Landmark Based Reward Shaping in Reinforcement Learning with Hidden States.
Alper DemirErkin ÇildenFaruk PolatPublished in: AAMAS (2019)
Keyphrases
- markov decision process
- reward shaping
- reinforcement learning
- hidden states
- hidden state
- hidden markov models
- state space
- optimal policy
- reinforcement learning algorithms
- function approximation
- conditional random fields
- reward function
- markov decision problems
- hidden variables
- exponential family
- model free
- optimal control
- multi agent
- supervised learning
- partially observable
- learning algorithm
- generative model
- temporal difference
- transfer learning
- missing values
- action space
- higher order
- action selection
- dynamic bayesian networks
- policy search
- bayesian networks