Heterogeneous Graph
Heterogeneous graphs, which consist of different types of nodes and edges, are powerful tools for modeling complex real-world systems with diverse relationships. Current research focuses on developing effective graph neural network (GNN) architectures, including transformers and attention mechanisms, to learn representations from these complex structures, often addressing challenges like heterophily (nodes of the same type having dissimilar neighbors) and over-smoothing. These advancements are improving performance in various applications, such as recommendation systems, fake news detection, and resource allocation in wireless networks, by leveraging the rich relational information encoded within heterogeneous graphs. The development of robust and scalable algorithms for heterogeneous graph learning is a significant area of ongoing research, with a focus on improving efficiency and accuracy for large-scale datasets.
Papers
Heterogeneous Graph Neural Networks using Self-supervised Reciprocally Contrastive Learning
Cuiying Huo, Dongxiao He, Yawen Li, Di Jin, Jianwu Dang, Weixiong Zhang, Witold Pedrycz, Lingfei Wu
HDGT: Heterogeneous Driving Graph Transformer for Multi-Agent Trajectory Prediction via Scene Encoding
Xiaosong Jia, Penghao Wu, Li Chen, Yu Liu, Hongyang Li, Junchi Yan