HINDSIGHT FORESIGHT RELABELING FOR META-REINFORCEMENT LEARNING

Keyphrases

• Delabeling 100%
• Meta-reinforcement Learning 100%
• Reinforcement Learning Algorithm 25%
• Performance Improvement 12%
• Easy-to-implement 12%
• Training Process 12%
• Learning Settings 12%
• Performance Efficiency 12%
• Sampling Efficiency 12%
• Shared Experiences 12%
• Reward Function 12%
• Asymptotic Performance 12%
• New Behaviors 12%
• New Task 12%
• Task Allocation 12%
• Reinforcement Learning Agent 12%
• Training Tasks 12%
• Reinforcement Learning Task 12%
• Meta-training 12%
• Multi-task RL 12%

Computer Science

• Reinforcement Learning 100%
• Training Process 16%
• Learning Agent 16%

Neuroscience

• Reinforcement Learning 100%
• Behavior (Neuroscience) 16%