Asymmetric Actor Critic for Image-Based Robot Learning.
Lerrel PintoMarcin AndrychowiczPeter WelinderWojciech ZarembaPieter AbbeelPublished in: Robotics: Science and Systems (2018)
Keyphrases
- actor critic
- reinforcement learning
- policy gradient
- approximate dynamic programming
- neuro fuzzy
- temporal difference
- optimal control
- gradient method
- function approximation
- reinforcement learning algorithms
- policy iteration
- average reward
- step size
- markov decision processes
- optimization methods
- variance reduction
- evaluation function
- multi agent
- neural network