Compositional Conservatism: A Transductive Approach in Offline Reinforcement Learning.
Yeda SongDongwook LeeGunhee KimPublished in: CoRR (2024)
Keyphrases
- reinforcement learning
- function approximation
- real time
- temporal difference
- learning algorithm
- reinforcement learning algorithms
- optimal policy
- dynamic programming
- transductive learning
- model free
- database
- semi supervised learning
- semi supervised
- support vector
- supervised learning
- state space
- learning process
- markov decision processes
- kernel matrix
- multi agent
- robotic control
- learning capabilities
- markov decision process
- reinforcement learning methods
- autonomous learning
- action selection
- optimal control
- learning problems
- transfer learning
- text classification
- machine learning
- neural network