Pseudo-MDPs and factored linear action models.
Hengshuai YaoCsaba SzepesváriBernardo Ávila PiresXinhua ZhangPublished in: ADPRL (2014)
Keyphrases
- action models
- partially observable
- state space
- markov decision processes
- factored markov decision processes
- algebraic decision diagrams
- markov decision problems
- goal directed
- ai planning
- optimal policy
- reinforcement learning
- decision problems
- reward function
- dynamical systems
- heuristic search
- factored mdps
- visual features