Policy Gradients for Probabilistic Constrained Reinforcement Learning.
Weiqin ChenDharmashankar SubramanianSantiago PaternainPublished in: CISS (2023)
Keyphrases
- reinforcement learning
- optimal policy
- action selection
- policy search
- markov decision process
- partially observable
- probabilistic model
- function approximation
- reinforcement learning algorithms
- policy iteration
- markov decision processes
- function approximators
- bayesian networks
- reinforcement learning problems
- approximate dynamic programming
- state space
- policy gradient
- state action
- actor critic
- model free
- decision problems
- partially observable environments
- probabilistic logic
- multi agent
- markov decision problems
- learning algorithm
- control policy
- model free reinforcement learning
- reward function
- infinite horizon
- long run
- dynamic programming
- sufficient conditions
- generative model
- learning problems
- conditional probabilities
- machine learning
- probability theory
- partially observable domains
- evaluation function
- uncertain data
- temporal difference learning
- belief networks
- stochastic games
- action space
- temporal difference