This is the repository presented in the paper "Annotating Online Misogyny" from ACL 2021.
Annotated Corpus in Danish sampled from Twitter, Facebook, Reddit.
For more information, see the stromberg.ai/publication/aom.
Access to the data can be granted under NDA for research purposes. Hugging Face is used to mediate access; apply at strombergnlp/bajer_danish_misogyny.
By filling the form you submit a request for access to data and annotation.
The repository exists of:
- data
- dataset
- out-of-context-posts (sorted out)
- annotation
- codebook
- data_collection_keyword_list
- additional data
- Danish slurs: extending Reddit survey list from Sigurbergsson, Derczynski* on Danish known slurs (free Google search for annotators)
- Translations of posts from IberEval/Evalita (English) to Danish
- counter-examples stereotypes: transforming Danish stereotypical posts to their counter-example (total ~30 posts, tasks turned out to be too challenging)
- additional information from the annotation task
- feedback annotators and motivation
Lastly, feel free to reach out regarding any enquiries around the project.
Please cite:
Zeinert, P., Inie, N., Derczynski, L., 2021. Annotating Online Misogyny, in: Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). Presented at the ACL-IJCNLP 2021, Association for Computational Linguistics, Online, pp. 3181–3197.
Bibtex:
@inproceedings{zeinert_annotating_2021,
address = {Online},
title = {Annotating {Online} {Misogyny}},
booktitle = {Proceedings of the 59th {Annual} {Meeting} of the {Association} for {Computational} {Linguistics} and the 11th {International} {Joint} {Conference} on {Natural} {Language} {Processing} ({Volume} 1: {Long} {Papers})},
publisher = {Association for Computational Linguistics},
author = {Zeinert, Philine and Inie, Nanna and Derczynski, Leon},
month = aug,
year = {2021},
pages = {3181--3197},
doi = {http://dx.doi.org/10.18653/v1/2021.acl-long.247}
}