Semi-Supervised Node Classification on Graphs: Markov Random Fields vs. Graph Neural Networks.
Binghui WangJinyuan JiaNeil Zhenqiang GongPublished in: AAAI (2021)
Keyphrases
- markov random field
- graph structure
- semi supervised
- directed graph
- neural network
- graph construction
- semi supervised classification
- graphical structure
- pattern recognition
- belief propagation
- finding the shortest path
- discriminative random fields
- pairwise
- weighted graph
- energy minimization
- random fields
- graphical models
- graph cuts
- higher order
- undirected graph
- image restoration
- min cut
- image segmentation
- potential functions
- semi supervised learning
- parameter estimation
- supervised learning
- energy function
- mrf model
- low level vision
- markov networks
- conditional random fields
- text classification
- map inference
- random graphs
- higher order cliques
- graph model
- model selection
- information extraction
- maximum a posteriori
- loopy belief propagation
- knn
- bayesian networks
- support vector
- potts model
- object recognition
- active learning
- hidden markov models
- probabilistic model
- edge weights
- maximum likelihood
- spanning tree
- labeled data
- unlabeled data