Should Graph Neural Networks Use Features, Edges, Or Both?
Lukas FaberYifan LuRoger WattenhoferPublished in: CoRR (2021)
Keyphrases
- neural network
- directed graph
- random walk
- low level
- image features
- artificial neural networks
- directed acyclic graph
- feature extraction
- self organizing maps
- weighted graph
- neural network model
- graph matching
- extracted features
- graph theory
- image structure
- structural information
- connected components
- co occurrence
- pattern recognition