Mapping for Disasters with AI

Red Cross volunteers Dale Kunce and Maryjane Rose have seen first-hand the devastation of natural disasters like Hurricane Maria. Having accurate maps of affected regions can mean life or death in evacuating people to safety, and getting critical supplies to those in need. Yet many of the most vulnerable regions in the world have no accurate maps at all, and volunteers require weeks or even months to manually get updated maps in place. Intel AI Lab, MILA, and CrowdAI found a way to cut that time down significantly using a deep learning image segmentation model.