As a privacy-preserving collaborative paradigm, federated graph learning (FGL) enables distributed training of graph neural networks (GNNs) without exposing raw graph data. Subgraph-FL has become the dominant FGL paradigm, yet most studies focus on overall nod... ...
