UMUTeam at SemEval-2023 Task 10: Fine-grained detection of sexism in English

Proceedings of the The 17th International Workshop on Semantic Evaluation (SemEval-2023), pages 589–594 July 13-14, 2023 ©2023 Association for Computational Linguistics UMUTeam at SemEval-2023 Task 10: Fine-grained detection of sexism in English Ronghao Pan1, José Antonio García-Díaz1, Salud María Jiménez Zafra2, Rafael Valencia-García1 1 Facultad de Informática, Universidad de Murcia, Campus de Espinardo, 30100 Murcia, Spain {ronghao.pan@um.es, joseantonio.garcia8, valencia}@um.es 2 Computer Science Department, SINAI, CEATIC, Universidad de Jaén, 23071, Spain sjzafra@ujaen.es Abstract In this manuscript, we describe the participa- tion of UMUTeam in the Explainable Detection of Online Sexism shared task proposed at Se- mEval 2023. This task concerns the precise and explainable detection of sexist content on Gab and Reddit, i.e., developing detailed classifiers that not only identify what is sexist, but also explain why it is sexism. Our participation in the three EDOS subtasks is based on extending new unlabeled sexism data in the Masked Lan- guage Model task of a pre-trained model, such as RoBERTa-large to improve its generalization capacity and its performance on classification tasks. Once the model has been pre-trained with the new data, fine-tuning of this model is performed for different specific sexism classifi- cation tasks. Our system has achieved excellent results in this competitive task, reaching top 24 (84) in Task A, top 23 (69) in Task B, and top 13 (63) in Task C. 1 Introduction The term sexism refers to any abuse or negative sen- timent directed at women on the basis of their sex, or on the basis of their sex combined with one or more identity attributes. It should be noted that sex- ism is a growing problem on the Internet and can have harmful effects on women and other marginal- ized groups, making online spaces less accessible and unwelcoming, which can perpetuate asymme- tries and social injustices. Automated tools are already widely used today to identify and evaluate sexist content online. It is important that these tools provide specific ratings and explain why content is considered sexist to improve the interpretability, re- liability, and comprehensibility of decisions made by automated tools. However, most only provide generic high-level rating without further explana- tion. Thus, the Explainable Detection of Online Sexism (EDOS) shared task (Kirk et al., 2023), pro- posed at SemEval 2023, supports the development of English-language models for sexism detection that are more accurate and explainable, with pre- cise classifications of sexist content on Gab and Reddit. To this end, the task is divided into three hierarchical subtasks: • Binary sexism detection (Task A): a binary classification where systems have to predict whether a post is sexist or not sexist. • Category of sexism (Task B): for posts which are sexist, a four-class classification where systems have to predict one of four categories (threats, derogation, animosity, and prejudiced discussions). • Fine-grained vector of sexism (Task C): for post which are sexist, an 11-class classifica- tion where systems have to predict one of 11 fine-grained vectors1 (1.1 threats of harm, 1.2 incitement and encouragement of harm, 2.1 descriptive attacks, …

Content truncated to 500 words