SemEval-2023 Task 10: Explainable Detection of Online Sexism
Authors:
- Bertie VidgenBVBertie Vidgen
Preprint
Abstract
Online sexism is a widespread and harmful phenomenon. Automated tools can assist the detection of sexism at scale. Binary detection, however, disregards the diversity of sexist content, and fails to provide clear explanations for why something is sexist. To address this issue, we introduce SemEval Task 10 on the Explainable Detection of Online Sexism (EDOS). We make three main contributions: i) a novel hierarchical taxonomy of sexist content, which includes granular vectors of sexism to aid explainability; ii) a new dataset of 20,000 social media comments with fine-grained labels, along with larger unlabelled datasets for model adaptation; and iii) baseline models as well as an analysis of the methods, results and errors for participant submissions to our task.Keywords
All Related Versions
This publication has 44 references indexed in Scilit:
- TwHIN-BERT: A Socially-Enriched Pre-trained Language Model for Multilingual Tweet Representations at Twitter2023
- Power to the People? Opportunities and Challenges for Participatory AI2022
- Hidden behind the obvious: Misleading keywords and implicitly abusive language on social mediaOnline Social Networks and Media, 2022
- Participatory Design of AI Systems: Opportunities and Challenges Across Diverse Users, Relationships, and Application Domains2022
- Explainable Abusive Language Classification Leveraging User and Network Data2021
- Categorizing Sexism and Misogyny through Neural ApproachesACM Transactions on the Web, 2021
- Exploring Misogyny across the Manosphere in Reddit2019
- Data Statements for Natural Language Processing: Toward Mitigating System Bias and Enabling Better ScienceTransactions of the Association for Computational Linguistics, 2018
- Inside the Right-Leaning Echo Chambers: Characterizing Gab, an Unmoderated Social System2018
- What is Gab2018