Continual State Representation Learning for Reinforcement Learning using Generative Replay.
Hugo Caselles-DupréMichaël Garcia OrtizDavid FilliatPublished in: CoRR (2018)
Keyphrases
- reinforcement learning
- learning process
- supervised learning
- state space
- learning algorithm
- online learning
- learning capabilities
- learning problems
- unsupervised learning
- knowledge acquisition
- learning systems
- active learning
- function approximation
- learning classifier systems
- temporal difference learning
- machine learning
- actor critic
- learning agents
- eligibility traces
- temporal difference
- action selection
- image representation
- dynamic programming