Bad Apple Challenge
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/endlessinertia/causal-contextual-bandits
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是一份旨在通过用户互动验证DSSApple应用程序在诊断苹果疾病有效性方面的用户研究。它包含了诸如基于图像的上下文和基于相似性的上下文等环境特征，并通过应用辍学概率来模拟用户互动中的变异性。该数据集规模涉及515个挑战性会话，以及163名非专家参与者的数据。研究任务是通过多臂老虎机方法对苹果疾病进行顺序诊断。
This dataset is a user study designed to validate the efficacy of the DSSApple application for apple disease diagnosis through user interactions. It includes environmental contextual features such as image-based context and similarity-based context, and simulates variability in user interactions by applying dropout probabilities. The dataset encompasses data from 515 challenging sessions and 163 non-expert participants. The core research task is to perform sequential diagnosis of apple diseases using the multi-armed bandit method.
提供机构：
DSSApple team
