Batch Active Learning with Graph Neural Networks via Multi-Agent Deep Reinforcement Learning.
Yuheng ZhangHanghang TongYinglong XiaYan ZhuYuejie ChiLei YingPublished in: AAAI (2022)
Keyphrases
- reinforcement learning
- multi agent
- active learning
- neural network
- batch mode
- pattern recognition
- function approximation
- transfer learning
- learning algorithm
- fuzzy logic
- random walk
- exploration exploitation
- semi supervised
- supervised learning
- function approximators
- learning agents
- multi agent environments
- genetic algorithm
- weighted graph
- model free
- single agent
- control policy
- multiagent systems
- learning classifier systems
- graph theory
- graph representation
- graph model
- incremental learning
- graph structure
- learning capabilities
- state space
- artificial neural networks
- cooperative
- reinforcement learning algorithms
- multi agent reinforcement learning
- reinforcement learning agents
- graph matching
- random sampling
- directed acyclic graph
- machine learning
- neural network model
- back propagation
- active exploration
- active learning strategies
- training set
- learning process
- labeled data
- selective sampling
- learning strategies
- autonomous agents
- markov decision processes
- directed graph
- learning problems
- temporal difference
- action selection
- data sets