Towards A Unified Policy Abstraction Theory and Representation Learning Approach in Markov Decision Processes.
Min ZhangHongyao TangJianye HaoYan ZhengPublished in: CoRR (2022)
Keyphrases
- markov decision processes
- decision theoretic planning
- optimal policy
- partially observable
- reinforcement learning
- state abstraction
- state space
- model based reinforcement learning
- infinite horizon
- policy iteration
- transition matrices
- macro actions
- decision processes
- markov decision process
- reward function
- finite state
- learning tasks
- decision problems
- supervised learning
- learning algorithm
- stochastic games
- markov games
- finite horizon
- action selection
- planning under uncertainty
- actor critic
- factored mdps