StARformer: Transformer With State-Action-Reward Representations for Robot Learning.
Jinghuan ShangXiang LiKumara KahatapitiyaYu-Cheol LeeMichael S. RyooPublished in: IEEE Trans. Pattern Anal. Mach. Intell. (2023)
Keyphrases
- state action
- reinforcement learning
- average reward
- evaluation function
- stochastic games
- action space
- function approximators
- reward function
- optimal policy
- state transitions
- policy gradient
- markov decision process
- markov decision processes
- function approximation
- long run
- belief state
- state space
- search space
- kernel matrix
- learning algorithm
- stochastic processes
- fuzzy logic