QALD-10: Wikidata Dump for KGQA Benchmark Evaluation

QALD-10 Wikidata Dump is published to enable fair and replicable evaluation of Knowledge Graph Question Answering systems. The data supports the newly released QALD-10 Challenge benchmark. Ricardo Usbeck from Universität Hamburg authored this resource for researchers to set up a local Wikidata endpoint.

Use Cases

• Benchmarking KGQA system performance based on the QALD-10 challenge questions
• Training KGQA models using structured Wikidata knowledge
• Setting up a local Wikidata endpoint for controlled experimentation

Strengths

• Designed specifically for fair and replicable evaluation of KGQA systems
• Associated with the QALD-10 Challenge, a recognized benchmark

Limitations

• Column-level documentation is absent; field semantics must be inferred after download
• Row count is unknown, which may limit suitability assessment

Provenance

Source

Wikidata

Collection Method

Data dump published for benchmark evaluation