Multi-label Categorization of Accounts of Sexism using a Neural Framework

Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, pages 1642–1652, Hong Kong, China, November 3–7, 2019. c⃝2019 Association for Computational Linguistics 1642 Multi-label Categorization of Accounts of Sexism using a Neural Framework Pulkit Parikh1, Harika Abburi1, Pinkesh Badjatiya1, Radhika Krishnan1, Niyati Chhaya1,2, Manish Gupta1∗, Vasudeva Varma1 1IIIT-Hyderabad, India 2Adobe Research, India {pulkit.parikh,harika.a,pinkesh.badjatiya}@research.iiit.ac.in, radhika.krishnan@iiit.ac.in, nchhaya@adobe.com, {manish.gupta,vv}@iiit.ac.in Abstract Sexism, an injustice that subjects women and girls to enormous suffering, manifests in bla- tant as well as subtle ways. In the wake of growing documentation of experiences of sex- ism on the web, the automatic categorization of accounts of sexism has the potential to as- sist social scientists and policy makers in uti- lizing such data to study and counter sexism better. The existing work on sexism classifica- tion, which is different from sexism detection, has certain limitations in terms of the cate- gories of sexism used and/or whether they can co-occur. To the best of our knowledge, this is the first work on the multi-label classification of sexism of any kind(s), and we contribute the largest dataset for sexism categorization. We develop a neural solution for this multi- label classification that can combine sentence representations obtained using models such as BERT with distributional and linguistic word embeddings using a flexible, hierarchical ar- chitecture involving recurrent components and optional convolutional ones. Further, we lever- age unlabeled accounts of sexism to infuse domain-specific elements into our framework. The best proposed method outperforms several deep learning as well as traditional machine learning baselines by an appreciable margin. 1 Introduction Sexism, discrimination on the basis of one’s sex, prevails in our society in numerous forms, caus- ing immense suffering to women and girls. Online forums have enabled victims of sexism to share their experiences freely and widely by facilitat- ing anonymity and connecting far-away people. A meaningful categorization of these accounts of sexism can play a part in analyzing sexism with a view to developing sensitization programs, sys- temic safeguards, and other mechanisms against ∗The author is also an applied researcher at Microsoft. this injustice. Given the rising volume of such information on digital media, automated sexism categorization can aid social scientists and policy makers in combating sexism by conducting such analyses efficiently. While sexism is detected as a category of hate in some of the hate speech classification work (Bad- jatiya et al., 2017; Waseem and Hovy, 2016), it does not perform sexism classification. Except the work on categorizing sexual harassment by Kar- lekar and Bansal (2018), the prior work on clas- sifying sexism assumes the categories to be mu- tually exclusive (Anzovino et al., 2018; Jha and Mamidi, 2017). Moreover, the existing category sets number between 2 to 5. In this paper, we fo- cus on the new problem of the multi-label cate- gorization of an account of sexism reporting any type(s) of sexism. We create a dataset comprising 13023 accounts of sexism, including first-person accounts from survivors, each tagged …

Content truncated to 500 words