Multi-Agent Reinforcement Learning with General Utilities via Decentralized Shadow Reward Actor-Critic

Keyphrases

• General Utility 100%
• Multi-agent Reinforcement Learning 100%
• Actor-critic 100%
• Occupancy Measure 60%
• Stationary Point 40%
• Cumulative Return 40%
• Learning-based 20%
• High Probability 20%
• Local Utility 20%
• Globally Optimal 20%
• State Action 20%
• Optimal Policy 20%
• Neighborhood Information 20%
• Gradient Update 20%
• Nonlinear Function 20%
• Long-term State 20%
• Local Gradient 20%
• Risk Sensitivity 20%
• Policy Parameters 20%
• Policy Evaluation 20%
• Weighted Averaging 20%

Computer Science

• Multi-Agent Reinforcement Learning 100%
• Stationary Point 66%
• Nonlinear Function 33%
• Weighted Averaging 33%
• Risk Sensitivity 33%