Abstract: The node classification in graphs aims to predict the categories of unlabeled nodes utilizing a small set of labeled nodes. However, weighted graphs often contain noisy edges and anomalous ...
Abstract: As an important component of intelligent transportation systems, vehicle-to-everything (V2X) provides critical communication support for road safety through real-time interactive and ...
We aim to build a pre-trained Graph Neural Network (GNN) model on molecules without human annotations or prior knowledge. Although various attempts have been proposed to overcome limitations in ...