StARformer: Transformer with State-Action-Reward Representations for Visual Reinforcement Learning.
Jinghuan ShangKumara KahatapitiyaXiang LiMichael S. RyooPublished in: ECCV (39) (2022)
Keyphrases
- state action
- reinforcement learning
- average reward
- evaluation function
- function approximators
- continuous state
- markov decision process
- reward function
- policy gradient
- action space
- function approximation
- reinforcement learning algorithms
- markov decision processes
- state transitions
- state space
- multi agent
- learning algorithm
- optimal policy
- model free
- stochastic games
- machine learning
- temporal difference
- dynamic programming
- belief state
- learning automata
- neural network
- partially observable
- action selection
- long run
- belief revision
- semi supervised
- learning process