Data

The following datasets are made publicly available in order to encourage reproducibility of results, comparison between different approaches, novel applications and new scientific developments:

Crowdsourcing datasets from Amazon Mechanical Turk
  • LabelMe (multi-class classification; annotations from Amazon Mechanical Turk)
  • MovieReviews (regression; annotations from Amazon Mechanical Turk)
  • 2003 CONLL NER task (sequence labelling; annotations from Amazon Mechanical Turk)
Citation: