Uncover Sexual Harassment Patterns from Personal Stories by Joint Key Element Extraction and Categorization

Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, pages 2328–2337, Hong Kong, China, November 3–7, 2019. c⃝2019 Association for Computational Linguistics 2328 Uncover Sexual Harassment Patterns from Personal Stories by Joint Key Element Extraction and Categorization Yingchi Liu1, Quanzhi Li1, Marika Cifor2, Xiaozhong Liu3, Qiong Zhang1 and Luo Si1 1Alibaba Group, US 2University of Washington, USA 3Indiana University Bloomington, USA {yingchi.liu, quanzhi.li, qz.zhang,luo.si}@alibaba-inc.com mcifor@uw.edu, liu237@indiana.edu Abstract The number of personal stories about sexual harassment shared online has increased expo- nentially in recent years. This is in part in- spired by the #MeToo and #TimesUp move- ments. Safecity is an online forum for peo- ple who experienced or witnessed sexual ha- rassment to share their personal experiences. It has collected >10,000 stories so far. Sex- ual harassment occurred in a variety of situ- ations, and categorization of the stories and extraction of their key elements will provide great help for the related parties to understand and address sexual harassment. In this study, we manually annotated those stories with la- bels in the dimensions of location, time, and harassers’ characteristics, and marked the key elements related to these dimensions. Further- more, we applied natural language processing technologies with joint learning schemes to automatically categorize these stories in those dimensions and extract key elements at the same time. We also uncovered significant pat- terns from the categorized sexual harassment stories. We believe our annotated data set, pro- posed algorithms, and analysis will help peo- ple who have been harassed, authorities, re- searchers and other related parties in various ways, such as automatically filling reports, en- lightening the public in order to prevent fu- ture harassment, and enabling more effective, faster action to be taken. 1 Introduction Sexual violence, including harassment, is a perva- sive, worldwide problem with a long history. This global problem has finally become a mainstream issue thanks to the efforts of survivors and advo- cates. Statistics show that girls and women are put at high risk of experiencing harassment. Women have about a 3 in 5 chance of experiencing sexual harassment, whereas men have slightly less than 1 in 5 chance (Quinnipiac, 2017; EEOC, 2018; Goldstein, 2018). While women in developing countries are facing distinct challenges with sexual violence (Lea et al., 2017), however sexual vio- lence is ubiquitous. In the United States, for exam- ple, there are on average >300,000 people who are sexually assaulted every year (Morgan and Tru- man, 2018). Additionally, these numbers could be underestimated, due to reasons like guilt, blame, doubt and fear, which stopped many survivors from reporting (Griffith, 2018). Social media can be a more open and accessible channel for those who have experienced harassment to be empow- ered to freely share their traumatic experiences and to raise awareness of the vast scale of sexual harassment, which then allows us to understand and actively address abusive behavior as part of larger efforts to prevent future sexual harassment. The deadly …

Content truncated to 500 words