Towards automatic learning resources organization via knowledge graphs

• Autores: Rubén Francisco Manrique Piramanrique
• Directores de la Tesis: Olga Marino Drews (dir. tes.)
• Lectura: En la Universidad de los Andes (Colombia) ( Colombia ) en 2019
• Idioma: español
• Enlaces
  • Tesis en acceso abierto en: hdl.handle.net
• Resumen

  • The Web increasingly permeates the activities of the human being. Learning has not been the exception, and we often learn on our own with resources from websites and online platforms. Due to the large volume of resources on the Web and the absence of a tutor's guidance, self-directed learners face two major challenges: to make an effective selection of learning resources and to organize them in a pedagogically useful way. Given a learning goal, effective selection refers to the selection of the resources that address the concepts and the topic related to the goal. On the other hand, the organization of the resources refers to the sequence in which these resources should be studied taking into account the prerequisite relationships of the concepts (i.e. complex concepts require the understanding of more basic ones). This research addresses these problems and proposes methods that support the selection and organization of learning resources. Our proposal is based on Knowledge Graphs belonging to the Linked Open Data initiative. We hypothesize that these processes (selection and organization) can be facilitated by introducing the background knowledge present in these Knowledge Graphs. In recent years, different institutions started to publish and share their knowledge in specific domains in the form of data structured under Semantic Web standards, with the aim of promoting interconnected knowledge, reuse and discovery. With the constant increase of these Knowledge Graphs, we assume that the processes of selection and organization can be more effective, automatic and generalizable to multiple domains. To demonstrate this, we propose the construction of a set of algorithms (i) to construct semantic representations of resources and learning goals based on their textual content, (ii) to infer relations of prerequisite between concepts of a Knowledge Graph and (iii) to select and sequence a set of resources based on the two previous processes.