As the number of textbooks soars, people may be stuck into thousands of books when learning knowledge. In order to provide a concise yet comprehensive picture for learning, we propose a novel framework, called MM4Books, to automatically build metro maps for efficient knowledge learning by summarizing massive electronic textbooks. We represent each book in digital libraries as a sequence of chapters, and then obtain learning objects by clustering the semantically similar chapters via an unsupervised clustering method to create a learning graph, and then build the metro map by applying an integer linear programming-based technique to select a collection of high informative and fluent but low redundant learning paths from the learning graph. To the best of our knowledge, it is the first work to address this task. Experiments show that our proposed approach outperforms all the state-of-the-art baseline approaches, and we also implemented a practical MM4Books system to prove that users can really benefit from the proposed approach for knowledge learning.

中文翻译：

---

通过汇总大量电子教科书来进行有效知识学习的Metro地图

随着教科书数量的猛增，人们在学习知识时可能会陷入成千上万本的书中。为了提供一个简洁而全面的学习画面，我们提出了一个新颖的框架，称为MM4Books，通过汇总大量的电子教科书来自动构建地铁地图以进行有效的知识学习。我们将数字图书馆中的每本书都按章节顺序表示，然后通过无监督聚类方法将语义相似的章节聚类以创建学习图，从而获得学习对象，然后通过应用基于整数线性规划的技术来构建地铁地图从学习图中选择信息丰富，流利但冗余程度低的学习路径的集合。据我们所知，这是解决此任务的第一项工作。实验表明，我们提出的方法优于所有最新的基线方法，并且还实现了实用的MM4Books 证明用户可以真正从建议的知识学习方法中受益。