A Spectral Analysis of Graph Neural Networks on Dense and Sparse Graphs.
Luana RuizNingyuan Teresa HuangSoledad VillarPublished in: ICASSP (2024)
Keyphrases
- spectral analysis
- densely connected
- spectral methods
- neural network
- graph laplacian
- quasi cliques
- weighted graph
- directed acyclic
- graph structure
- graph mining
- graph representation
- graph construction
- graph databases
- gaussian graphical models
- graph matching
- graph theory
- labeled graphs
- directed graph
- graph clustering
- graph model
- bipartite graph
- graph structures
- adjacency matrix
- undirected graph
- graph partitioning
- pattern recognition
- dynamic graph
- graph data
- autoregressive
- spectral features
- edge weights
- community structure
- neighborhood graph
- graphical models
- random walk
- filter bank
- community detection
- spanning tree
- structured data
- high dimensional
- pairwise
- connected graphs
- speech signal
- denoising
- feature space