MAHALO: Unifying Offline Reinforcement Learning and Imitation Learning from Observations.
Anqi LiByron BootsChing-An ChengPublished in: ICML (2023)
Keyphrases
- imitation learning
- reinforcement learning
- reinforcement learning methods
- function approximation
- reinforcement learning algorithms
- state space
- real time
- control problems
- model free
- optimal policy
- machine learning
- learning capabilities
- markov decision processes
- learning algorithm
- optimal control
- humanoid robot
- multi agent
- learning problems
- robotic systems
- maximum margin
- xml documents
- supervised learning
- temporal difference
- probabilistic model
- learning process
- action space
- function approximators