AI-Driven Early Warning for Epidemic Risk Using Satellite, Mobility and Social Data

Abstract

The research develops an AI-based forecast model for epidemic outbreaks with multi-source data integration. Accurate epidemic risk prediction combines satellite imagery, human mobility and social media data. The risk clusters are effectively detected by a hybrid deep learning model that employs CNNs and GNNs. The model helps detect malaria, dengue and influenza outbreaks early anywhere in the world. Actionable insights on resource-efficient epidemic response are useful to government agencies. Faster directed responses in turn benefit public health preparedness. It is important that ethical data use is used for responsible AI deployment. It provides future research to expand to more data sources and increase the scalability.