Planning with Exploration: Addressing Dynamics Bottleneck in Model-based Reinforcement Learning.
Xiyao WangJunge ZhangWenzhen HuangQiyue YinPublished in: CoRR (2020)
Keyphrases
- model based reinforcement learning
- markov decision processes
- dynamical systems
- markov decision problems
- partially observable markov decision processes
- reinforcement learning
- partially observable
- relational domains
- planning problems
- finite state
- multi relational
- decision theoretic
- state space
- dynamic programming
- action selection
- optimal policy
- blocks world
- probabilistic model
- relational databases