Learning Synthetic Environments and Reward Networks for Reinforcement Learning.
Fabio FerreiraThomas NierhoffAndreas SälingerFrank HutterPublished in: ICLR (2022)
Keyphrases
- model free
- reinforcement learning
- function approximation
- learning algorithm
- reinforcement learning methods
- eligibility traces
- supervised learning
- transfer learning
- machine learning
- learning process
- temporal difference learning
- multi agent
- real world
- connectionist networks
- learning agent
- dynamic environments
- learning systems
- state action
- mobile robot
- prior knowledge
- learning environment
- policy gradient
- policy search
- multi agent environments
- neural network
- partially observable environments