Dynamic regret of policy optimization in non-stationary environments

Keyphrases

• Policy Optimization 100%
• Dynamic Regret 100%
• Non-stationary Environments 100%
• Non-stationarity 42%
• Regret 28%
• Transition Kernel 28%
• Adaptation 14%
• Two-component 14%
• Changing Environment 14%
• Full Information 14%
• Reinforcement Learning 14%
• Optimization Algorithm 14%
• Near-optimal 14%
• Near-optimality 14%
• Reinforcement Learning Algorithm 14%
• Regret Bounds 14%
• Reward Feedback 14%
• Free Policy 14%
• Regret Analysis 14%
• Fixed Transition 14%
• Model-based Reinforcement Learning 14%
• Episodic MDP 14%

Computer Science

• Optimization Policy 100%
• Reinforcement Learning 100%
• Model-Free Reinforcement Learning 50%
• Optimization Algorithm 50%