Exchangeable Input Representations for Reinforcement Learning.
John MernDorsa SadighMykel J. KochenderferPublished in: CoRR (2020)
Keyphrases
- reinforcement learning
- learning algorithm
- function approximation
- markov decision processes
- machine learning
- multi agent
- autonomous learning
- multiple representations
- robot control
- action selection
- higher level
- state space
- input data
- optimal policy
- supervised learning
- learning classifier systems
- symbolic representation
- model free
- artificial neural networks
- partially observable
- decision making
- information systems
- real time