Fast Probabilistic Policy Reuse via Reward Function Fitting.
Jinmei LiuZhi WangChunlin ChenPublished in: IJCNN (2022)
Keyphrases
- reward function
- inverse reinforcement learning
- optimal policy
- markov decision processes
- generative model
- state space
- reinforcement learning
- reinforcement learning algorithms
- partially observable
- markov decision process
- policy search
- probabilistic model
- multiple agents
- control policies
- transition probabilities
- hierarchical reinforcement learning
- state variables
- bayesian networks
- initially unknown
- decision problems
- average reward
- state action
- markov decision problems
- machine learning
- maximum likelihood
- dynamic programming
- optimal solution