POPGym: Benchmarking Partially Observable Reinforcement Learning.
Steven D. MoradRyan KortvelesyMatteo BettiniStephan LiwickiAmanda ProrokPublished in: CoRR (2023)
Keyphrases
- partially observable
- reinforcement learning
- markov decision processes
- state space
- partial observability
- partially observable domains
- dynamical systems
- decision problems
- function approximation
- markov decision problems
- partially observable environments
- optimal policy
- hidden state
- partial observations
- reinforcement learning algorithms
- belief space
- temporal difference
- reward function
- action models
- belief state
- multi agent
- learning algorithm
- action space
- model free
- infinite horizon
- transfer learning
- learning capabilities
- partially observable markov decision processes
- markov decision process
- action selection
- orders of magnitude
- initially unknown
- sufficient conditions
- bayesian networks