ACSMKRHR at SemEval-2023 Task 10: Explainable Online Sexism Detection(EDOS)

Proceedings of the The 17th International Workshop on Semantic Evaluation (SemEval-2023), pages 724–732 July 13-14, 2023 ©2023 Association for Computational Linguistics ACSMKRHR at SemEval-2023 Task 10: Explainable Online Sexism Detection(EDOS) Rakib Hossain Rifat Abanti Chakraborty Shruti Marufa Kamal Farig Sadeque BRAC University, Dhaka, Bangladesh rakib.hossain.rifat@g.bracu.ac.bd, abanti.chakraborty.shruti@g.bracu.ac.bd, marufa.kamal1@g.bracu.ac.bd, farig.sadeque@bracu.ac.bd Abstract People are expressing their opinions online for a lot of years now. Although these opinions and comments provide people an opportunity of expressing their views, there is a lot of hate speech that can be found online. More specifi- cally, sexist comments are very popular affect- ing and creating a negative impact on a lot of women and girls online. This paper describes the approaches of the SemEval-2023 Task 10 competition for Explainable Online Sexism De- tection (EDOS). The task has been divided into 3 subtasks, introducing different classes of sex- ist comments. We have approached these tasks using the bert-cased and uncased models which are trained on the annotated dataset that has been provided in the competition. Task A pro- vided the best F1 score of 80% on the test set, and tasks B and C provided 58% and 40% re- spectively. 1 Introduction Sexism is prejudice or discrimination based on one’s sex or gender. Sexism can affect anyone, but it primarily affects women and girls who face nega- tive sentiment or abuse based on their gender com- bined with one or more other identity attributes (e.g. Black women, Muslim women, Trans women). It has been linked to stereotypes and gender roles and may include the belief that one sex or gender is in- trinsically superior to another(Wikipedia contribu- tors, 2023). Social media platforms have promoted the spread of hate speech through anonymization and accessibility, spurring greater study into devel- oping automatic algorithms to recognize these sorts of writings. According to a study(Duggan, 2017) on online harassment, women experience harass- ment due to their gender twice as often as males do. Bullying online can lead to depression, more- over, a study(Fulper et al., 2014) related to rape cases found a connection between the number of sexist tweets and the number of rapes in the United States of America. An online poll conducted by Amnesty International across eight high-income countries in 2017 revealed that 23% of women had experienced some form of abuse or harassment on social media platforms(Center on Gender Eq- uity and Health (GEH) – UC San Diego, 2023). It can cause harm to targeted women, render online places inhospitable and inaccessible, and maintain societal inequities and asymmetries. Social media text can be inconsistent and mixed with sarcasm often times which can make it challenging to detect sexist texts. The categorization of sexism differs from and may be complemented by the detection of hate speech. Although sexist comments may be re- garded as hate speech and there is a lot of research conducted regarding hate speech detection, sexism sentences have a broader aspect, and studies on sexism detection have a huge scope. Despite us- ing various automated tools in …

Content truncated to 500 words