Active Learning for Graph Neural Networks via Node Feature Propagation.
Yuexin WuYichong XuAarti SinghYiming YangArtur DubrawskiPublished in: CoRR (2019)
Keyphrases
- active learning
- neural network
- graph structure
- directed graph
- undirected graph
- edge weights
- pattern recognition
- multilayer perceptron
- random walk
- nodes of a graph
- graph theory
- back propagation
- fuzzy logic
- structured data
- path length
- preprocessing stage
- selective sampling
- feed forward
- overlapping communities
- graph representation
- learning algorithm
- graph clustering
- weighted graph
- training set
- artificial neural networks
- bipartite graph
- neural network model
- image features
- connected components
- tree structure
- point distribution
- finding the shortest path
- semi supervised
- data sets
- experimental design
- decision trees
- graph mining
- directed acyclic graph
- random sampling
- feature vectors
- self organizing maps
- learning strategies
- semi supervised learning