LGAT: a light graph attention network focusing on message passing for semi-supervised node classification.
Chengcheng SunFanrong MengChenhao LiXiaobin RuiZhixiao WangPublished in: Computing (2024)
Keyphrases
- message passing
- singly connected
- semi supervised
- belief propagation
- junction tree
- semi supervised classification
- distributed systems
- probabilistic reasoning
- belief networks
- probabilistic inference
- supervised learning
- approximate inference
- class labels
- overlapping communities
- machine learning
- shared memory
- inference in graphical models
- graphical models
- computer vision
- sum product
- factor graphs
- distributed shared memory
- sum product algorithm
- directed graph
- text classification
- undirected graph
- active learning
- feature extraction
- bayesian networks
- reinforcement learning
- training set
- pairwise
- model selection
- wireless sensor networks
- probability distribution
- complex networks
- multi view
- multi label
- graph cuts
- semi supervised learning
- labeled data