GraphCBAL: Class-Balanced Active Learning for Graph Neural Networks via Reinforcement Learning.
Chengcheng YuJiapeng ZhuXiang LiPublished in: CoRR (2024)
Keyphrases
- active learning
- neural network
- reinforcement learning
- fuzzy logic
- artificial neural networks
- random walk
- imbalanced data classification
- genetic algorithm
- function approximators
- bipartite graph
- function approximation
- learning strategies
- supervised learning
- back propagation
- learning process
- pattern recognition
- graph theory
- machine learning
- temporal difference
- graph representation
- multi agent
- multilayer perceptron
- transfer learning
- graph matching
- state space
- graph model
- experimental design
- dynamic programming
- learning capabilities
- learning algorithm
- selective sampling
- rare class
- training examples