Abstract: In unsupervised graph anomaly detection, existing methods usually focus on detecting outliers by learning local context information of nodes, while often ignoring the importance of global ...

“This is the record that’s been the most cathartic. The title came to me early in the process and became an inescapable idea. It seemed to sum up the way I think a lot of people are feeling,” Marr ...

Abstract: Detecting small objects in drone imagery is challenging due to low resolution and background blending, leading to limited feature information. Multiscale feature fusion can enhance detection ...