Frogger Game Scenario

该数据集是一个基于“青蛙过河”游戏的简单场景，旨在通过用户研究来评估可解释强化学习（XRL）框架在帮助人类理解不同强化学习代理能力方面的实用性。游戏中包含了诸如青蛙的受控移动、关卡进度以及部分可观测性等机制，这些都会影响代理对环境的感知。任务是通过用户研究来评估通过解释强化学习的能力。

This dataset features a simple scenario based on the classic Frogger game, designed to evaluate, via user studies, the practical utility of eXplainable Reinforcement Learning (XRL) frameworks in helping humans understand the capabilities of diverse Reinforcement Learning (RL) agents. The game includes mechanisms such as the frog's controlled movement, level progression, and partial observability, all of which affect the agents' perception of the environment. The core task of this dataset is to evaluate, through user studies, the efficacy of using explanatory approaches to facilitate human comprehension of RL agents' capabilities.