Refining Latent Homophilic Structures over Heterophilic Graphs for Robust Graph Convolution Networks.
Chenyang QiuGuoshun NanTianyu XiongWendi DengDi WangZhiyang TengLijuan SunQimei CuiXiaofeng TaoPublished in: CoRR (2023)
Keyphrases
- graph structures
- graph structure
- directed graph
- highly connected
- community discovery
- average degree
- graph mining
- graph representation
- graph theory
- edge weights
- weighted graph
- graph clustering
- structured data
- graph databases
- graph partitioning
- graph matching
- graph theoretic
- small world
- graph model
- graph layout
- real world graphs
- fully connected
- adjacency matrix
- graph theoretical
- graph classification
- random graphs
- series parallel
- random walk
- spanning tree
- social graphs
- graph construction
- graph search
- bipartite graph
- complex structures
- social networks
- labeled graphs
- complex networks
- latent variables
- network analysis
- graph isomorphism
- structural pattern recognition
- dynamic networks
- dense subgraphs
- massive graphs
- graph data
- degree distribution
- betweenness centrality
- graph mining algorithms
- structural patterns
- planar graphs
- subgraph isomorphism
- protein interaction networks
- relational structures
- graph kernels
- real world networks
- directed acyclic graph
- undirected graph
- social network analysis
- introduce a general framework
- network structure
- community detection
- dynamic graph
- power law
- graph patterns
- graph properties
- graph representations