Browsing by Author "Wang, Pei"

Now showing 1 - 3 of 3
  • Thumbnail Image
    Generative Chatbot Framework for Cybergrooming Prevention
    Wang, Pei (Virginia Tech, 2021-12-20)
    Cybergrooming refers to the crime of establishing personal close relationships with potential victims, commonly teens, for the purpose of sexual exploitation or abuse via online social media platforms. Cybergrooming has been recognized as a serious social problem. However, there have been insufficient programs to provide proactive prevention to protect the youth users from cybergrooming. In this thesis, we present a generative chatbot framework, called SERI (Stop cybERgroomIng), that can generate simulated conversations between a perpetrator chatbot and a potential victim chatbot. To realize the simulation of authentic conversations in the context of cybergrooming, we take deep reinforcement learning (DRL)-based dialogue generation to simulate the authentic conversations between a perpetrator and a potential victim. The design and development of the SERI are motivated to provide a safe and authentic chatting environment to enhance the youth's precautionary awareness and sensitivity of cybergrooming while any unnecessary ethical issues (e.g., the potential misuse of the SERI) are removed or minimized. We developed the SERI as a preliminary platform that the perpetrator chatbot can be deployed in social media environments to interact with human users (i.e., youth) and observe the conversations that the youth users respond to strangers or acquaintances when they are asked for private or sensitive information by the perpetrator. We evaluated the quality of conversations generated by the SERI based on open-source, referenced, and unreferenced metrics as well as human evaluation. The evaluation results show that the SERI can generate authentic conversations between two chatbots compared to the original conversations from the used datasets in perplexity and MaUde scores.
  • Thumbnail Image
    SERI: Generative Chatbot Framework for Cybergrooming Prevention
    Wang, Pei; Guo, Zhen; Huang, Lifu; Cho, Jin-Hee (2021-11-11)
    Cybergrooming refers to a crime to lure potential victims, particularly youth, by establishing personal trust relationships with them for sexual abuse or exploitation. Although cybergrooming is recognized as one of the serious social issues, there has been a lack of proactive programs to protect the youth. In this paper, we present a generative chatbot framework, called SERI Stop cybERgroomIng), that can generate authentic conversations between a perpetrator chatbot and a potential victim chatbot. The SERI is designed to provide a safe and authentic environment for enhancing youth's sensitivity and awareness of subtle cues of cybergrooming without exposing unnecessary ethical issues caused by potentially offensive or upsetting languages. The SERI is developed as a pre-stage before the perpetrator chatbot is deployed to chatting with an actual human youth user to observe how the youth user can respond to a stranger or acquaintance asking for sensitive or private information. Hence, to evaluate the quality of the conversations generated by the SERI, we use open-source, referenced, and unreferenced metrics to assess the generated conversations automatically. In addition, we evaluated the quality of the conversation based on the human evaluation method. Our results show that the SERI can generate authentic conversations between the two chatbots compared to the original conversations from the used dataset in perplexity and MaUde scores.
  • Thumbnail Image
    Text Mining-based Social-Psychological Vulnerability Analysis of Potential Victims To Cybergrooming: Insights and Lessons Learned
    Guo, Zhen; Wang, Pei; Cho, Jin-Hee; Huang, Lifu (ACM, 2023-05)
    Cybergrooming is a serious cybercrime that primarily targets youths through online platforms. Although reactive predator detection methods have been studied, proactive victim protection and crime prevention can also be achieved through vulnerability analysis of potential youth victims. Despite its significance, vulnerability analysis has not been thoroughly studied in the data science literature, while several social science studies used survey-based methods. To address this gap, we investigate humans’ social-psychological traits and quantify key vulnerability factors to cybergrooming by analyzing text features in the Linguistic Inquiry and Word Count (LIWC). Through pairwise correlation studies, we demonstrate the degrees of key vulnerability dimensions to cybergrooming from youths’ conversational features. Our findings reveal that victims have negative correlations with family and community traits, contrasting with previous social survey studies that indicated family relationships or social support as key vulnerability factors. We discuss the current limitations of text mining analysis and suggest cross-validation methods to increase the validity of research findings. Overall, this study provides valuable insights into understanding the vulnerability factors to cybergrooming and highlights the importance of adopting multidisciplinary approaches.