Extreme Q-Learning: MaxEnt RL without Entropy.
Divyansh GargJoey HejnaMatthieu GeistStefano ErmonPublished in: CoRR (2023)
Keyphrases
- reinforcement learning
- function approximation
- reinforcement learning algorithms
- model free
- state space
- maximum entropy
- multi agent
- optimal policy
- action selection
- rl algorithms
- learning algorithm
- mutual information
- temporal difference
- temporal difference learning
- information theory
- reinforcement learning methods
- cooperative
- markov decision processes
- information theoretic
- state action
- information entropy
- policy iteration
- actor critic
- dynamic programming
- multi agent reinforcement learning
- sequential decision problems
- temporal difference methods
- td learning
- exploration strategy
- action space
- machine learning
- optimal control
- learning problems
- continuous state
- continuous state spaces
- conditional random fields
- hierarchical reinforcement learning
- continuous state and action spaces