Learned Belief Search: Efficiently Improving Policies in Partially Observable Settings.
Hengyuan HuAdam LererNoam BrownJakob N. FoersterPublished in: CoRR (2021)
Keyphrases
- partially observable
- belief state
- decision problems
- search algorithm
- markov decision problems
- state space
- belief space
- markov decision processes
- search space
- dynamical systems
- search strategy
- partially observable markov decision processes
- reward function
- search strategies
- partial observability
- reinforcement learning
- partial observations
- partially observable domains
- domain specific
- computational complexity
- partially observable environments