PhD Student · Information Science
University of Pittsburgh,
School of Computing and Information
Studying how extremist communities evolve online, gender gaps in hateful content, and AI-driven interventions against misinformation — combining computational methods with social theory to build a safer information environment.
Extremist communities increasingly rely on social media to sustain and amplify divisive discourse. This study analyzes ten years of Facebook activity by hate groups related to the Israel–Palestine conflict, focusing on anti-Semitic and Islamophobic ideologies. We find that higher participation centralization is associated with greater user engagement, suggesting key actors sustain group activity over time. Our narrative frame detection models reveal centralized Islamophobic groups use more uniform messaging, while centralized anti-Semitic groups show greater framing diversity — providing a foundation for tailored counter-strategies.
Not all online discussions about conspiracy theories promote them — some aim to debunk them. This work establishes a classification scheme based on authors' perspectives, expressed through narrative elements or references to known theories. We train a BERT-based model and compare it to GPT, finding significant flaws in GPT's logical reasoning. The first large-scale classification of active conspiracy-related Reddit forums finds that only one-third of posts are genuinely conspiratorial — illuminating how LLMs handle nuanced contextual comprehension tasks.
I am a PhD student in Information Science at the University of Pittsburgh, studying how extremist online communities evolve, gender gaps in the circulation of hateful content online, and how AI-driven interventions can counter health misinformation. My research combines computational methods — network analysis, NLP, and large language models — with social theory to understand what makes harmful content persuasive. As an Indonesian woman, first-generation graduate student, and mother, I bring a perspective that is often absent from tech research.