Manifold Embeddings for Model-Based Reinforcement Learning under Partial Observability.
Keith BushJoelle PineauPublished in: NIPS (2009)
Keyphrases
- partial observability
- model based reinforcement learning
- markov decision processes
- partially observable
- reinforcement learning
- partially observable markov decision processes
- markov decision problems
- markov decision process
- state space
- planning problems
- finite state
- belief state
- belief space
- dynamical systems
- optimal policy
- function approximation
- planning under uncertainty
- decision problems
- multi agent
- infinite horizon
- average cost
- domain independent
- learning agent
- heuristic search
- domain knowledge
- single agent
- supervised learning
- machine learning