Ruddit: Norms of offensiveness for English Reddit comments

Open Access
Authors
  • R. Hada
  • S. Sudhir
  • P. Mishra
  • H. Yannakoudakis
Publication date 2021
Host editors
  • C. Zong
  • F. Xia
  • W. Li
  • R. Navigli
Book title The 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing
Book subtitle ACL-IJCNLP 2021 : proceedings of the conference : August 1-6, 2021
ISBN (electronic)
  • 9781954085527
Event The Joint Conference of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (ACL-IJCNLP 2021)
Volume | Issue number 1
Pages (from-to) 2700-2717
Number of pages 18
Publisher Stroudsburg, PA: The Association for Computational Linguistics
Organisations
  • Interfacultary Research - Institute for Logic, Language and Computation (ILLC)
Abstract

Warning: This paper contains comments that may be offensive or upsetting. On social media platforms, hateful and offensive language negatively impact the mental well-being of users and the participation of people from diverse backgrounds. Automatic methods to detect offensive language have largely relied on datasets with categorical labels. However, comments can vary in their degree of offensiveness. We create the first dataset of English language Reddit comments that has fine-grained, real-valued scores between -1 (maximally supportive) and 1 (maximally offensive). The dataset was annotated using Best-Worst Scaling, a form of comparative annotation that has been shown to alleviate known biases of using rating scales. We show that the method produces highly reliable offensiveness scores. Finally, we evaluate the ability of widely-used neural models to predict offensiveness scores on this new dataset.

Document type Conference contribution
Note With supplementary video
Language English
Published at https://doi.org/10.18653/v1/2021.acl-long.210
Other links https://paperswithcode.com/paper/ruddit-norms-of-offensiveness-for-english https://www.scopus.com/pages/publications/85113605704
Downloads
2021.acl-long.210v2 (Final published version)
Supplementary materials
Permalink to this page
Back