BriefGPT.xyz
Jun, 2024
等变离线强化学习
Equivariant Offline Reinforcement Learning
HTML
PDF
Arsh Tangri, Ondrej Biza, Dian Wang, David Klee, Owen Howell...
TL;DR
通过使用有限数量的演示，本研究探讨了在离线强化学习中使用$SO(2)$-等变神经网络的可能性，并通过实验证明了等变性如何提高低数据情况下的离线学习算法。
Abstract
sample efficiency
is critical when applying learning-based methods to
robotic manipulation
due to the high cost of collecting expert demonstrations and the challenges of on-robot policy learning through online Re
→