Adversarial Weight Perturbation Improves Generalization in Graph Neural Network.
Yihan WuAleksandar BojchevskiHeng HuangPublished in: CoRR (2022)
Keyphrases
- neural network
- edge weights
- graph theory
- weight update
- artificial neural networks
- graph representation
- weighted graph
- random walk
- graph structure
- pattern recognition
- fuzzy logic
- connected components
- neural nets
- neural network model
- graph mining
- multi layer
- training algorithm
- bipartite graph
- neural network is trained
- directed graph
- back propagation
- genetic algorithm
- bp neural network
- graph matching
- feed forward
- fuzzy neural network
- graph partitioning
- activation function
- knn
- support vector
- weight vector
- graph clustering
- dependency graph
- graph based algorithm
- image segmentation