Cheat MAB
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/whitemech/markov-abstractions-code-ijcai22
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是一个多臂老虎机问题，其中一个臂能提供更高的奖励，但更难以识别。实验结果是基于5次训练的平均值，并且每1.5万集进行一次评估。该数据集通过参数k进行规模调整，k的增加会提升领域复杂性。任务涉及的是带有探索策略的强化学习。
This dataset is a multi-armed bandit problem, where one arm yields higher rewards but is more difficult to identify. Experimental results are based on the average of 5 training runs, with evaluations conducted every 15,000 episodes. This dataset is scalable via the parameter k, where an increase in k elevates the complexity of the domain. The task involves reinforcement learning with exploration strategies.
提供机构：
Authors of the paper.
