Training A Foundation Model to Represent Graphs as Vectors

Qi Feng, Jicong Fan

Keywords: graph neural network, graph classification

Abstract: This paper aims to train a graph foundation model that is able to represent any graph as a vector preserving structural and semantic information useful for downstream graph-level tasks such as graph classification and graph clustering. To learn the features of graphs from diverse domains while maintaining strong generalization ability to new domains, we propose a multi-graph-based feature alignment method, which constructs weighted graphs using the attributes of all nodes in each dataset and then generates consistent node embeddings. To enhance the consistency of the features from different datasets, we propose a density maximization based mean alignment method. The original graphs and generated node embeddings are fed into a graph neural network to achieve discriminative graph representations in contrastive learning. More importantly, to enhance the information preservation from node-level representations to the graph-level representation, we construct a reference distribution module without using any pooling operation. The experimental results of zero-shot/few-shot graph classification and graph clustering show that our model outperforms strong competitors.

Primary Area: learning on graphs and other geometries & topologies

Submission Number: 12276