Reducing estimation bias via triplet-average deep deterministic policy gradient

• المؤلفون الرئيسيون: WU, Dongming, DONG, Xingping, SHEN, Jianbing, HOI, Steven C. H.
• التنسيق: text
• اللغة: English
• منشور في: Institutional Knowledge at Singapore Management University 2020
• الموضوعات: Averaging technology; deep reinforcement learning (DRL); estimation bias; triplet networks; Numerical Analysis and Scientific Computing; Software Engineering; Theory and Algorithms
• الوصول للمادة أونلاين: https://ink.library.smu.edu.sg/sis_research/5920; https://ink.library.smu.edu.sg/context/sis_research/article/6923/viewcontent/tnnls19ReducingBias_av.pdf

The overestimation caused by function approximation is a well-known property in Q-learning algorithms, especially in single-critic models, which leads to poor performance in practical tasks. However, the opposite property, underestimation, which often occurs in Q-learning methods with double critics...