AI RESEARCH
Beyond the Laplacian: Doubly Stochastic Matrices for Graph Neural Networks
arXiv CS.LG
•
ArXi:2604.15069v1 Announce Type: new Graph Neural Networks (GNNs) conventionally rely on standard Laplacian or adjacency matrices for structural message passing. In this work, we substitute the traditional Laplacian with a Doubly Stochastic graph Matrix (DSM), derived from the inverse of the modified Laplacian, to naturally encode continuous multi-hop proximity and strict local centrality.