#YouToo? Detection of Personal Recollections of Sexual Harassment on Social Media

Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 2527–2537 Florence, Italy, July 28 - August 2, 2019. c⃝2019 Association for Computational Linguistics 2527 #YouToo? Detection of Personal Recollections of Sexual Harassment on Social Media Arijit Ghosh Chowdhury∗ Manipal Institute of Technology arijit10@gmail.com Ramit Sawhney∗ Netaji Subhas Institute of Technology ramits.co@nsit.net.in Rajiv Ratn Shah MIDAS, IIIT-Delhi rajivratn@iiitd.ac.in Debanjan Mahata Bloomberg dmahata@bloomberg.net Abstract The availability of large-scale online so- cial data, coupled with computational meth- ods, can help us answer fundamental ques- tions relating to our social lives, particularly our health and well-being. The #MeToo trend has led to people talking about personal expe- riences of harassment more openly. This work attempts to aggregate such experiences of sex- ual abuse to facilitate a better understanding of social media constructs and to bring about social change. It has been found that disclo- sure of abuse has positive psychological im- pacts. Hence, we contend that such informa- tion can be leveraged to create better cam- paigns for social change by analyzing how users react to these stories and can be used to obtain a better insight into the consequences of sexual abuse. We use a three-part Twitter- Specific Social Media Language Model to seg- regate personal recollections of sexual harass- ment from Twitter posts. An extensive com- parison with state-of-the-art generic and spe- cific models along with a detailed error analy- sis explores the merit of our proposed model. 1 Introduction Global estimates indicate that about 1 in 3 women worldwide has experienced either physical and/or sexual intimate partner violence or non-partner sexual violence in their lifetime 1. The hashtag #MeToo has been prevalent on Twitter as a cam- paign centered around sharing stories of sexual ha- rassment in the act of solidarity with other victims and spreading awareness of a widespread and en- demic issue. With vast amounts of people sharing their recollections of sexual harassment on the In- ternet, it is important that we make scientific use * denotes equal contribution. 1https://www.who.int/news-room/fact- sheets/detail/violence-against-women of this data to increase awareness and enable real- world change. Manually sorting and comprehend- ing the information shared in these stories is an ar- duous task. Hence our work can serve as the miss- ing link between online activism and real change. Health information seeking and sharing prac- tices online have been known in helping people cope with mental health problems (De Choud- hury and De, 2014). Studies have shown that on- line forums and support groups provide a con- ducive environment allowing people to get con- nected with others who share similar stories, thus act as a path to obtaining help and advice around mental health problems (Eysenbach et al., 2004). Moreover, self-disclosure is therapeutic for men- tal health communities (Johnson and Ambrose, 2006). Our study proposes a Twitter-Specific Social Media Language Model for the aggregation of tweets containing personal stories of sexual ha- rassment. Manikonda et al. (2018) have carried out a preliminary analysis of the user engagement, discussion topics, word …

Content truncated to 500 words