Conditional Mutual Information for Disentangled Representations in Reinforcement Learning.
Mhairi DunionTrevor McInroeKevin Sebastian LuckJosiah P. HannaStefano V. AlbrechtPublished in: CoRR (2023)
Keyphrases
- conditional mutual information
- reinforcement learning
- mutual information
- selection criterion
- state space
- learning algorithm
- reinforcement learning algorithms
- temporal difference
- model free
- function approximation
- robotic control
- higher level
- symbolic representation
- dynamic programming
- high dimensional
- multiple representations
- learning problems
- image registration
- optimal control
- evolutionary algorithm
- learning process
- function approximators
- policy search
- feature selection
- neural network