Generalisation in Lifelong Reinforcement Learning through Logical Composition.
Geraud Nangue TasseSteven JamesBenjamin RosmanPublished in: ICLR (2022)
Keyphrases
- reinforcement learning
- markov decision processes
- state space
- function approximation
- optimal policy
- model free
- learning activities
- transfer learning
- robotic control
- reinforcement learning methods
- robot control
- learning capabilities
- reinforcement learning algorithms
- dynamic programming
- learning process
- multi agent
- learning algorithm
- learning scenarios
- radial basis function
- supervised learning
- action selection
- lifelong learning
- markov decision process
- action space
- temporal difference learning
- logical reasoning
- music composition
- machine learning