The Optimal Approximation Factors in Misspecified Off-Policy Value Function Estimation

Keyphrases

• Optimal Approximation 100%
• Misspecified 100%
• Approximation Factor 100%
• Policy Values 100%
• Value Function Estimation 100%
• Learning Problems 16%
• Aliasing 16%
• State Space 16%
• L-norm 16%
• Partial Coverage 16%
• Blowing 16%
• Function Approximation 16%
• Reinforcement Learning 16%
• Misspecification 16%
• Weighted L2-norm 16%
• State Distribution 16%
• Infinite Data 16%
• Finite Approximation 16%

Mathematics

• Upper Bound 100%
• Function Value 100%
• Approximation of Function 50%
• Finiteness 50%
• Open Question 50%
• Multiplicative 50%
• Misspecification 50%
• State Distribution 50%