Federated Hypergraph Learning with Hyperedge Completion

Hypergraph neural networks enhance conventional graph neural networks by capturing high-order relationships among nodes, which proves vital in data-rich environments where interactions are not merely pairwise. As data complexity and interconnectivity grow, it is common for graph-structured data to be split and stored in a distributed manner, underscoring the necessity of federated learning on subgraphs. In this work, we propose FedHGN, a novel algorithm for federated hypergraph learning. Our algorithm utilizes subgraphs of a hypergraph stored on distributed devices to train local HGNN models in a federated manner:by collaboratively developing an effective global HGNN model through sharing model parameters while preserving client privacy. Additionally, considering that hyperedges may span multiple clients, a pre-training step is employed before the training process in which cross-client hyperedge feature gathering is performed at the central server. In this way, the missing cross-client information can be supplemented from the central server during the node feature aggregation phase. Experimental results on seven real-world datasets confirm the effectiveness of our approach and demonstrate its performance advantages over traditional federated graph learning methods.