TRS Research Seminar by Dr. Zeinab Noorian

Title: User-centric Modeling of Online Hate through the lens of Psycholinguistic Patterns and Behaviors in Social Media

Description: Hate speech in social media is a growing problem that reinforces racial discrimination and mistrust between people, leading to physical crimes, violence, and fragmentation in the world communities. Although previous studies showed the potential of user profiling in hate speech detection in social media, there has not been a thorough analysis on users’ characteristics and dispositions to understand the development of hate attitude among social media users. In an attempt to bridge this gap, we introduce generalizable methodology to hate speech detection that investigates a role of a wide range of psycholinguistic characteristics and behavioral traits in characterizing and distinguishing users prone to post hate speech in social media– including word usage pattern, polarization, emotional expression, topical interests, personality traits, readability, communication style, social engagement, posting trends, and information quality analysis. Through extensive observational studies on a large-scale dataset, our findings reveal significant statistical differences on most dimensions of psycholinguistic attributes and online activities of hateful-to-be users. We further develop a classifier and demonstrate that features derived from user timelines are strong indicators for automatically predicting the onset of hateful behavior with, motivating social media policy makers to further investigate the effectiveness of user characterization for preventive interventions of hate speech spread in social media.

Bio: Zeinab Noorian is an assistant professor in the Information Technology Management Department. Her research spans information retrieval, natural language processing and predictive analytics for big data. She has a strong interest in computational social science questions and social media analysis, specially to understand the causal impact of different phenomena on mental health, behavior and decision making process of people reflected in online communities.