LanczosNet: Multi-Scale Deep Graph Convolutional Networks.
Renjie LiaoZhizhen ZhaoRaquel UrtasunRichard S. ZemelPublished in: ICLR (Poster) (2019)
Keyphrases
- multiscale
- deep learning
- fully connected
- average degree
- graph representation
- graph theoretic
- small world
- graph theory
- highly connected
- scale space
- social networks
- random walk
- degree distribution
- complex networks
- community discovery
- graph structures
- spanning tree
- graph databases
- edge weights
- betweenness centrality
- graph layout
- directed edges
- graph model
- directed acyclic graph
- natural images
- graph matching
- coarse to fine
- structured data
- network structure
- directed graph
- learning algorithm
- wavelet transform
- image representation
- social graphs
- overlapping communities
- citation networks
- maximum flow
- neural network
- phase transition
- path length
- graph structure
- weighted graph
- graph partitioning