Expressiveness and Approximation Properties of Graph Neural Networks.
Floris GeertsJuan L. ReutterPublished in: ICLR (2022)
Keyphrases
- neural network
- graph properties
- pattern recognition
- random walk
- back propagation
- genetic algorithm
- artificial neural networks
- fuzzy logic
- structural properties
- graph representation
- desirable properties
- graph structure
- neural nets
- expressive power
- structured data
- special case
- error bounds
- rough sets
- fault diagnosis
- training process
- graph model
- graph partitioning
- markov random field