SafeScreen: Detecting Early Signs of Domestic Abuse on Social Media Using NLP

Our speaker, Paule Marcelle Elanga, holds a Master’s degree in Artificial Intelligence with a specialization in Data Science. She is passionate about applied AI, particularly in projects that drive positive social impact. She will be presenting:
SafeScreen: Detecting Early Signs of Domestic Abuse on Social Media Using NLP and Psychological Features.
Summary: SafeScreen is an AI-based system designed to detect early signs of domestic abuse through the analysis of social media posts. The model combines transformer-based language models (BERT and DistilBERT) with handcrafted psychological features derived from trauma psychology. This hybrid approach enables the early identification of abuse-expressing and abuse-normalizing language. The research highlights the potential of AI for social good, particularly in providing early intervention and support for vulnerable individuals.