How to Spend Your Robot Time: Bridging Kickstarting and Offline Reinforcement Learning for Vision-based Robotic Manipulation.
Alex X. LeeColine DevinJost Tobias SpringenbergYuxiang ZhouThomas LampeAbbas AbdolmalekiKonstantinos BousmalisPublished in: CoRR (2022)
Keyphrases
- manipulation tasks
- reinforcement learning
- real robot
- parallel robot
- mobile robot
- robotic systems
- real time
- vision system
- robot navigation
- robot control
- human robot interaction
- object manipulation
- motion planning
- humanoid robot
- robotic arm
- imitation learning
- service robots
- function approximation
- autonomous robots
- robot behavior
- end effector
- robotic control
- hand eye
- markov decision processes
- multi robot
- robot arm
- visual landmarks
- robotic platform
- degrees of freedom
- computer vision
- robotic tasks
- dynamic programming
- human computer interaction
- reinforcement learning algorithms
- state space
- perceptual aliasing
- vision sensors
- machine learning
- autonomous learning
- temporal difference
- model free
- multi agent
- learning algorithm
- robotic manipulator
- robot programming
- action selection
- optimal policy
- dynamic environments