RL-CycleGAN: Reinforcement Learning Aware Simulation-to-Real.
Kanishka RaoChris HarrisAlex IrpanSergey LevineJulian IbarzMohi KhansariPublished in: CVPR (2020)
Keyphrases
- reinforcement learning
- real robot
- function approximation
- state space
- temporal difference
- robocup soccer
- optimal policy
- reinforcement learning algorithms
- partially observable domains
- temporal difference learning
- control problems
- simulation data
- rl algorithms
- optimal control
- simulation model
- markov decision processes
- dynamic programming
- multi agent
- direct policy search
- actor critic
- multi agent reinforcement learning
- autonomous learning
- action selection
- real world
- adaptive control
- model free
- transfer learning
- learning algorithm
- real environment
- reinforcement learning methods
- continuous state
- learning classifier systems
- supervised learning