Apr, 2023
基于模型的离线强化学习中的不确定性驱动轨迹截断
Uncertainty-driven Trajectory Truncation for Model-based Offline Reinforcement Learning
Junjie Zhang, Jiafei Lyu, Xiaoteng Ma, Jiangpeng Yan, Jun Yang...
TL;DRTATU is proposed to address the issue of uncertainty in synthetic samples for model-based offline RL algorithms and has been shown to improve the performance of various RL algorithms on the D4RL benchmark.