Lower and upper bounds for numbers of linear regions of graph convolutional networks.
Hao ChenYu Guang WangHuan XiongPublished in: Neural Networks (2023)
Keyphrases
- lower and upper bounds
- lower bound
- upper bound
- upper and lower bounds
- lagrangian relaxation
- graph structure
- social networks
- directed graph
- graph structures
- random walk
- edge weights
- graph layout
- valid inequalities
- connected components
- image regions
- small world
- deep learning
- fully connected
- graph theory
- branch and bound
- complex networks
- network structure
- input image
- data streams
- community discovery
- data mining