Overestimation
المبالغة في التقدير
حالة تكون فيها قيم Q المقدَّرة أعلى بشكل منهجي من القيم الحقيقية، غالباً بسبب انحياز التعظيم الناتج عن استخدام عامل max مع تقديرات مشوَّشة.
A condition where estimated Q-values are systematically higher than true values, typically caused by maximization bias arising from applying the max operator to noisy estimates.
Also translated asالتقدير المفرط، المبالغة التقديرية
First appears in this corpus in: Deep Reinforcement Learning with Double Q-Learning (2016)
Appears in these papers
- Conservative Q-Learning for Offline Reinforcement Learning2020in the sky ✦
- Conservative Q-Learning for Offline Reinforcement Learning2020in the sky ✦
- Deep Reinforcement Learning with Double Q-Learning2016in the sky ✦
- Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor2018in the sky ✦
- TD3: Addressing Function Approximation Error in Actor-Critic Methods2018in the sky ✦