Towards Interpretable Reinforcement Learning with Constrained Normalizing Flow Policies.
Finn RietzErik SchaffernichtStefan HeinrichJohannes A. StorkPublished in: CoRR (2024)
Keyphrases
- reinforcement learning
- optimal policy
- policy search
- markov decision process
- control policies
- function approximation
- state space
- partially observable markov decision processes
- policy gradient methods
- markov decision processes
- cooperative multi agent systems
- flow patterns
- reinforcement learning algorithms
- model free
- decision problems
- reinforcement learning agents
- hierarchical reinforcement learning
- markov decision problems
- reward function
- fitted q iteration
- learning algorithm
- dynamic programming
- continuous state
- machine learning
- control policy
- function approximators
- policy iteration
- learning problems
- finite state
- sufficient conditions
- multiagent reinforcement learning
- approximate policy iteration
- optimal control
- temporal difference