Improving exploration efficiency of deep reinforcement learning through samples produced by generative model.
Dayong XuFei ZhuQuan LiuPeiyao ZhaoPublished in: Expert Syst. Appl. (2021)
Keyphrases
- generative model
- reinforcement learning
- probabilistic model
- mixture model
- discriminative learning
- function approximation
- prior knowledge
- bayesian framework
- posterior probability
- em algorithm
- temporal difference
- discriminative models
- data sets
- learning process
- topic models
- expectation maximization
- markov decision processes
- active exploration
- training set
- pitman yor process
- action selection
- latent dirichlet allocation
- transfer learning
- markov chain
- learning algorithm