Balancing policy constraint and ensemble size in uncertainty-based offline reinforcement learning.
Alex BeesonGiovanni MontanaPublished in: Mach. Learn. (2024)
Keyphrases
- reinforcement learning
- optimal policy
- policy search
- sequential decision making
- action selection
- markov decision process
- machine learning
- learning algorithm
- reward function
- function approximators
- state space
- function approximation
- partially observable
- real time
- markov decision processes
- partially observable environments
- control policies
- ensemble methods
- partial observability
- reinforcement learning algorithms
- policy gradient
- actor critic
- random forest
- decision problems
- dynamic programming
- learning process
- reinforcement learning problems
- neural network
- partially observable domains
- state action
- policy iteration
- temporal difference
- random forests
- base classifiers
- training data
- decision making
- feature selection