RL-CycleGAN: Reinforcement Learning Aware Simulation-To-Real.
Kanishka RaoChris HarrisAlex IrpanSergey LevineJulian IbarzMohi KhansariPublished in: CoRR (2020)
Keyphrases
- reinforcement learning
- real robot
- function approximation
- state space
- reinforcement learning algorithms
- model free
- simulation data
- temporal difference
- control problems
- markov decision processes
- rl algorithms
- learning algorithm
- learning problems
- robocup soccer
- action space
- transfer learning
- optimal policy
- autonomous learning
- supervised learning
- learning agents
- simulation model
- function approximators
- reinforcement learning methods
- learning process
- mathematical model
- real environment
- multi agent
- control system
- real world
- monte carlo
- approximate dynamic programming
- multi agent reinforcement learning
- neural network