Community Perspective

Understanding community perspectives is crucial for mitigating harmful online content and biases in automated systems. Current research focuses on leveraging large language models, graph transformer networks, and community-specific classifiers to analyze online discussions and identify harmful language, biases, and differing worldviews within various online communities. This work aims to improve the accuracy and fairness of automated systems by incorporating contextual information and community norms, leading to more nuanced and equitable applications in areas such as hate speech detection, abusive language classification, and peer review processes.