Memoryless Exact Solutions for Deterministic MDPs with Sparse Rewards.
Joshua R. BertramPeng WeiPublished in: CoRR (2018)
Keyphrases
- markov decision processes
- reinforcement learning
- fully observable
- state space
- reward function
- dynamic programming
- optimal policy
- stationary policies
- finite state
- factored mdps
- markov decision problems
- sparse data
- planning under uncertainty
- high dimensional
- average reward
- decision theoretic planning
- reinforcement learning algorithms
- finite horizon
- decision diagrams
- partially observable
- planning problems
- sparse representation
- policy iteration
- discounted reward
- compressive sensing
- infinite horizon
- vector quantizer
- function approximation
- sparse coding
- image classification
- expected reward
- semi markov decision processes
- multiarmed bandit