Academic journal article Educational Technology & Society

Generic Educational Knowledge Representation for Adaptive and Cognitive Systems

Academic journal article Educational Technology & Society

Generic Educational Knowledge Representation for Adaptive and Cognitive Systems

Article excerpt

Introduction

The development of specifications, standards and tools related to the Semantic Web (Hendler, 2001; Dolog & Nejdl, 2007) has led to the development of ontologies with the objective of globally interconnecting knowledge. Applying this knowledge to education requires the development of browser technologies that automatically filter knowledge in order to generate an appropriate offer for each educational objective. Existing e-learning management systems (LMS) use instructional knowledge coded with different levels of granularity and associated with multimedia resources such as ADL-SCORM or IMS-CC. However, research and development centers propose adaptive models that range from basic instructional rules to implicit and complex representations related to the needs of each particular implementation. Currently, efforts are focused on the creation of models, specifications and standards that increase computational integration (Kozaki, Hayashi, Sasajima, Tarumi, & Mizoguchi, 2008; Dietze, Gugliotta, & Domingue, 2007; De Roure & Hendler, 2004) using ontologies (Gaeta, Orciuoli, & Ritrovato, 2009; Zitko, Stankov, Rosic, & Grubisic, 2009; Isotani & Mizoguchi, 2008; Boyce & Pahl, 2007; W3C, 2004) in what is called the Intelligent/Semantic Web (Devedzic, 2004; Zhong, Liu, & Yao, 2002) or Web3.0 (Zeldman, 2006).

This approach leads to the emerging paradigm of open-corpus knowledge (Sosnovsky, 2009). It offers a wealth of knowledge described at a semantic level to educational systems, which greatly improves the capabilities of automatic systems. However, the interoperability of learning platforms is restricted to the exchange of information in domain and student models that follow specifications. We believe that the interoperability of educational systems on the Grid requires a framework to facilitate interaction at a functional level and to overcome the limitations of the specifications. The LTI project (IMS, 2009) is an example of such efforts.

Work on Intelligent Educational Systems (IESs) is traditionally divided into two main paradigms (Nicholas & Martin, 2008): Intelligent Tutoring Systems (ITSs) (Murray, 1999; Wenger, 1987) and Adaptive Hypermedia Systems (AHSs) (Dolog, 2008; Brusilovsky, 1996). The former are designed to guide students in acquiring specific skills to complement the primary instructional method and use Constraint-Based Modeling from Ohlsson's theory. The system works by correcting the learner's practice errors and guiding him/her to improve; the WETAS system is an example (Martin, Mitrovic, & Suraweera, 2008). AHSs, in contrast, are directed to a complete theoretical instruction through learning concepts. They use intelligent systems to adapt the sequence, presentation and contents according to the student model. New educational systems incorporate the latest advances in cognitive science. Some of them employ estimated cognitive parameters of learning, while others develop computing architectures that simulate the mental processing of students. Cognitive control is the ability to integrate information from a multitude of sources and use that information to flexibly guide behavior. From the characteristics described above, we developed a framework for educational systems called the Cognitive Ontology of Educational Systems (COES) whose main objective is to provide functional interoperability. The teaching-learning process is a complete and indivisible entity in which numerous real and abstract elements are artificially bound together by functions in different areas: emotional, cognitive, instructional, behavioral, etc. The traditional IES architecture (De Bra, Houben, & Wu, 1999; Murray, 1998) has kept a functional division in which a processor uses the rules of a pedagogical/adaptive model to select content from a domain model. However, many systems encode pedagogical models through specific operational rules within the domain model. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.