Adversarially Trained Actor Critic for Offline Reinforcement Learning

Keyphrases

• Actor-critic 100%
• Behavioral Policy 40%
• Continuous Control Tasks 20%
• Coverage-based 20%
• Data Coverage 20%
• Deep Reinforcement Learning (deep RL) 20%
• Function Approximation 20%
• General Functioning 20%
• Hyperparameters 40%
• Learning Implementation 20%
• Low Quality Data 20%
• Model-free Algorithms 20%
• No Regret 20%
• Offline Reinforcement Learning 100%
• Policy Actors 20%
• Reinforcement Learning Algorithm 20%
• Stackelberg Game 20%
• Two-player Games 20%

Computer Science

• Complex Environment 33%
• Continuous Control 33%
• Deep Reinforcement Learning 33%
• Function Approximation 33%
• Large Data Set 33%
• Leaning Parameter 66%
• Reinforcement Learning 100%
• stackelberg game 33%