Abstract:

Ontologies are essential tools for the semantic representation of knowledge in a machine-processable format to achieve the goals of personalized e-learning. Unfortunately, ontology development is difficult, time-consuming, and as a result, they are rarely used in practical e-learning systems. Therefore, other tools for semantic knowledge modeling are needed. In this work, we will explore the joint use of ontologies and related semantic technologies, such as concept maps (CM) and knowledge graphs (KGs) for personalized learning and training.

CMs are frequently used in e-learning during teaching, learning, knowledge assessment and self-assessment. CMs provide useful and facilitative instruments in many stages of planning, developing and carrying out personalized and adaptive education. CMs can be applied for modeling an e-learning course and in this way help course designers in eliciting and determining the needed learning content and in systematizing it. CMs can be also used to specify the relationships among the concepts and also among resources and places where they are defined or explained. The relations, representing pedagogical issues, such as prerequisite or corequisite relations among concepts, can also be captured by CMs, providing information on the instructional sequence. In this way, CMs provide a structure for creating and sequencing units of learning. CMs can also be used for searching, recommending or inserting media resources that are associated with concepts. CMs can be developed easier due to the presence of many easy to use concept maps development tools. The main drawback of concept maps is related to the difficulties of its usage by software systems.

KGs represent Linked Open Data and are closely related to ontologies. Recently, Resource Description Framework (RDF) KGs also have been used in personalized e-learning for semantic description of learning objects and predict students’ performance based on the description of the whole learning ecosystem. We will explore ways, technologies and tools for representing KGs, possibilities and outcomes of its usage in e-learning domain. Personal Knowledge Graphs (PKGs) are small-sized user-centric KGs, mainly used for personalised representation of user data and interests. Big, well-established encyclopedic KGs, such as DBpedia are used in the internet for representing the structure of knowledge in machine-prosesable way.

One of our goals in this paper is to explore and find possibilities of usage both concept maps and ontologies (via mapping) by software systems for personalized tutoring or recommendation of learning content. We propose a methodology for usage of both concept maps and ontologies in personalized e-learning systems to support active and adaptive learning and teaching (including selection of learning strategies, and learning paths) for individuals as well as for groups of learners. We will investigate the applicability of well working ontology mapping methods in the process of mapping various types of concept maps for supporting adaptive and personalized resource recommendation and learning. We will also explore the relationships between specifics of mapped CMs (including scientific domain or labeling language) and possibilities to achieve high quality of automated mapping in the context of personalized educational systems. Our main aim is to find ways and positive results of jointly use of ontologies, KGs and CMs in intelligent personalized learning or resource recommendation systems.

Keywords:

Intelligent educational system, Personalization, Knowledge Graphs, Concept maps, e-learning, Ontologies.