Description
Natural disasters and extreme weather events, from hurricanes to earthquakes to wildfires, devastate countless cities, towns, and villages every year. They ravage the infrastructure of affected regions, including homes, power lines, and water supplies, incur immense economic loss, and cause thousands of deaths. In turn, this causes perpetually negative health outcomes. These are challenges faced particularly by the developing world. In this presentation, I will outline my current work in developing interpretable convolutional neural networks using PyTorch to assess building damage from satellite imagery. Further, I will outline previous works in the field to contextualize this research as well as propose future areas of work.