Model-Based Reinforcement Learning for Offline Zero-Sum Markov Games

Keyphrases

• Sample Complexity 100%
• Model-based Reinforcement Learning 100%
• Zero-sum Markov Game 100%
• Nash Equilibrium 50%
• Markov 50%
• Complexity Bounds 50%
• Approximate Nash Equilibrium 50%
• Targeting Accuracy 50%
• Zero-sum 50%
• Offline Data 50%
• Variance Reduction 50%
• Value Iteration 50%
• Bernstein 50%
• Infinite Horizon 50%
• Minimax Optimality 50%
• S-states 50%
• Distribution Shift 50%
• Sample Splitting 50%
• Bounds for Zeros 50%
• Lower Confidence Bound 50%
• Markov Games 50%

Mathematics

• Optimality 100%
• Nash Equilibrium 100%
• Approximates 50%
• Minimax 50%
• Variance Reduction 50%
• Lower Confidence Bound 50%

Computer Science

• Nash Equilibrium 100%
• Model-Based Reinforcement Learning 100%
• Confidence Bound 50%
• Variance Reduction 50%