Policy Interpretation for Partially Observable Monte-Carlo Planning: a Rule-based Approach.
Giulio MazziAlberto CastelliniAlessandro FarinelliPublished in: AIRO@AI*IA (2020)
Keyphrases
- monte carlo
- partially observable
- state space
- markov decision problems
- reinforcement learning
- markov chain
- decision problems
- markov decision processes
- partially observable environments
- infinite horizon
- dynamical systems
- partial observability
- planning domains
- partially observable domains
- belief space
- belief state
- reward function
- partially observable markov decision process
- optimal policy
- monte carlo simulation
- monte carlo tree search
- partially observable markov decision processes
- particle filter
- fully observable
- long run
- planning problems
- optimal strategy
- game tree
- ai planning
- machine learning
- inverse reinforcement learning
- decision theoretic
- learning algorithm