未提供具体数据集名称

本文探讨了构建视频故事理解问答数据集的层次化方法，提出了记忆容量、逻辑复杂度和DIKW金字塔三个标准。数据集旨在通过不同难度的问答任务，评估机器对视频故事的理解能力。通过这些标准，可以构建一个三维的视频理解地图，用于评估机器智能的发展阶段。数据集的应用领域主要集中在视频理解技术，特别是机器如何通过问答任务来理解和解释视频内容。

This paper explores a hierarchical methodology for building question-answering datasets for video story understanding, and proposes three criteria: memory capacity, logical complexity, and the DIKW Pyramid. The dataset is designed to evaluate machine comprehension of video stories through question-answering tasks with varying difficulty levels. Using these criteria, a three-dimensional video understanding map can be constructed to assess the developmental stages of machine intelligence. The application domains of this dataset primarily center on video understanding technologies, particularly how machines comprehend and interpret video content via question-answering tasks.