Obstacle Tower
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/Miffyli/rl-action-space-shaping
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是在障碍塔环境中使用的，旨在测试具有修改后动作空间的强化学习效果。同时，该数据集还测试了不同的动作空间配置，以评估智能体的表现。所涉及的任务是强化学习。
This dataset is employed in the Obstacle Tower environment, with the dual purposes of evaluating the performance of reinforcement learning (RL) systems with modified action spaces, and testing diverse action space configurations to assess the performance of AI agents. All tasks encompassed by this dataset belong to the field of reinforcement learning.
提供机构：
OpenAI
