Generalized Advantage Estimation (GAE)

التقدير المعمّم للميزة

أسلوب لتقدير الميزة (مقدار تفوّق فعل على المتوسط) يوازن بين الانحياز والتباين عبر معلمة λ، بحساب المتوسط الموزون أُسّياً لتقديرات الميزة بآفاق زمنية مختلفة.

A method for estimating advantages (how much better an action is than average) that balances bias and variance through a λ parameter, computing an exponentially-weighted average of advantage estimates at different time horizons.

Also translated asتقدير الميزة المعمّم، GAE

First appears in this corpus in: Policy Gradient Methods for Reinforcement Learning with Function Approximation (1999)

Appears in these papers

High-Dimensional Continuous Control Using Generalized Advantage Estimation2016 in the sky ✦
Learning Dexterous In-Hand Manipulation2019 in the sky ✦
Policy Gradient Methods for Reinforcement Learning with Function Approximation1999 in the sky ✦
Proximal Policy Optimization Algorithms2017 in the sky ✦