On Provable Benefits of Depth in Training Graph Convolutional Networks.
Weilin CongMorteza RamezaniMehrdad MahdaviPublished in: CoRR (2021)
Keyphrases
- average degree
- graph representation
- graph theory
- small world
- network structure
- social networks
- restricted boltzmann machine
- fully connected
- graph mining
- graph structure
- complex networks
- random walk
- community structure
- highly connected
- recurrent networks
- biological networks
- edge weights
- training examples
- graph structures
- network size
- graph matching
- graph layout
- neural network
- betweenness centrality
- training set
- community discovery
- dynamic networks
- structured data
- graph partitioning
- degree distribution
- overlapping communities
- connected components
- directed graph
- bipartite graph
- depth information
- directed edges