HGC-Herd: Efficient Heterogeneous Graph Condensation via Representative Node Herding

This article introduces a method called HGC-Herd for efficiently condensing heterogeneous graphs. The core idea is to select representative nodes to reduce the graph's complexity. The use of 'herding' suggests an iterative process of selecting nodes that best represent the overall graph structure. The focus on heterogeneous graphs indicates the method's applicability to complex data with different node and edge types. The efficiency claim suggests a focus on computational cost reduction.