More Like this

1. Hierarchical Attention Networks for Cyberbullying Detection on the Instagram Social Network

   https://doi.org/10.1137/1.9781611975673.27  

   Cheng, Lu ; Guo, Ruocheng ; Silva, Yasin N. ; Hall, Deborah L. ; Liu, Huan ( May 2019 , Proceedings of the ... SIAM International Conference on Data Mining)

   Cyberbullying has become one of the most pressing online risks for young people and has raised serious concerns in society. The emerging literature identifies cyberbullying as repetitive acts that occur over time rather than one-off incidents. Yet, there has been relatively little work to model the hierarchical structure of social media sessions and the temporal dynamics of cyberbullying in online social network sessions. We propose a hierarchical attention network for cyberbullying detection that takes these aspects of cyberbullying into account. The primary distinctive characteristics of our approach include: (i) a hierarchical structure that mirrors the structure of a social media session; (ii) levels of attention mechanisms applied at the word and comment level, thereby enabling the model to pay different amounts of attention to words and comments, depending on the context; and (iii) a cyberbullying detection task that also predicts the interval of time between two adjacent comments. These characteristics allow the model to exploit the commonalities and differences across these two tasks to improve the performance of cyberbullying detection. Experiments on a real-world dataset from Instagram, the social media platform on which the highest percentage of users have reported experiencing cyberbullying, reveal that the proposed architecture outperforms the state-of-the-art method. 

   more » « less
2. Improving Cyberbullying Detection with User Interaction

   https://doi.org/10.1145/3442381.3449828  

   Ge, S. ; Cheng, L. ; Liu, H. ( April 2021 , The Web Conference)

   null (Ed.)

   Cyberbullying, identified as intended and repeated online bullying behavior, has become increasingly prevalent in the past few decades. Despite the significant progress made thus far, the focus of most existing work on cyberbullying detection lies in the independent content analysis of different comments within a social media session. We argue that such leading notions of analysis suffer from three key limitations: they overlook the temporal correlations among different comments; they only consider the content within a single comment rather than the topic coherence across comments; they remain generic and exploit limited interactions between social media users. In this work, we observe that user comments in the same session may be inherently related, e.g., discussing similar topics, and their interaction may evolve over time. We also show that modeling such topic coherence and temporal interaction are critical to capture the repetitive characteristics of bullying behavior, thus leading to better predicting performance. To achieve the goal, we first construct a unified temporal graph for each social media session. Drawing on recent advances in graph neural network, we then propose a principled graph-based approach for modeling the temporal dynamics and topic coherence throughout user interactions. We empirically evaluate the effectiveness of our approach with the tasks of session-level bullying detection and comment-level case study. Our code is released to public. 

   more » « less
3. Mitigating Bias in Session-based Cyberbullying Detection: A Non-Compromising Approach

   https://doi.org/10.18653/v1/2021.acl-long.168  

   Cheng, L. ; Mosallanezhad, A. ; Silva, Y. N. ; Hall, D. L. ; Liu, H. ( August 2021 , The Joint Conference of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (ACL-IJCNLP))

   null (Ed.)

   The element of repetition in cyberbullying behavior has directed recent computational studies toward detecting cyberbullying based on a social media session. In contrast to a single text, a session may consist of an initial post and an associated sequence of comments. Yet, emerging efforts to enhance the performance of session-based cyberbullying detection have largely overlooked unintended social biases in existing cyberbullying datasets. For example, a session containing certain demographic-identity terms (e.g., “gay” or “black”) is more likely to be classified as an instance of cyberbullying. In this paper, we first show evidence of such bias in models trained on sessions collected from different social media platforms (e.g., Instagram). We then propose a context-aware and model-agnostic debiasing strategy that leverages a reinforcement learning technique, without requiring any extra resources or annotations apart from a pre-defined set of sensitive triggers commonly used for identifying cyberbullying instances. Empirical evaluations show that the proposed strategy can simultaneously alleviate the impacts of the unintended biases and improve the detection performance. 

   more » « less
4. Unsupervised Cyberbullying Detection via Time-Informed Deep Clustering

   Cheng, L. ; Shu, K. ; Wu, S. ; Silva, Y. ; Hall, D. ; Liu, H. ( January 2020 , The 29th ACM International Conference on Information and Knowledge Management (CIKM))

   Social media is a vital means for information-sharing due to its easy access, low cost, and fast dissemination characteristics. However, increases in social media usage have corresponded with a rise in the prevalence of cyberbullying. Most existing cyberbullying detection methods are supervised and, thus, have two key drawbacks: (1) The data labeling process is often labor-intensive and time-consuming; (2) Current labeling guidelines may not be generalized to future instances because of different language usage and evolving social networks. To address these limitations, this work introduces a principled approach for unsupervised cyberbullying detection. The proposed model consists of two main components: (1) A representation learning network that encodes the social media session by exploiting multi-modal features, e.g., text, network, and time. (2) A multi-task learning network that simultaneously fits the time intervals and estimates the bullying likelihood based on a Gaussian Mixture Model. The proposed model jointly optimizes the parameters of both components to overcome the shortcomings of decoupled training. Our core contribution is an unsupervised cyberbullying detection model that not only experimentally outperforms the state-of-the-art unsupervised models, but also achieves competitive performance compared to supervised models. 

   more » « less
5. Sexist Slurs: Reinforcing Feminine Stereotypes Online

   https://doi.org/10.1007/s11199-019-01095-z  

   Felmlee, Diane ; Inara Rodis, Paulina ; Zhang, Amy ( November 2019 , Sex Roles)

   Social media platforms are accused repeatedly of creating environments in which women are bullied and harassed. We argue that online aggression toward women aims to reinforce traditional feminine norms and stereotypes. In a mixed methods study, we find that this type of aggression on Twitter is common and extensive and that it can spread far beyond the original target. We locate over 2.9 million tweets in one week that contain instances of gendered insults (e.g., “bitch,” “cunt,” “slut,” or “whore”)—averaging 419,000 sexist slurs per day. The vast majority of these tweets are negative in sentiment. We analyze the social networks of the conversations that ensue in several cases and demonstrate how the use of “replies,” “retweets,” and “likes” can further victimize a target. Additionally, we develop a sentiment classifier that we use in a regression analysis to compare the negativity of sexist messages. We find that words in a message that reinforce feminine stereotypes inflate the negative sentiment of tweets to a significant and sizeable degree. These terms include those insulting someone’s appearance (e.g., “ugly”), intellect (e.g., “stupid”), sexual experience (e.g., “promiscuous”), mental stability (e.g., “crazy”), and age (“old”). Messages enforcing beauty norms tend to be particularly negative. In sum, hostile, sexist tweets are strategic in nature. They aim to promote traditional, cultural beliefs about femininity, such as beauty ideals, and they shame victims by accusing them of falling short of these standards. Harassment on social media constitutes an everyday, routine occurrence, with researchers finding 9,764,583 messages referencing bullying on Twitter over the span of two years (Bellmore et al. 2015). In other words, Twitter users post over 13,000 bullying-related messages on a daily basis. Forms of online aggression also carry with them serious, negative consequences. Repeated research documents that bullying victims suffer from a host of deleterious outcomes, such as low self-esteem (Hinduja and Patchin 2010), emotional and psychological distress (Ybarra et al. 2006), and negative emotions (Faris and Felmlee 2014; Juvonen and Gross 2008). Compared to those who have not been attacked, victims also tend to report more incidents of suicide ideation and attempted suicide (Hinduja and Patchin 2010). Several studies document that the targets of cyberbullying are disproportionately women (Backe et al. 2018; Felmlee and Faris 2016; Hinduja and Patchin 2010; Pew Research Center 2017), although there are exceptions depending on definitions and venues. Yet, we know little about the content or pattern of cyber aggression directed toward women in online forums. The purpose of the present research, therefore, is to examine in detail the practice of aggressive messaging that targets women and femininity within the social media venue of Twitter. Using both qualitative and quantitative analyses, we investigate the role of gender norm regulation in these patterns of cyber aggression. 

   more » « less