TOWARDS SEMANTIC WEB ENHANCED LEARNING
Danail Dochev and Gennady Agre
Institute of Information Technologies – Bulgarian Academy of Sciences, Acad. G. Bonchev 2, Sofia, Bulgaria
Keywords: Technology Enhanced Learning (TEL), Semantic Web Services (SWS), Service oriented Architecture
(SOA).
Abstract: The paper discusses an approach for facilitating the authoring process for Technology Enhanced Learning
by providing service oriented access to learning materials based on their content. This service-oriented
approach is based on Semantic Web Service technologies. It reflects current work in the frame of an on-
going Bulgarian research project SINUS “Semantic Technologies for Web Services and Technology
Enhanced Learning”.
1 INTRODUCTION
The research and development efforts in the field of
Technology Enhanced Learning (TEL) address the
exploitation of the huge potential of Information and
Communication Technologies to unbound and
multiply the opportunities for accessing learning in
general. While learning human knowledge must be
constructed – not by the teacher (or courseware
author) for the student, but by the student himself
with the teacher’s or peers’ assistance. Technology
has to provide information and create situations,
enabling activities for constructing learner’s
knowledge. Current TEL systems and tools facilitate
the development of various learning situations,
reducing substantially the time and space limitations
on the learning process, supporting more pro-active
and knowledge-pulling types of learning
architectures (Aceto et al, 2007).
Learning content has always been regarded as
keystone for all learning situations in classical edu-
cation as well as in e-earning activities. That’s why
the authoring of learning materials is one of the most
important and labour-intensive activities in the
modern TEL practice. The present paper discusses a
way for facilitating the authoring process by
providing service oriented access to learning
materials based on their content. The paper reflects
the current work accomplished in the frame of an
on-going national research project SINUS “Semantic
Technologies for Web Services and Technology
Enhanced Learning” (http://sinus.iinf.bas.bg
).
2 SOME TENDENCIES IN TEL
The continuous growth of multiple digital libraries
makes the problem to open enormous existing digital
resources to be easily available for learning needs
more significant and actual. Education-focused
digital archives are expected to support the reuse of
resources for the creation of new learning materials.
This involves re-purposing - integrating and relating
existing resources into a new context. A learning
context has many dimensions including various and
difficult to coordinate social and cultural factors: the
learner’s educational system, the learner’s cognitive
abilities, his/her prior knowledge, learning style,
cultural preferences.
The need to specify and separate learning resour-
ces from the information about the context of their
usage led to creation of various kinds of metadata
schemas. This work has focused around the notion
of learning object (LO), as a conceptual base, capab-
le to guarantee interoperability to the rapidly grow-
ing number of Web-based educational applications.
In the TEL tools based on current e-Learning stan-
dards (IMS-LD, LAMS etc.), a LO can be conside-
red as a static and monolithic block, since once crea-
ted, it is rather difficult to change or modify its inner
resources and/or to add/remove services and resour-
ces at run-time. The traditional approaches to create
a LO typically rely on expertises of the institutional
designer only, and have practically no capability to
reuse existing blocks.
In the last years there is an increasing interest for
semantic descriptions of learning materials as a
212
Dochev D. and Agre G. (2009).
TOWARDS SEMANTIC WEB ENHANCED LEARNING.
In Proceedings of the International Conference on Knowledge Management and Information Sharing, pages 212-217
DOI: 10.5220/0002330202120217
Copyright
c
 SciTePress

mean to make TEL systems more flexible and help-
ful for end users. By accessing semantically anno-
tated and adequately structured knowledge from
digital archives lecturers/authors should participate
in ‘open source’ dynamic content development from
massive, dynamically growing learning resources.
This tendency fits well to the new TEL research
tendencies aiming at more learner-centric, interest-
oriented, pro-active and knowledge-pulling types of
learning architectures.
2.1 The LOGOS Approach
The FP6 IST project LOGOS “Knowledge-on-de-
mand for ubiquitous learning”, developed by 15
partner organizations from 8 countries, concerned
innovative developments for all the e-Learning
processes components – resources, services,
communication spaces (www.logosproject.com).
The project focused on learning resources by add-
ressing the transformation of digitised information
from existing large-scale repositories into learning
content, adequately enhancing and facilitating the
learners’ knowledge building. One of the main re-
sults of the project was the implementation of a
platform for ubiquitous (any place, any time, perso-
nalized) learning, combining an authoring studio for
creation of learning materials from existing digital
repositories with emphasis on semantic annotation
and access, and facilities for cross-media coursewa-
re delivery through different communication chan-
nels (IP-based, digital TV and/or mobile devices)
The architecture of LOGOS authoring studio is
based on hierarchy of information objects:
• Media objects - ‘raw’ multimedia (MM) ob-
jects, catalogued with some technical character-
istics to enable multiple channel delivery;
• Digital objects - media objects, annotated with
technical and administrative, as well as with se-
mantic metadata (based on domain ontologies);
• Learning objects (incl. assessment objects) -
digital objects, enriched with educational
metadata (LOM);
• Courseware objects - graphs of learning
activities associated with learning objects.
The LOGOS authoring process is sketched on
Figure 1. Some lessons learnt from LOGOS practice
of authoring learning materials are:
• Investigations to unify the mechanisms and
tools for access to information objects and
processing of semantic information would
increase the effectiveness of TEL platforms
development and maintenance by creating
“lightweight” versions of authoring tools, based
on well defined and well understood use cases.
• It seems desirable to use and combine limited-
world domain ontologies, reflecting viewpoints
on the domain for specific well defined user
groups in order to increase the effectiveness of
the preparation of semantic resources.
• The manual high labour-intensive work to
annotate resources should be facilitated by
providing templates and finding similarities
with existing annotations.
• There is definite need to increase the usability
of the authoring tools making them more
dedicated and end-user-friendly in order to
support effectively the indexing and re-
purposing of learning content.
• The authoring tools have to be validated and
experimented with different user groups to fit
well enough with the significant use cases for
the different user roles of the TEL platform.
Figure 1: LOGOS authoring process.
The limitations of current approaches for
authoring learning content can be summarized as
follows:
• High knowledge and labour intensiveness of
learning content preparation.
• Rigid schemes of metadata standards,
considering only objective (factual) and static
(created only once) metadata
TOWARDS SEMANTIC WEB ENHANCED LEARNING
213

• Limited reusability across different learning
contexts, and metadata standards.
• Limited dynamic adaptability to actual learning
context.
These limitations result in high development
costs and lack of enough high quality adaptable e-
Learning materials available for mass usage.
Currently a number of research teams are
exploring virtualisation mechanisms by which each
resource is virtualised as a service. Such attempts to
use SWS technology aims to exploit common
mechanisms and tools to reuse the learning materials
and other already developed building resources by
enabling automatic search and late binding of res-
ources and services.
3 SWS TECHNOLOGY
Web services define a new paradigm for the Web, in
which a network of computer programs becomes the
consumer of information. However, Web service
technologies only describe the syntactical aspects of
a Web service and, therefore, only provide a set of
rigid services that cannot adapt to a changing
environment without human intervention. Realiza-
tion of the full potential of the Web services requires
further technological advances in service inter-
operation, discovery, choreography and
orchestration. A possible solution to these problems
is likely to be provided by combining the Semantic
Web technology (ontologies) with Web services.
The result is Semantic Web Services (SWS), which
are self-contained, self-describing, semantically
marked-up software resources that can be published,
discovered, composed and executed across the Web
in a task driven semi-automatic way. SWS can cons-
titute a solution to the integration problem, as they
enable dynamic, scalable and reusable cooperation
between different systems and organizations.
3.1 Core Initiatives and Projects
There are two major initiatives working on
developing a world-wide standard for the semantic
description of Web services. The first one is OWL-S
(www.daml.org/services/owl-s/) a collaborative
effort by BBN Technologies, Carnegie Mellon
University, Nokia, Stanford University, SRI
International and Yale University. OWL-S is
intended to enable automation of Web service
discovery, invocation, composition, interoperation
and execution monitoring by providing appropriate
semantic descriptions of services. The second one is
the Web Service Modeling Ontology (WSMO) - a
European initiative intending to create an ontology
for describing various aspects related to SWS and to
solve the integration problem. WSMO has been
under development over the past four years and has
been adopted in several IST FP-6 Integrated Projects
such as DIP (dip.semanticweb.org/), SEKT
(sekt.semanticweb.org/), Knowledge Web
(knowledgeweb.semanticweb.org/), ASG (asg-
platform.org/), INFRAWEBS (www.infrawebs.eu)
and LUISA (www.luisa-project.eu) by consortia
including in total more than 70 academic and
industrial partners.
WSMO (www.wsmo.org) is conceptually
grounded on Modelling Framework (WSMF)
(Fensel and Bussler 2002) using Web Service
Modeling Language (WSML) for describing its four
main components - ontologies providing the formal
semantics to the information used by all other
components, goals specifying objectives that a client
may have when consulting a Web service, Web
services representing the functional part which must
be semantically described in order to allow their
semi-automated use and mediators used as
connectors providing interoperability facilities
among the rest of components.
At the moment practical application of SWS
technologies is still rather restricted due to several
reasons, some of which are high complexity of both
OWL-S and WSMO, the lack of standard domain
ontologies and unavailability of mature tools
supporting WSMO or OWL-S.
3.2 INFRAWEBS Project
A technology-oriented step for overcoming some of
the above-mentioned problems was proposed by IST
FP-6 project INFRAWEBS (Agre et al. 2009). It
focused on developing a Semantic Service
Engineering Framework enabling creation,
maintenance and execution of WSMO-based SWS,
and supporting SWS applications within their life-
cycle. Being strongly conformant to the current
specification of various elements of WSMO
(ontologies, goals, semantic services and mediators),
the INFRAWEBS Framework managed with the
complexity of creation of such elements by:
• Identifying different types of actors (users) of
Semantic Web Service Technologies;
• Clarifying different phases of the Semantic
Service Engineering process, and
• Developing a specialized software toolset
oriented to the identified user types and
KMIS 2009 - International Conference on Knowledge Management and Information Sharing
214

intended for usage in all phases of the SWS
Engineering process.
The INFRAWEBS Framework has been
implemented on the top of an extensible Enterprise
Service Bus (ESB) middleware (Mule) that exposes
the public methods of the INFRAWEBS
components and can be extended by any future
components or services. Such Integrated
INFRAWEBS Framework (IIF) can be seen as the
underlying P2P infrastructure for communication
and integration of all the INFRAWEBS components,
and at the same time the unique selling point for
exposing the functionality of such components to the
external world in form of services. All these features
have allowed recognizing INFRAWEBS as one of
the first frameworks for semantic service
engineering that covers the whole SWS life-cycle
and allows creation of complex semantically-
enabled applications.
4 AN APPROACH TO SEMANTIC
WEB ENHANCED LEARNING
As it has been already mentioned SWS technology
promises to provide a possible solution to increase
the effectiveness of preparation and use of adaptable
learning content by changing the current data- and
metadata-based paradigm to TEL to a dynamic
service-oriented approach (Dietze et al., 2007).
The current national research project SINUS -
Semantic Technologies for Web Services and
Technology Enhanced Learning (sinus.iinf.bas.bg)
aims at providing a framework for development
TEL-oriented applications based on SWS
technology. It stands on methods and tools
developed as well as on lessons learnt in LOGOS
and INFRAWEBS projects.
4.1 SINUS Conceptual Architecture
The Conceptual Architecture of SINUS is an
adaptation of the INFRAWEBS Semantic SOA-
based architecture (Agre 2009) towards TEL
applications. The architecture defines two main
elements: SINUS Design-time and Run-time
Environments (Figure 2).
The Design-time Environment consists of a
decentralized network of nodes or peers connecting
with the rest of peers through the integrated
infrastructure. Each peer consists of a centralized
bundle of components within a single server and
offers a native Java support for integration, and a
Web Service interface, which allows the easy
incorporation of the components.
Design
Design
-
-
time Environment
time Environment
Semantic
Repository
Digital
Library
Information
Layer
Information Indexer and Retriever
S WS
Editor
Ontology
Editor
Semantic Goal
Editor
SWS
Composition
Methods
Logic-based
Matching
Similarity
Evaluation Methods
Learning
Methods
Metadata
Editor
Problem-solving
Method Layer
Tool
Layer
Semantic WS
Executor
Semantic WS
Discoverer
Semantic WS
Composer
Run
Run
-
-
time Environment
time Environment
Deployment
Layer
Semantic Studio
Semantic
Repository
Digital
Library
Information
Layer
Information Indexer and Retriever
S WS
Editor
Ontology
Editor
Semantic Goal
Editor
SWS
Composition
Methods
Logic-based
Matching
Similarity
Evaluation Methods
Learning
Methods
Metadata
Editor
Problem-solving
Method Layer
Tool
Layer
Semantic WS
Executor
Semantic WS
Discoverer
Semantic WS
Composer
Run
Run
-
-
time Environment
time Environment
Deployment
Layer
Semantic Studio
Figure 2: SINUS Conceptual Architecture.
The components of a peer are organized in two
directions: (i) problem solving based on semantic
information (or Logic-based problem solving) versus
problem solving based on non-semantic information
(similarity-based problem solving); and (ii) different
types of information needed for both kinds of
problem solving.
The SINUS Design-time Environment proposes:
• Information structures for storing and
retrieving semantic (ontology-based) and non-
semantic (metadata-based) information:
• Semantic Repository enables efficient
storage and retrieval of all elements of the
Semantic Web (WSMO objects): goals,
ontologies and SWS. From TEL point of
view the Repository stores all resources
annotated by ontologies used in TEL
process – digital objects, learning objects,
learning designs and ontologies themselves.
• Digital Library – contains digital resources
with some initial metadata descriptions that
can be further used for creating ontology
annotated resources.
• Information Indexer and Retriever – con-
tains a special representation of both Se-
mantic (WSMO) and non-semantic res-
ources of the platform. Such a
representation allows effective similarity-
based search and retrieval of all resources
based on their content.
• Tools for creation and maintenance of semantic
and non-semantic resources:
• Semantic Studio – an integrated tool set
aiming at designing all elements of
WSMO-based Semantic Web objects by
TOWARDS SEMANTIC WEB ENHANCED LEARNING
215

effective reuse of already existing semantic
and non-semantic descriptions stored in the
Semantic Repository and Digital Library.
The Semantic Studio contains Ontology
Editor aiming at creating ontologies in a
user-friendly manner; SWS Editor
providing graphical ontology-based way for
creating and composition of WSMO-based
SWS and Semantic Goal Editor providing
means for creation of WSMO-based
reusable learning goals and goals templates
(designs) used for developing SWS-based
TEL applications.
• Metadata Editor aims at creating initial
metadata annotations.
• Methods used for creating and maintaining
Semantic and non-semantic objects:
• Combination of TEL-specific and logic-
based methods for object discovery.
• TEL-specific decision-support methods for
dynamic service composition.
• Several methods for calculating object
similarity – structural, linguistic, statistical,
fuzzy, etc.
The SINUS Design-time Environment will adapt
and enhance such tools as INFRAWEBS Designer
(Agre and Dilov 2008) and INFRAWEBS
Organization Memory (Andonova et al. 2007),
which have been evaluated as an achievement of the
INFRAWEBS project that has a potential impact on
the adoption of SWS on a larger scale.
The SINUS Run-time Environment is
responsible for communication with different users
and peers of the framework. It will provide a
middleware for discovery, dynamic composition,
execution and monitoring of SWS.
4.2 Matching Authoring Process to
SINUS Architecture
The SINUS Architecture allows natural
implementation on a layered approach for creation
TEL applications. First layer is devoted to creation
of “basic” semantic and non-semantic multimedia
resources. It is assumed that operation on this layer
will be accomplished by two types of users –
annotators and knowledge managers. The annotators
will use the Metadata Editor for initial annotation of
raw multimedia objects (created by them or found
somewhere in WWW). The annotated objects will
be published in SINUS Digital Library and become
available for further semantic annotations for all
users of the framework.
The knowledge managers will use the Ontology
Editor of the Semantic Studio for creating WSML
ontologies to be further used in the process of
semantic annotation. The created ontologies will be
published in a local (belonging to a concrete SINUS
peer) Semantic Repository, which, in its turn, makes
them available for all other peers of the framework.
The second layer deals with the creation of se-
mantically annotated digital objects and learning
objects – combination of such digital objects addi-
tionally annotated by educational metadata (LOM).
From technology point of view both types of objects
will be represented in a uniform way – as SWS. For
this the users of the framework (playing the role of
Educationalists) will use the SWS Editor of the
Semantic Studio, the domain ontologies stored in the
Semantic Repository and a WSML version of LOM
(s-lom) ontology, developed under LUISA project.
As usual, all developed objects will be made
available for all SINUS peers by publishing them in
a local Semantic Repository.
The third layer is intended for creating learning
goals and learning designs describing possible goals
of the end-user of a SWS Leaning application. At
design-time these objects (represented as WSMO
goals) are created by the Goal Editor of the Semantic
Studio, which intensively uses some formally
represented domain and learning strategy ontologies.
These goals are the basic blocks for formulating
end-user queries to a SWS Learning application in
run-time.
In Run-time each end-user goal is dynamically
decomposed on sub-goals, for which the
corresponding sets of matching Semantic Services
(learning or semantic digital objects) will be
discovered. The resulted dynamically composed
complex process may be stored again in the
Semantic Repository as a new complex SWS or be
executed by the Framework Run-time Environment.
It should be mentioned that the process of
publishing of each newly created object (semantic or
non-semantic) is passed through the indexing the
object description done by the Information Indexer
and Retriever component of the Framework
implemented as Web service. The process of
creating semantic objects is facilitated by ability to
reuse the existing descriptions of similar objects
stored in the Semantic Repository. The search of
such similar objects is realized by the Information
Indexer and Retriever based on object indexes
created during the process of publishing the objects.
KMIS 2009 - International Conference on Knowledge Management and Information Sharing
216

5 DISCUSSION AND FUTURE
WORK
In this paper we have proposed an approach for
developing a new Semantic SOA-based framework
oriented to TEL applications facilitating reusability
and repurposing of learning objects. The approach is
based on analyzing and exploiting the advantages of
SWS technology in the automation of learning
object discovery, selection and composition within a
distributed service architecture seamlessly integrated
through ontologies.
One of the most promising approaches for
creating Learning Management Systems using SWS
architecture has been proposed in the frame of the
LUISA project. The approach has reused and
adapted the SWS Framework previously developed
in the frame of another IST project – DIP, which
was implemented in parallel with INFRAWEBS
Project. Being based on the same SWS
methodological framework (WSMO), the projects
have created different SOA-based architectures
caused by the different objectives to be solved.
Our approach aims at developing a new domain-
specific Semantic SOA-based architecture, which
will be focused on further development and refining
the INFRAWEBS Framework for semantic service
engineering that covers the whole semantic Web
service life-cycle and allows creation of complex
semantically-enabled applications (Agre et al. 2009).
The implementation of the proposed approach
needs intensive research in several directions:
• Developing new application-oriented methods
and end-user oriented tools for description of
SWS and Goals. Problems for representation of
Semantic Web Services and Goals are among
the hottest research aspects of SWS technology
for which no general solutions have been found
yet. The efforts will be oriented to refining and
advancing the original graphical and ontology-
based approach developed in the frame of
INFRAWEBS project (Agre and Dilov 2007).
• Developing new methods for dynamic
composition of Semantic Web Services suited
for TEL. The problem of dynamic composition
of Web services is the core of the Service-
oriented computing paradigm. There are a very
few really implemented and rather restricted
approaches for this still open problem. Most of
them are based on OWL-S Framework. Our
research on dynamic composition of SWS will
be oriented to further development of an
original data-driven SWS composition
approach, which was recognized as “the only
fully automated Functional Level-based
Composition planner for WSMO yet”.
ACKNOWLEDGEMENTS
This work is partially funded by Bulgarian NSF
under the project D-002-189 SINUS “Semantic
Technologies for Web Services and Technology
Enhanced Learning”.
REFERENCES
Agre G., T. Pariente Lobo, Z. Marinova, J. Nern, A. Mic-
sik, A. Boyanov, T. Atanasova, J. Scicluna, J. López-
Cobo, E. Tzafestas. 2009. INFRAWEBS – A Frame-
work for Semantic Service Engineering. Chapter 12
In: E. di Nitto, A.-M Sassen (Eds.) At your service:
Service Engineering in the Information Society
Technologies Program. The MIT Press, Cambridge,
Mass. USA.
Agre G. and I. Dilov, 2007. How to Create a WSMO-
Based Semantic Service without Knowing WSML. In:
LNCS 4832, pp. 217–235, 2007. Springer-Verlag
Berlin Heidelberg.
Aceto S., C. Delrio, C. Dondi (Eds.), 2007. e-Learning for
Innovation. Executive Summary of Helios Yearly Re-
port,2007, ISBN: 2-930429-13-5, p.25.
Andonova G., G. Agre, H.-Joachim Nern, A. Boyanov.
2006. Fuzzy Concept Set Based Organizational
Memory as a Quasi Non-Semantic Component within
the INFRAWEBS Framework. In: Proc. of
IPMU2006, July 2-7, Paris, France, pp. 2268-2275.
Arapi P., Moumoutzis N., Mylonakis M., Christodoulakis
S., 2008. Design, Implementation and Experimental
Evaluation of a Pedagogy-driven Framework to Sup-
port Personalised Learning Experiences. In: P. Stock-
inger, D. Dochev (eds.) Proc. of 2th LOGOS
Workshop “Cross-Media and Personalised Learning
Applications with Intelligent Content” Varna, 2008,
pp. 7-25.
Arroyo S. 2006. A Semantic Service-Based Micro-
Learning Framework. In : T. Hug, M. Lindner, P. A.
Bruck (eds.) Micromedia & e-Learning 2.0: Gaining
the Big Picture. Proc. of Microlearning Conference
2006, Innsbruck University Press, pp. 260-272.
Dietze, S., Gugliotta, A., Domingue, J., 2007. Towards
adaptive E-Learning Applications based on Semantic
Web Services. In: Proc. of TENCompetence Workshop
on Service Oriented Approaches and Lifelong Compe-
tence Development Infrastructures, January, 2007,
UK.
Fensel D. and C. Bussler 2002. The Web Service
Modeling Framework WSMF, Electronic Commerce:
Research and Applications, 1(2): 113-137.
TOWARDS SEMANTIC WEB ENHANCED LEARNING
217