Rectifying Reinforcement Learning for Reward Matching.
Haoran HeEmmanuel BengioQingpeng CaiLing PanPublished in: CoRR (2024)
Keyphrases
- reinforcement learning
- reinforcement learning algorithms
- matching algorithm
- function approximation
- state space
- markov decision processes
- pattern matching
- matching scheme
- reward function
- model free
- multi agent
- graph matching
- optimal control
- matching process
- learning algorithm
- partially observable environments
- markov decision process
- partially observable
- reinforcement learning methods
- eligibility traces
- robotic control
- temporal difference
- transfer learning
- keypoints
- optimal policy
- machine learning
- action selection
- learning problems
- single agent
- policy iteration
- function approximators
- average reward
- feature points
- learning process
- neural network