt The University of Manchester, UK Hao Feng North China University of Science & Technology, China Giorgos Stoilos Babylon Health, UK Damir Juric Babylon Health, UK Mohammad Khodadadi Babylon Health, UK |
| Pseudocode | Yes | Figure 1: The combination rule for eliminating A sig C(N) from a set N of clauses in A-reduced form. Figure 2: The combination rule for eliminating r sig R(N) from a set N of clauses in r-reduced form. |
| Open Source Code | Yes | The tool, along with the prototype for forgetting and the test data sets, can be downloaded via http://www.cs.man.ac.uk/ schmidt/publications/aaai19/. |
| Open Datasets | Yes | We evaluated it on a snapshot of the NCBO Bio Portal repository4 taken in March 2017 (Matentzoglu and Parsia 2017), containing 396 OWL API compatible ontologies. ... Our target are SNOMED CT and NCIt ontologies. |
| Dataset Splits | No | The paper describes evaluating the method on existing ontologies and their versions but does not specify a traditional train/validation/test split for a model or algorithm. |
| Hardware Specification | Yes | The experiments were run on a desktop computer with an Intel Core i7-4790 processor, four cores running at up to 3.60 GHz, and 8 GB of DDR3-1600 MHz RAM. |
| Software Dependencies | Yes | we implemented a prototype in Java using the OWL API Version 3.5.63 |
| Experiment Setup | Yes | The prototype was evaluated for three settings: forgetting 10% (199), 30% (597) and 50% (995) of concept and role names in the signature of each ontology. ... A timeout of 1000 seconds was imposed on each run. |