Abstract: We propose a general attack framework based on evolutionary algorithms to quickly and efficiently generate low-perturbation adversarial samples for 3D point cloud data. Specifically, we ...

Abstract: Real-time precipitation retrieval is crucial for timely warnings of extreme weather events such as heavy rain or floods. Geostationary satellite observations combined with machine learning ...