Robust Training of Graph Neural Networks via Noise Governance.
Siyi QianHaochao YingRenjun HuJingbo ZhouJintai ChenDanny Z. ChenJian WuPublished in: WSDM (2023)
Keyphrases
- neural network
- training algorithm
- training process
- image noise
- feedforward neural networks
- neural network training
- information technology
- geometric distortions
- fuzzy logic
- graph representation
- salt pepper
- graph theory
- multi layer perceptron
- noisy environments
- fault diagnosis
- signal to noise ratio
- artificial neural networks
- information systems
- graph structure
- feed forward neural networks
- graph model
- recurrent networks
- estimation error
- noise level
- neural nets
- recurrent neural networks
- neural network model
- missing data
- genetic algorithm
- digital government
- pattern recognition
- training set
- active learning
- radial basis function network
- registration errors
- training samples
- training examples
- back propagation
- connected components
- directed graph
- graph mining
- noise reduction