PAGAR: Imitation Learning with Protagonist Antagonist Guided Adversarial Reward.
Weichao ZhouWenchao LiPublished in: CoRR (2023)
Keyphrases
- imitation learning
- reinforcement learning
- multi agent
- robotic systems
- humanoid robot
- reinforcement learning methods
- maximum margin
- long run
- learning algorithm
- temporal difference
- model free
- function approximation
- state space
- dynamic programming
- average reward
- training data
- real time
- transfer learning
- vision system
- probabilistic model
- reinforcement learning algorithms
- bayesian networks