EvenNet: Ignoring Odd-Hop Neighbors Improves Robustness of Graph Neural Networks.
Runlin LeiZhen WangYaliang LiBolin DingZhewei WeiPublished in: NeurIPS (2022)
Keyphrases
- neural network
- random walk
- pattern recognition
- nearest neighbor
- graph theory
- graph representation
- graph model
- back propagation
- directed graph
- structured data
- multi layer
- weighted graph
- graph structure
- multilayer perceptron
- neural nets
- bipartite graph
- graph matching
- connected components
- directed acyclic graph
- graph theoretic
- feed forward
- self organizing maps
- sufficient conditions
- network architecture
- fuzzy logic
- random graphs