Reinforcement Learning Dynamics in the Infinite Memory Limit.
Wolfram BarfussPublished in: AAMAS (2020)
Keyphrases
- reinforcement learning
- function approximation
- dynamical systems
- state space
- reinforcement learning algorithms
- memory usage
- model free
- markov decision processes
- dynamic model
- robotic control
- optimal control
- main memory
- optimal policy
- memory requirements
- data sets
- learning process
- artificial neural networks
- multi agent
- computing power
- collective behavior
- low memory
- dynamic programming
- partially observable
- learning agent
- temporal difference learning
- policy search
- learning algorithm