Published August 1, 2025 | Version v1
Dataset Open

Learning problems of 5 QA systems on QALD10, QALD9+DB, and QALD9+WK

Authors/Creators

  • 1. ROR icon Paderborn University

Description

This project provides learning problems of 5 QA Systems on QALD10, QALD9+DB, and QALD9+WK in JSON format. Each one is the performance of a QA system and the positive and negative examples represent questions that the systems answered correctly (F1 >= 0.5) or incorrectly (F1 < 0.5), respectively.

The following JSON snippet shows an example of a learning problem for the MST5 on QALD10 dataset:
{
  "problems": {
    "MST5_QALD10": {
      "positive_examples": [
        "https://github.com/KGQA/QALD-10/blob/main/data/qald_10/qald_10.json#Q0",
        ...
      ],
      "negative_examples": [
        "https://github.com/KGQA/QALD-10/blob/main/data/qald_10/qald_10.json#Q1",
        ...
      ]
    }
  }
}

Three knowledge graphs built for QALD 9 plus DBpedia, QALD 9 plus Wikidata and QALD10 can be found here.

The three knowledge graphs are used to solve the learning problems using concept learner provided by ontolearn.



Files

TF_Json.zip

Files (7.6 kB)

Name Size Download all
md5:df01249a8855bf737df8583c048bbb44
7.6 kB Preview Download