ASGN: An Active Semi-supervised Graph Neural Network for Molecular Property Prediction.
Zhongkai HaoChengqiang LuZheyuan HuHao WangZhenya HuangQi LiuEnhong ChenCheekong LeePublished in: CoRR (2020)
Keyphrases
- semi supervised
- neural network
- prediction model
- protein function prediction
- label propagation
- semi supervised classification
- graph construction
- prediction accuracy
- artificial neural networks
- graph properties
- labeled data
- semi supervised learning
- multi view
- three dimensional
- back propagation
- graph representation
- neural network model
- graph structure
- bipartite graph
- neural network is trained
- graph model
- genetic algorithm
- global consistency
- graph matching
- supervised learning
- structured data
- elman network
- pattern recognition
- active learning
- drug design
- fuzzy logic
- random walk
- unlabeled data
- multi layer perceptron
- co training
- spanning tree
- directed acyclic graph
- prediction error
- recurrent neural networks
- graph theory
- learning algorithm
- protein structure