离线学习

如果我们已经有大量的实验数据，如何从中学出最佳策略？这就是离线学习。

课程材料

• 滑铁卢 CS885 RL PPT
• Berkeley CS285 Lec 15: Offline Reinforcement Learning (Part 1), slides, Youtube Video
• Berkeley CS285 Lec 16: Offline Reinforcement Learning (Part 2), slides, Youtube Video
• 上海交大伯禹增强学习 Lec 11 离线强化学习

论文

斯坦福 CS 224r 论文 Offline RL

• Conservative Q-Learning for Offline Reinforcement Learning. Kumar et al. (2020)
• COMBO: Conservative Offline Model-Based Policy Optimization. Yu et al. (2021)
• Offline Reinforcement Learning with Implicit Q-Learning. Kostrikov et al. (2021)

滑铁卢 论文

• Levine, Kumar, Tucker, Fu (2021) Offline reinforcement learning: Tutorial, review, and perspectives on open problems, arxiv.
• Kumar, Zhou, Tucker, Levine (2020) Conservative Q-Learning for Offline Reinforcement Learning, NeurIPS.

练习

• 上海交大伯禹增强学习 练习 第18章-离线强化学习.ipynb
• CS886 练习 2，conservative Q-Learning
• 伯克利 CS285 HW 5: Exploration and Offline reinforcement learning, Website
• 斯坦福 CS224R DRL HW 3: Offline Mujoco

课本材料

N/A