Abstract: Graph Neural Networks (GNNs) have been gaining more attention due to their excellent performance in modeling various graph-structured data. However, most of the current GNNs only consider ...

Diffusion processes, characterised through operators such as the graph Laplacian, foster the development of diffusion maps which yield robust, multi-scale representations.

First, a dynamic graph structure is created using a weight-sharing GCN feature encoder, transforming discrete changes into topological relationships between nodes. Then, the powerful feature ...