Action-Quantized Offline Reinforcement Learning for Robotic Skill Learning.
Jianlan LuoPerry DongJeffrey WuAviral KumarXinyang GengSergey LevinePublished in: CoRR (2023)
Keyphrases
- skill learning
- action space
- reinforcement learning
- state space
- markov decision processes
- real time
- action selection
- real valued
- robot arm
- real robot
- mobile robot
- optimal policy
- single agent
- robotic systems
- stochastic processes
- motor skills
- multiple agents
- activity recognition
- markov decision process
- dynamic programming
- learning process
- machine learning