IBM.Next: QuerioDALI       


Spyros Kotoulas photoVanessa Lopez  photo photoPierpaolo Tommasi photo

IBM.Next: QuerioDALI - Evaluation

QuerioDALI has been evaluated using the following state of the art benchmarks over DBpedia, Freebase and biomedical ontologies:

  • QALD-5 test set (50 questions over DBpedia): F-measure of 0.64
  • QALD-4 train set (25 questions over biomedical ontologies) : F-measure of 072
  • Free917 dataset over Freebase (100 questions) : F-measure of 0.85

 Detailed evaluation results at:

For more info about QALD benchmarks: Question Answering over Linked Data campaigns