Self-Training Of Graph Neural Networks Using Similarity Reference For Robust Training With Noisy Labels.
Hyoungseob ParkMinki JeongYoungeun KimChangick KimPublished in: ICIP (2020)
Keyphrases
- neural network
- training set
- training process
- training algorithm
- training examples
- label noise
- noisy environments
- training data
- similarity measure
- pattern recognition
- directed graph
- distance measure
- feedforward neural networks
- label propagation
- multi layer
- graph structure
- labelled data
- similarity function
- genetic algorithm
- semi supervised learning
- fuzzy logic
- backpropagation algorithm
- supervised learning
- similarity graph
- labeled graphs
- learning algorithm
- semi supervised classification
- active learning
- similarity matrix
- multi label
- graph databases
- random walk
- back propagation
- structured data
- graph theory
- cost sensitive
- feed forward