Academic Curriculum Generation using Wikipedia for External Knowledge

Academic Curriculum Generation using Wikipedia for External Knowledge Anurag Reddy Muthyala IIIT Hyderabad India anurag.reddy@research.iiit.ac.in & Vikram Pudi IIIT Hyderabad India vikram@iiit.ac.in November 15, 2022 Abstract In this paper, we address the problem of auto- matic academic curriculum generation. A curricu- lum outlines definitive topics with their sub-topics and enables teachers and students to form an over- all idea of the course outcomes and goals, and a plan of what to teach and learn to achieve those goals. Automatic curriculum generation is rel- evant in modern times with the ever increasing, rapidly changing, digitally-available academic con- tent, that is too large for manual processing by human teams. Using Wikipedia as an external knowledge-base, along with a pipeline of standard components, we show that it is possible to generate human-interpretable 2-level topic hierarchies. We show that our approach works on publicly available textbooks, by first removing their title-structure, and then automatically regenerating a 2-level title structure that is on-par. 1 Introduction We address the problem of automatic academic curriculum generation. We treat a curriculum as one that outlines definitive topics with their sub- topics, in order to enable teachers and students to form an overall idea of the course outcomes and goals, along with a plan of what to teach and learn to achieve those goals. Automatic curriculum generation is relevant in modern times with the ever increasing, rapidly changing, digitally-available academic content, that is too large for manual processing by human teams. The need for automation is crucially felt in interdisciplinary fields [Jacobi(2014)], and to personalize content and presentation for individual student needs and flow [Katuk and Ryu(2010)]. We formulate the problem as generating a 2- level human-interpretable topic hierarchy consist- ing of module titles and the topics within those modules. This caters to the most common require- ment of most academic curricula. However, we add that this formulation is not restrictive as it is possible, when needed, to devise methods to gen- erate deeper hierarchies using the base method for 2-level hierarchies through recursive application. We aim to implement a subject-centered gen- erative model that generates topics based on the domain knowledge instead of the learner’s ability. This ensures that we generate a uniform structure for all learners, which is the typical goal of a cur- riculum. Our model is an unsupervised approach based on the probability distribution of words for topic generation. We incorporate the salient features necessary for generating curriculum from given set of documents. The primary objective is to generate a 2-level module-topic hierarchy following a data-driven ap- proach that does not depend on the academic do- main and discipline. Our model is a simple pipeline of standard components. In order to create a se- 1 mantic structure (titles) from the candidates that are generated, we use Wikipedia for external knowl- edge and links to Wikipedia pages as learning ob- jects to enhance learner’s curiosity. Employing the previously mentioned approaches, we generate module-topic hierarchies that are on par with hu- man generated ones, by using ideas …

Content truncated to 500 words