Harnessing Discrete Representations For Continual Reinforcement Learning.
Edan MeyerAdam WhiteMarlos C. MachadoPublished in: CoRR (2023)
Keyphrases
- reinforcement learning
- state space
- continuous state
- markov decision processes
- function approximation
- discrete geometry
- learning algorithm
- collective intelligence
- higher level
- continuous state and action spaces
- supervised learning
- real valued
- learning process
- optimal control
- artificial neural networks
- model free
- temporal difference
- representation scheme
- continuous domains
- control problems
- social networks
- machine learning
- robotic control
- real time