Lower and Upper Bounds for Numbers of Linear Regions of Graph Convolutional Networks.
Hao ChenYu Guang WangHuan XiongPublished in: CoRR (2022)
Keyphrases
- lower and upper bounds
- lower bound
- upper bound
- upper and lower bounds
- fully connected
- graph structure
- lagrangian relaxation
- graph theory
- graph structures
- small world
- random walk
- edge weights
- linear constraints
- image regions
- sparse coding
- community discovery
- graph layout
- input image
- valid inequalities
- normalized cut
- power law
- social networks
- search space
- bipartite graph