Swedish EAT: question classification

What the BERT? Fine-tuning KB-BERT for Question Classification. Unpublished manuscript, School of Electrical Engineering and Computer Science, KTH.

Related datasets II. USAGE Key applications Machine learning, EAT Classification Intended task(s)/usage(s) Evaluate models by standard classification Recommended evaluation measures Accuracy Dataset function(s) Testing Recommended split(s) Test only III. DATA Primary data* Text Language* Swedish Dataset in numbers* 5451 questions in training set, 500 in test set.

Nature of the content* Open ended factoid questions. Format* Comma-separated, four columns: text -- the open ended factoid question verbose label -- both the coarse-grained label and the fine-grained label formatted as COARSE:fine coarse label -- coarse-grained label fine label -- fine-grained label Data source(s)* Translated from the QAQC dataset (cogcomp.seas.upenn.edu/Data/QA/QC/) Data collection method(s)* -- Data selection and filtering* -- Data preprocessing* -- Data labeling* -- Annotator characteristics IV.

ETHICS AND CAVEATS Ethical considerations "Some outdated treatment of women (eg "Vilka är de sexigaste kvinnorna i världen?")" Things to watch out for V. ABOUT DOCUMENTATION Data last updated* 2021-07-27 Which changes have been made, compared to the previous version* First version Access to previous versions This document created* 2021-07-27 This document last updated* 2023-06-08 Where to look for further details Documentation template version* VI.

OTHER Related projects References