Offline Reinforcement Learning as Anti-Exploration.
Shideh RezaeifarRobert DadashiNino VieillardLéonard HussenotOlivier BachemOlivier PietquinMatthieu GeistPublished in: CoRR (2021)
Keyphrases
- reinforcement learning
- active exploration
- exploration strategy
- action selection
- model based reinforcement learning
- exploration exploitation
- function approximation
- reinforcement learning algorithms
- learning algorithm
- autonomous learning
- real time
- machine learning
- exploration exploitation tradeoff
- model free
- markov decision processes
- active learning
- temporal difference
- robotic control
- state space
- learning capabilities
- balancing exploration and exploitation
- database
- real world
- genetic algorithm
- information systems
- control policy
- objective function
- multi agent
- transfer learning
- optimal policy