Breaking the deadly triad with a target network
The deadly triad refers to the instability of a reinforcement learning algorithm when it employs off-policy learning, function approximation, and bootstrapping simultaneously. In this paper, we investigate the target network as a tool for breaking the deadly triad, providing theoretical support for...
Главные авторы: | Zhang, S, Yao, H, Whiteson, S |
---|---|
Формат: | Conference item |
Язык: | English |
Опубликовано: |
PMLR
2021
|
Схожие документы
-
Breaking the deadly triad in reinforcement learning
по: Zhang, S
Опубликовано: (2022) -
Austrian syndrome: The deadly triad
по: Yong Il Shin, и др.
Опубликовано: (2020-01-01) -
Austrian syndrome: Resurgence of an old and deadly triad
по: Kristian Rivera, и др.
Опубликовано: (2024-04-01) -
Insulin resistance, impaired postprandial lipid metabolism and abdominal obesity. A deadly triad.
по: Frayn, K
Опубликовано: (2002) -
The Role of Dark Triad of Personality, Material Preoccupations, and Sorrow of Being Compared in Deadly Sins
по: Ghasem Ahi
Опубликовано: (2019-09-01)