← Go back

Research Colloquium Presentation by Özge Erten

2 years ago

Automatic Rewriting of SPARQL Queries Between Versions of a Knowledge Graph

Özge Erten - ESR 13, Maastricht University

The presentation's main topic is to give brief information about research that is ongoing. When a new version of an RDF knowledge graph is released and published as a result of a set of changes, it might cause issues in the pipelines using the knowledge graph. As the result, some queries produce no results or incorrect results either due to the change in schema representation or data. An automatic approach to updating SPARQL queries instead of a manual approach would decrease spending time on updating SPARQL queries. In the scope of the research, the queries that were constructed for the previous version of the KG will be adapted to the new KG structure through an automated approach. And this research focuses on two main aspects of developing such an approach, one of them is to detect knowledge graph version changes and the other is to automatically update SPARQL queries. A baseline method will be used for capturing changes between two versions of an RDF KG, and a new method will be developed to first analyze the captured changes to eliminate that will only consider no data-loss schema changes, and secondly, SPARQL queries will be evaluated and the queries that fit the scope and limitations will be updated automatically.  

Please find the presentation file for the detailed explanation here