StructComp: Substituting propagation with Structural Compression in Training Graph Contrastive Learning

arXiv:2312.04865v3 Announce Type: replace
Abstract: Graph contrastive learning (GCL) has become a powerful tool for learning graph data, but its scalability remains a significant challenge. In this work, we propose a simple yet effective training framework called Structural Compression (StructComp) to address this issue. Inspired by a sparse low-rank approximation on the diffusion matrix, StructComp trains the encoder with the compressed nodes. This allows the encoder not to perform any message passing during the training stage, and significantly reduces the …

abstract arxiv become challenge compression cs.lg data framework graph graph data issue low propagation scalability simple tool training type work