Masked Attention is All You Need for Graphs

arXiv:2402.10793v1 Announce Type: new
Abstract: Graph neural networks (GNNs) and variations of the message passing algorithm are the predominant means for learning on graphs, largely due to their flexibility, speed, and satisfactory performance. The design of powerful and general purpose GNNs, however, requires significant research efforts and often relies on handcrafted, carefully-chosen message passing operators. Motivated by this, we propose a remarkably simple alternative for learning on graphs that relies exclusively on attention. Graphs are represented as node or edge …

abstract algorithm arxiv attention attention is all you need cs.ai cs.lg design flexibility general gnns graph graph neural networks graphs networks neural networks performance research speed type