Balancing policy constraint and ensemble size in uncertainty-based offline reinforcement learning.
Alex BeesonGiovanni MontanaPublished in: CoRR (2023)
Keyphrases
- reinforcement learning
- optimal policy
- sequential decision making
- policy search
- function approximation
- markov decision process
- neural network
- partial observability
- function approximators
- markov decision processes
- markov decision problems
- learning algorithm
- action selection
- partially observable environments
- partially observable
- dynamic programming
- training set
- state space
- reinforcement learning problems
- real time
- policy gradient
- rl algorithms
- reinforcement learning algorithms
- continuous state
- multi agent
- policy evaluation
- approximate dynamic programming
- actor critic
- transition model
- temporal difference
- partially observable domains