(Last updated: January 27, 2016)
Task 1: Multilingual question answering over RDF dataGiven a natural language question or keywords, retrieve the correct answer(s) from a given RDF repository.
- Dataset: DBpedia 2015
- Training questions: qald-6-train-multilingual.json
- Test questions: qald-6-test-multilingual.json (without annotations: qald-6-test-multilingual-questions.json)
Task 2: Hybrid question answering over RDF data and free textGiven a natural language question or keywords, determine the correct answer(s) by combining information from a given RDF repository and free text.
- Dataset: DBpedia 2015 (with free text abstracts)
- Training questions: qald-6-train-hybrid.json
- Test questions: qald-6-test-hybrid.json (without annotations: qald-6-test-hybrid-questions.json)
Task 3: Statistical question answering over RDF datacubesGiven a natural language question or keywords, retrieve the correct answer(s) from an RDF datacube repository.
- Dataset: LinkedSpending
- RDF dump: qbench2datasets.zip
- Endpoint: http://cubeqa.aksw.org/sparql
- Training questions: qald-6-train-datacube.json
- Test questions: qald-6-test-datacube.json (without annotations: qald-6-test-datacube-questions.json)
|Release of training data and instructions:||January 15, 2016|
|Paper submission deadline:||March 21, 2016|
|Notification of acceptance:||April 11, 2016||Release of test data:||April 11, 2016|
|Deadline for submission of system answers:||April 18, 2016|
|Release of evaluation results:||April 20, 2016|
|Submission of camera-ready papers:||April 24, 2016|
|ESWC challenge track:||May 31, 2016|