A Semantic Web Service Description of Learning Object

Chaker Ben Mahmoud

, Ikbel Azaiez

, Fathia Bettahar

and Faïez Gargouri

IResCoMath, University of Gabès, Gabès, Tunisia

ISIMS, University of Sfax, Sfax, Tunisia

Keywords: e-Learning, Learning Paths, Learning Service, Ontology, Semantic Web Services, OWLS, Web Services.

Abstract: How to find and identify learning objects according with the learner profile represents a main interest in the

quality of learning process. Thus, using the paradigm of Semantic Web Services ensure the independence

and reusability of learning object in a different context. In this paper, we propose an extension of OWLS

that encompass the description of the learning intention and the context of use that characterize a learning

object. We also describe the generic scenario of the publication and discovery process.

1 INTRODUCTION

Recent years, several works have dealt on adaptation

and personalization of learning content. In this

context, the learning objects became a fundamental

element to develop educational contents.

In fact, learning objects include several

problems. First, they cannot be effectively reused

because they are distributed between several places

and depends on the learning system. Besides, the

learning objects have a limit of cooperation which

presents a low probability of binding between

objects.

Therefore, many rules and metadata standards

have been proposed as a solution to overcome the

problem of accessibility and interoperability of

learning objects, and a lot of norms and standards

were created to achieve this. In this stage, several

standardization efforts have been launched including

LOM and SCORM (Lee et al., 2006). These

standard descriptions of learning resources focus on

the characterization of content rather than on its use.

In fact, the above standards have limitations in

the context of heterogeneous learning objects. In

addition, the definition of specialized courses

according to desired skills requires a composition of

learning objects to provide the learner with a

personalized learning course. The problems of

interoperability, reusability and composition of

learning content can be solved by using the

principles of Web service paradigm. Web services

are defined as open standards that provide a flexible

solution for integrating heterogeneous and dynamic

applications that enable interoperability between

different systems.

In this paper, we propose a semantic description

of learning services that encompasses the description

of the learning intention and the use of context that

characterizes a learning object. Then, we propose a

semantic service descriptor, based on our OWLS

extension, to enrich service registry.

The rest of the paper is organized as follows: In

Section II we present some related works. Then, we

give an overview of our approach in section III. We

describe, in section IV, the learner profile in the

form of ontology where it specifies four basic

concepts. In Section V we propose an extension of

OWLS to support learning object. In Section VI we

present the principle of learning semantic web

service publication and discovery. Finally, we finish

with a conclusion and some remarks/hints about

future work.

2 RELATED WORKS

(Padron et al., 2004) present a learning web services

framework to support the integration of newer,

complex learning processes and the dynamic

generation of content based communities of interest.

This framework is based on two elements of

construction: the first is The Learning Web Services,

supported by a basic Web Service architecture,

which allows to create, define and publish learning

objects that encapsulate different learning processes;

and the second is Learning Web Services

Mahmoud, C., Azaiez, I., Bettahar, F. and Gargouri, F..

A Semantic Web Service Description of Learning Object.

In Proceedings of the 7th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (IC3K 2015) - Volume 2: KEOD, pages 195-201

ISBN: 978-989-758-158-8

195

Composition, supported by an architecture for

services composition that allows to look for,

integrate, execute and redefine the learning

processes.

(Gutiérrez-Carreón et al., 2009) is interested to

the semantic description of services to ease the

discovery learning services based on semantic

matching process between educational service

features and user needs. To do this, each device is

controlled by a computer with internet connection.

The user can control the devices that are connected

to computers and the acquisition of data stored in

databases. The implementation of these features is

based on the semantic web, particularly on the use of

ontologies and metadata to annotate learning

services. Indeed, this system uses three types of

ontology domain ontology, an ontology representing

the goals and an ontology describing the learning

services. Ontology learning services is described

using the terms in WSMO (Web Service Modelling

Ontology). The ontology of the objectives described

aspects related to user needs. The domain ontology

defines the terminology and concepts of the subject

area that are used to describe the relevant aspects of

the objectives.

(Cho et al., 2008) focus on the description of the

context of services to adapt learning services to the

user. Moreover, taking into account the context of

the adaptation of services is based on a set of rules.

These are predicates that combine contextual

information and service descriptions to check their

relevance to a particular situation.

(Zniber et al., 2010) presented an approach to

build personalized pathways called POPS (Process-

Oriented Pedagogic Service) by composing services

dynamically. This approach is a conceptual

framework that defines a model for describing the

pedagogical services. This model of Pedagogical

Service provides a set of concepts to describe the

services. According to Zniber, a pedagogical service

is composed of three parts: "profile”, "structure" and

"behavior". The "Profile" describes the general

appearance of the pedagogical service. It

corresponds to the service interface and will be used

when searching for a match between the available

services and the learners’ intentions. The "structure"

part describes the organization of the process to

achieve the pedagogical objective. It is defined by a

process, an initial position and a final position. The

"behavior" element is the "executable" level of

service. It describes the use of the service by a

learner and it takes the form of an implementation

plan with activities and resources to be mobilized.

3 PROPOSED APPROACH

3.1 Motivation

The development of learning systems aims to

provide learners with courses adapted to their needs

and their profile. The challenge therefore is to make

the system more responsive to the request of the

student is based on learning object scattered on

several platforms.

In this context, we consider a learning system as

a set of Learning Semantic Web services where each

service represents a learning object that describes an

intention and context of use. It’s by composing

dynamically services learning that possible to build

custom course adapted to a given profile. The

description of these services and the formulation of

the request of learner are made by two ontologies:

objectives ontology and ontology of the domain

learning.

3.2 Overview of Our Approach

The architecture, shown in Figure 1, represents our

e-learning approach to provide learners with

learning paths adapted to their requests (Ben

Mahmoud et al., 2015), (Ben Mahmoud et al., 2014).

This approach consisted of three components:

 Learning data representation component of

learning in order to achieve the learner's learning

objectives.

 Formulation component of the learner Query

 The building component of the learning path that

satisfied a particular objective set by the learner.

Figure 1: Components of the learning approach

architecture.

In our proposed approach, we use ontologies

both for a semantic description of learning services

and to make easy their research and composition to

generate personalized learning paths. They are also

KEOD 2015 - 7th International Conference on Knowledge Engineering and Ontology Development

196

used for sharing and reuse of learning objects.

Indeed, the same ontologies are used to semantically

describe the learning services (point of view of the

course designer) and to define the learning

requirement (point of view of the learner). Thus, we

have used a domain ontology and objectives

ontology. The first is used to represent knowledge

about the domain of education and the second is to

provide knowledge about the learning objectives.

4 LEARNER PROFILE

In the learning systems, the learner model is

essential for the generation of personalized paths. It

is to take into consideration the concepts of learning

which the system must adapt. These concepts can be

different from one learner to another. At this stage,

there are an adaptable model and an adaptive model.

The adaptable model is modified only by the learner,

while the adaptive model is changed by the system

according to the learner's paths.

The term "model" is used to describe

informations related to learners. The authors use this

term to express the way to represent learners'

knowledge in a given system that to describe

knowledge of a particular learner in this system.

However, in our work, we are interested only to the

data of a learner and not their construction process.

For this reason, we prefer to use the term "profile" of

the learner, in order to separate this object of how to

create it.

Several standards describe the model of the

learner such as PAPI, IMS-LIP. Our learner profile

is based on the IMS-LIP standard (IMS-LIP, 2001),

which is a proposal of the consortium "IMS Global

Learning" made primarily to meet the need for

standardization of data relating to the description of

learners different learning systems. As the majority

of learning systems, knowledge learning expresses

the competence, educational goals, history learning

and preferences.

We describe the learner profile in the form of

ontology where it specifies four basic concepts

(existing in the IMS-LIP): Identification, Affiliation,

Accessibility and Competence (Figure 2).

 Accessibility: describes the general accessibility

such as: language skills, disabilities, eligibility

requirements and learning preferences.

 Competence: describes the Competence,

experience and knowledge.

 Affiliation: provides information on membership

in professional organizations.

 Identification: describes the demographic and

geographic data on the learner (name, age,

address, email, etc.)

Figure 2: Learner Profile.

5 OWLS-LO EXTENSION

In this section, we describe our proposed extension

to the OWLS recommended (Wang et al., 2013).

This extension allows the description of the

educational aspect of service learning in ontology

"ServiceLearning". It corresponds to the service

interface and is used when searching for a match

between the semantic learning services available on

the one hand, and the requests expressed by the

learners, on the other hand.

Figure 3: The ServiceLearning Ontology.

The Figure 3 illustrates our Learning extension

of OWLS. The proposed property “provide” is a

property of Service. The class "ServiceLearning"

corresponds to the respective range of this property.

Each instance of Service will provide a

ServiceLearning description. The ServiceLearning

represents the information needed to discover the

A Semantic Web Service Description of Learning Object

197

appropriate service in order to satisfy a specific

learning need.

This extension is composed of three basic

concepts: Leaning intention, context of use, and

required services (see Figure 4).

Figure 4: OWLS extension to represent LO.

5.1 The Intention

The intentional vision places the concept of service

to a higher level of abstraction where the service is

designed to lead to the satisfaction of a user’s

intention. This intention is what the user awaits in

performing a service (Jackson, 1995). Our

description of a service is as the learning intention

allowed defining the finality of service, without

going into the details of its use. It expressed an

intention that the learner sought to achieve. In this

context, Prat (Prat, 1997) have proposed a model for

the concept of intention which is derived from the

linguistic approach and inspired by the case

grammar of Fillmore (Fillmore, 1968) and

extensions of Dik (Dik et al., 1989).

The intention, in this model, is represented by a

verb, targets and different parameters that play

specific roles in relation to the verb. The verb

describes the action of the realization of the

intention, while the target is affected by the object

embodiment of the intention. The parameters (way,

direction, quantity and quality) are used to clarify

and express additional information.

In our approach, the intention was defined by a

learning objective (verb) and a concept of learning

domain (Target). The concept learning objective

depicted the types of learning objectives in

accordance with Bloom's taxonomy [bloom]. They

are expressed in terms of goals and organized in

levels. The definition of the objective falls within the

ontology of learning objectives. The learning

concept domain indicated the target of the learning

intention. The specification of the concept used the

terminology defined in the ontology of the

educational domain (Figure 5).

Figure 5: Intention Code.

For example, for the OWLS-LO (01) Service

("Define Class"), was characterized by the learning

objective "Define" defined in the ontology of

learning objectives and the concept of learning

"class" defined in the ontology of the educational

domain "Java".

5.2 The Context

The context provides a description of the

pedagogical aspect of the learning object as well as

the learning situation in which the service can be

used. To describe this aspect, we based on the

descriptions of the IEEE LOM (IEEE, 2007). Thus,

this context was a selection of properties of the

LOM allowing indexing learning objects

semantically and describing mainly the container but

not its contents.

It's represented by three types of knowledge:

affiliation, accessibility and competence. First, the

Affiliation describes the level of targeted studies for

learning object. Secondly, the accessibility mainly

describes the language in which the resource was

presented. Finally, the competence presents the level

of difficulty of this Learning Object relative to the

target audience: easy, medium, or difficult as it is

illustrated in Figure 6.

Figure 6: Context Code.

KEOD 2015 - 7th International Conference on Knowledge Engineering and Ontology Development

198

For example, for the OWLS-LO (01) Service

("Define Class"), referred to a learning object

whose context was characterized by a "University"

affiliation, language was "Fr" and competence

"Easy."

5.3 The Required Services

The required services are all the knowledge required

so that the learner can use the learning object. This

notion can describe a navigation strategy among the

reused learning objects. Indeed, for a learner, to

access the contents of an object check that it has a

body of knowledge needed to tackle it. Therefore,

the requirement for each service must be satisfied. In

our extension, we define the services required as the

set of coupled {Concept learning domain and

Concept learning objective} needed to use the

service learning during any one specifying the

weighting of each. (Figure 7)

Figure 7: Required service code.

For example, for the OWLS-LO (01) Service

("Define Class"), we could define two required

services: {Define, Attribute} with a weighting of 0.5

and {Define, Method} with a weighting of 0.5.

6 OWLS-LO EXTENSION

IMPLEMENTATION

6.1 Process of Publishing

In this section, we present the generic scenario of the

publication process (Figure 8). The instruction

designer presents the learning object to be

interoperable.

After the generation of the learning Web service,

appropriate to learning object, we passed to semantic

description of learning web service by using existing

ontologies. Finally, the learning Web service and

OWLS-LO were published in UDDI registry and

OWLS-LO repository, respectively.

6.2 Process of Discovery

After publishing the learning semantic Web services

both in OWLS-LO repository and in UDDI registry,

we proceeded to discover them. Thus, we present the

generic scenario of the discovery process (Figure 9).

The learner used Semantic Description Query to

describe their intention (Semantic Request) of

learning through browsing ontologies (Domain

ontology, Ontology of objective).

Once the request is submitted, the building path

module (Step of Construction of Learning Path)

extracts the intention and profile of learner and

proceeds, thereafter, to seek (step of Semantic

Matching) appropriate services allowing the

generation of a learning path. Indeed, this semantic

matching similarity proceed to match the intention

of learner with the intention learning of OWLS-LO,

on the one hand, and to match the use of context of

OWLS-LO with learner’s current profile, on the

other hand. After that, we calculate the importance

factor of all required services. Therefore, we select

the OWLS-LO having the highest score.

Finally, we presented a learning path to the

learner according to their goal.

Figure 8: Process of publishing.

A Semantic Web Service Description of Learning Object

199

Figure 9: Process of discovery.

6.3 Semantic Matching OWLS-LO

This section gives a global description of the

discovery process (Figure 10), starting from the

query submission to the OWLS-LO replies, by

emphasizing the main steps related to this

matchmaking.

Figure 10: OWLS-LO Discover Mechanism.

When the learner presented their request based

on an intention to be satisfied, the discovery process

was started. The discovery mechanism loaded all

OWLS-LO semantic description of the services and

launched the matching. In a first step, we proceeded

to match the learner’s intention with the intention

that the OWLS-LO service satisfied. Then we

calculated the importance factor of each service

required for this selected OWLS-LO. In the end, we

matched the educational context services with

learner’s current context. After getting a list of the

most appropriate OWLS-LO, we selected the service

having the highest score matching.

7 CONCLUSIONS

In this paper, we proposed an approach defined a

learning Web service for each learning object to

overcome the problems of interoperability and

accessibility of learning objects. This web service is

represented by an extension of OWLS composed of

three basic concepts: learning intention, use of

context and required services. This extension will be

used when searching for a match between the

semantic learning services available on the one

hand, and the requests expressed by the learners

according on her profile, on the other hand.

In order to appreciate the usefulness and the

efficiency of our approach, we intend to make the

description more meaningful and the service

discovery more precise and appropriate to the

learner‘s needs. Also, we expect to evaluate our

service discovery mechanism in a more interesting

scenario in future work.

REFERENCES

Lee, M., Tsai, J., Wang, T., 2006. An Ontological

Approach for Semantic-Aware Learning Object

Retrieval. In Sixth International Conference on

Advanced Learning Technologies, pp.208-210.

Padron, C. L., Torres, J., Dodero, J. M., Diaz, P., Aedo, I.,

2004. Learning Web services composition and learner

communities support for the deployment of complex

learning processes. In IEEE International Conference

on Advanced Learning Technologies. pp. 390-394.

KEOD 2015 - 7th International Conference on Knowledge Engineering and Ontology Development

200

Gutiérrez-Carreón, G., Daradoumis, T., Jorba, J., 2009.

Automatic composition of Learning Grid Portlets: a

comparison of syntactic and semantic approaches. In

International Journal of Grid and Utility Computing

(IJGUC), vol.1, no.4, pp.308-315.

Cho, D. J., Woo Hong, M. W., 2008. A Design of

Ontology Context Model in Ubiquitous Learning. In

12th WSEAS International Conference on

COMPUTERS, pp.844-848.

Zniber, N., 2010. Service-Oriented Model for Personalized

Learning Process Desig. These, Université Paul

Cézanne - Aix-Marseille III.

Ben Mahmoud, C., Azaiez, I., Bettahar, F., Abel, M.,

Gargouri, F., 2015. A Learning Semantic Web Service

for Generating Learning Paths. 14th IEEE/ACIS

International Conference on Computer and

Information Science, ICIS 2015, Las Vegas, NV, USA,

June 28 - July 1, 2015, pp 627—631.

Ben Mahmoud, C., Bettahar, F., Abel, M., Gargouri, F.,

2014. A New Approach Based on Learning Services to

Generate Appropriate Learning Paths", In 16th

International Conference on Enterprise Information

Systems, Portugal, pp.643-646.

IMS-LIP. (2001). IMS Learner Information Packaging

Information Model Specification, Retrieved 24 June,

2014,

http://www.imsglobal.org/profiles/lipinfo01.html.

Wang, H., Gibbins, N., Payne, T., Patelli, A., 2013. A

Survey of Semantic Web Services Formalisms. In

Ninth International Conference on Semantics,

Knowledge and Grids, pp.135–142.

Jackson, M., 1995. Software requirements specifications: a

lexicon of practice, principles and prejudices. ACM

Press/Addison-Wesley Publishing. New York, NY,

USA, 228 p.

Prat, N., 1997. Goal Formalisation and Classification for

Requirements Engineering. In third International

Workshop on Requirements Engineering: Foundation

for Software Quality.

Fillmore, C. J., 1968. The case for case. In Bach et

Harms", eds. (New York: Holt, Rinehart, and

Winston), Universals in Linguistic Theory, 88 p.

Dik, S. C., 1989. The theory of functional grammar. (Foris

Publications). Dordrecht, Holland ; Providence, RI,

U.S.A.: Foris Publications, p. 497.

IEEE, 2007. Draft standard for Learning Object Metadata

(LOM) (IEEE 1484.12.1-2002), IEEE Learning

Technology Standards Committee (LTSC).

A Semantic Web Service Description of Learning Object

201