Structural-aware motif-based prompt tuning for graph clustering

Mingchen Sun, Mengduo Yang, Yingji Li, Dongmei Mu, Xin Wang, Ying Wang

Abstract: Highlights•We present structural-aware motif-based prompt tuning framework for efficient clustering.•We explore graph motif as structure prompt to augment original graph structure.•We design a task token prompt to narrow the gap between pretrain and downstream tasks.•We validate the effectiveness of our model on six real-world datasets.