Stefan presented zelph, an innovative open-source semantic network system designed to improve Wikidata’s data quality, at the Wikimedia Hackathon 2025. The presentation demonstrated how zelph processes the complete Wikidata dataset (1.7 TB), detects logical contradictions, and makes automatic inferences based on existing relationships.

Particularly highlighted was zelph’s ability to define inference rules within the network itself and perform meta-reasoning about relationships. A live demonstration illustrated the practical application using real examples from Wikidata.

For more information about zelph, visit https://zelph.org and https://github.com/acrion/zelph