INTEGRATING SEMANTIC WEB REASONING INTO

LEARNING OBJECT METADATA

Shang-Juh Kao, I-Ching Hsu

Department of Computer Science, National Chung-Hsing University, Taiwan

Keywords: LOM, Semantic Web, SCORM, OWL.

Abstract: One of important functions of Learning Object Metadata (LOM) is to associate XML-based metadata with

learning objects. The inherent problem of LOM is that it’s XML specified, which emphasizes syntax and

format rather than semantic and knowledge representation. Hence, it lacks the semantic metadata to provide

reasoning and inference functions. These functions are necessary for the computer-interpretable descriptions

that are critical in the reusability and interoperability of the distributed learning objects. This paper aims at

addressing this shortage, and proposes a multi-layered semantic framework to allow the reasoning and

inference capabilities to be added to the conventional LOM. To illustrate how this framework work, we

developed a Semantic-based Learning Objects Annotations Repository (SLOAR) that offers three different

approaches to locate relevant learning objects for an e-learning application - LOM-based metadata,

ontology-based reasoning, and rule-based inference.

1 INTRODUCTION

The Sharable Content Object Reference Model

(SCORM) (SCORM, 2004) is developed and

extended based on IEEE Learning Object Metadata

(LOM) (IEEE LOM, 2002). The LOM in SCORM is

used to describe SCORM-compliant learning objects

in a consistent fashion such that they can be

identified, categorized, searched for and discovered

within and across systems to further facilitate

sharing and reuse.

The inherent problem of LOM is that it is based

on XML, which lays stress on syntax and format

rather than semantic and knowledge representation.

Hence, LOM exhibits the advantage of data

transformations and digital libraries, but it lacks the

semantic metadata to provide reasoning and

inference functions. These functions are necessary

for the computer-interpretable descriptions, which

are critical in the area of learning objects reusability,

autoexec course generation, dynamic course

decomposition, learning object mining, etc.

To this problem, a mapping from LOM to

statements in an RDF model has been defined

(Nilsson, 2003). However, RDF alone doesn't share

some basic common structures that help to describe

classes of learning objects and types of relationship

between learning objects. Thus, we need more

flexibilities and facilities for expressing meaning

and semantics than what has in RDF. The Semantic

Web provides a catalytic solution to this problem.

To enhance the knowledge representation of the

XML-based markup language, the traditional

Semantic Web approach is to upgrade the original

XML-based to ontology-based markup language.

The upgrade mentioned above from XML-based

LOM to RDF-based LOM is an example. The main

problem with this approach is that the original

XML-based markup language is replaced with the

new ontology-based markup language. In this paper,

we propose a novel integration approach that

combines the first four layers of Semantic Web

stack, including URI layer, XML layer (LOM),

ontology layer, and rule layer. In the multi-layered

semantic framework, Semantic Web technologies

can be integrated with LOM to enhance the

computer reasoning, and the original LOM still exist

to cooperate with ontologies and rules. That is, the

multi-layered semantic framework does not change

the original schema of LOM. Hence, the existing

LOM and SCORM metadata documents can

continue to be used.

In order to demonstrate the feasibility of this

multi-layered semantic framework, an application

system of the Semantic-Based Learning Objects

Annotations Repository (SLOAR) is developed to

dynamically provide the information of relevant

learning objects for course creators. SLOAR

supports three different approaches to finding

relevant learning objects, including LOM-based

metadata, ontology-based reasoning, and rule-based

inference. Such dynamic finding is desirable for a

number of reasons. Firstly, it is customized for each

181

Kao S. and Hsu I. (2006).

INTEGRATING SEMANTIC WEB REASONING INTO LEARNING OBJECT METADATA.

In Proceedings of the Eighth International Conference on Enterprise Information Systems - SAIC, pages 181-184

DOI: 10.5220/0002445101810184

 SciTePress

individual learning object, based on what metadata

and knowledge the learning object has shown so far.

Secondly, because the content or category of a

learning object may keep changing, dynamic finding

provides more up-to-date suggestions than a static

design. Thirdly, as the number of learning objects

may be large, adding suggestion links may become

cumbersome for the course developer. Lastly, it can

also be used at run-time to help in the decision of

what content model component to deliver to the

learner.

This paper is organized as follows. The next section

presents the multi-layered semantic framework.

Section 3 gives the architecture of SLOAR. In

Section 4, we illustrate how the SLOAR can employ

Semantic Web technologies to provide different

approaches for finding relevant learning objects.

2 MULTI-LAYERED SEMAMTIC

FRAMEWORK FOR LOM

In this paper, we propose a novel integration

approach to combine the first four layers of

Semantic Web stack, including URI layer (learning

objects), XML layer (Metadata, LOM), Ontology

layer (OWL), and Rule layer (Rule Markup

Language, RuleML), as shown in Figure 1.

The URI layer is composed of learning objects,

which can be identified by URI. The XML layer is

composed of LOM metadata that are XML-based

metadata for describing learning objects. The

Ontology layer provides OWL-based ontologies,

which can enhance LOM to Semantic-based

metadata, hence, improve reasoning capabilities of

LOM. The Rule layer supports more complex

inference than the Ontology layer, and builds

RuleML rules on top of OWL ontologies. The

ontology is based on description logics to provide

sound and decidable reasoning. In contrast, the rule

is a logic program, which can complement ontology

to support more complex rule-based inferences.

ml:XML

ml:HTML

ml:XHTML

Markup Ontology

property

ml:application

cu-1 cu-2

if jm:XMLParser(ml:XML,ml:JAXP) and jm:using(jm:JAXP,jm:DOM)

then jm:treeMode(ml:XML,jm:DOM)

cu-4

Ontology Layer

(OWL-based

ontology base)

Rules Layer

(RuleML)

XML Layer

(SCORM-LOM

classification,

relation

Metadata)

learning object ID relationclass

ml:standard

instance

URI Layer

(Learning

Objects)

JAXP for XML XML Advance

XHTML

Introduction

entity

http://../jaxp.htm http://../xml.htm http://../html.htm

Java DOM

http://../jdom.htm

cu-3

ml:isversionOf

ml:XMLParser

Java Ontology

jm:treeMode

jm:DOM

jm:Java

jm:API

jm:JAXP

jm:API

jm:using

learning object

Figure 1: Multi-layered semantic framework.

3 SYSTEM ARCHITECTURE

The basic function of SLOAR is to provide the

information of relevant learning objects for course

creators. It supports three different approaches for

finding relevant learning objects, including LOM-

based metadata, ontology-based reasoning, and rule-

based inference. In SLOAR, each learning object is

associated with a classification metadata to quote

extra semantic from a specific ontology class and is

associated with a relation metadata to quote extra

semantic from a specific ontology property. These

ontologies are implemented in OWL that can be

integrated into LOM, and as a result, the semantic

capabilities of LOM were greatly improved.

The core components of SLOAR include the

annotation base, knowledge base, search agent, and

inference agent. The flow-oriented SLOAR

architecture is depicted in Figure 2, as described in

the following:

˙ Annotations base: is a learning object

annotations repository that is composed of LOMs. A

LOM is an XML document containing a set of

markup elements to describe the learning objects..

˙ Knowledge base: is developed by the Semantic

Web standard to support reasoning tasks. The

knowledge is grouped into two categories: ontology

layer inference using OWL-based ontologies and

rules layer inference using RuleML logic program.

˙ Search agent: is a search engine that supports for

a XPath query on the learning objects metadata base.

˙ Inference agent: is an intelligent agent that is

implemented based on a JESS-based rule engine

(JESS, 2005) and supports a XSLT processor.

Figure 2: SLOAR Architecture.

The information flow of the SLOAR, as listed in

the Figure 2, can be realized and started as follows.

1. The requester sends a learning object with URL

to the search engine.

2. This step is the LOM-based metadata approach.

The search agent relies on the request to query

the LOM Base to finding all relevant LOM-

ICEIS 2006 - SOFTWARE AGENTS AND INTERNET COMPUTING

182

based metadata documents of the learning

object.

3. The search engine sends these LOM-based

metadata documents to invoke the inference

agent.

4. This step is the ontology-based reasoning

approach. The inference agent conducts the

following tasks to infer for semantic links.

4.1 It retrieves and parses the relevant OWL-based

ontologies quoted by the classification and

relation tags.

4.2 It utilizes the OWL2Jess.xsl XSLT style sheet

(OWL2Jess, 2005) for transforming these

semantics of OWL-based ontologies into JESS-

based rules.

4.3 It infers the semantic relevant learning objects

from these JESS-based rules.

5. The rule-based inference is performed. The

inference agent conducts the following tasks to

infer the relevant learning objects.

5.1 It relies on the relevant ontologies, mentioned

on the step 4.2, to query the rule base to retrieve

relevant RuleML-based rules.

5.2 It utilizes the RuleMLTransform.xsl XSLT style

sheet (RuleML2Jess, 2002) for transforming

these RuleML-based rules into JESS-based

rules.

5.3 It infers the rule relevant learning objects from

these JESS-based rules.

6. Finally, the inference agent passes the

information of relevant learning objects,

including LOM-based, ontology-based, and

rule-based learning objects to the requester.

4 USAGE SCENARIO OF SLOAR

To explicitly demonstrate how SLOAR works, a

usage scenario of locating relevant learning objects

is presented in the following.

4.1 Multiple-Layered Conceptions

Figure 1 provides a concrete example of how

multiple-layered semantic framework is employed

by SLOAR. In the URI layer, there are a number of

learning objects, including XML Advance, XHTML

Introduction, JAXP for XML, Java DOM, etc. In the

XML layer, each learning object is described with a

LOM that consists of classification metadata and

relation metadata.

The user interface of SLOAR is shown in the

Figure 3. A course creator selects a learning object

and then presses the "Query" button. The SLOAR

will rely on the learning object to invoke search

agent and inference agent to produce information of

relevant learning objects. In the next section, we will

base on this example to explain SLOAR how to

support different approaches to finding relevant

learning objects.

Figure 3: A query result on SLOAR with course ID is cu-

4.2 Different Approaches

When the inference agent receives LOMs from the

search agent, it sequentially invokes different

approaches for locating relevant learning objects.

4.2.1 LOM-based Metadata

According to the learning object ID (i.e. cu-1)

received, the search engine finds all relevant LOMs

in the annotations base. Since all LOMs are actually

XML documents, this corresponds to performing an

XPath query on each LOM, looking for learning

object whose identifier has the same value as "cu-1".

The search results consist of two LOMs. The former

is the cu-1's LOM that consists of an outbound link

from cu-1 to cu-4, as shown in Figure 4. The latter is

the cu-2's LOM that consists of an outbound link

from cu-2 to cu-1. The LOM-based approach only

depends on the cu-1's LOM that exhibits a number

of metadata. First, the file element describes the

URL address of cu-1. Second, the classification

element is used to describe where cu-1 quotes a

particular ontology class. Third, the relation element

is used to describe features that define the

relationship between cu-1 and other learning objects.

For example, the kind element describes where the

relationship quotes a particular ontology property,

and the resource element describes where cu-1 links

to a particular learning object.

This kind of approach is that directly extracts

data from the original LOM to produce the relevant

information of learning objects, so we call it LOM-

based metadata approach. The inference agent

extracts data from the relation metadata of cu-1 to

show that there is an XMLParser relation from cu-1

to cu-4. The output result of ontology-based

reasoning is shown in (A) of Figure 3.

INTEGRATING SEMANTIC WEB REASONING INTO LEARNING OBJECT METADATA

183

<metadata><lom>……………….

<value>http://…/markup.owl#XML</value></purpose>

</classification>

<value>http://…/java.owl#XMLParser</value></kind>

<catalog>learning object ID</catalog>

</resource></identifier></relation> ……………

</lom></metadata></resource>

Figure 4: The partial LOM code of the learning object cu-

4.2.2 Ontology-based Reasoning

The inference agent depends on semantics of

Markup ontology and cu-2's LOM to reason the

following facts.

1. The cu-2 learning object is an instance of

XHTML class.

2. There is a standard relation from cu-2 to cu-1.

3. The application property is an inverse of

standard property (see Figure 1).

Base on the above facts, inference agent can reason

that there is an application relation from cu-1 to cu-

2. The output result of ontology-based reasoning is

shown in (B) of Figure 3.

4.2.3 Rule-based Inference

This inference agent relies on the previous inference

results, LOM-based metadata documents and

RuleML rules to perform the following tasks.

1. It converts the LOMs to JESS-based facts, as

shown in Figure 5.

(assert (triple (predicate "http://../markup.owl#standard")

(subject "cu-2") (object "cu-1")))

(assert (triple (predicate "http:/./markup.owl#XMLParser")

(subject "cu-1") (object "cu-4")))

(assert (triple (predicate "http://../java.owl#using")

(subject "cu-4") (object "cu-3")))

Figure 5: The JESS-based facts.

2. It converts the RuleML rule (see Figure 1) to

JESS-based rule, as shown in Figure 6.

(defrule XMLparserMode

(triple (predicate "http://../DO/markup.owl#XMLParser")

(subject ?x) (object ?y))

(triple (predicate "http://…/DO/java.owl#using")

(subject ?y) (object ?z))

(assert

(triple (predicate "http://…/DO/markup.owl#treeMode")

(subject ?x) (object ?z))))

Figure 6: The JESS-based rule.

3. It relies on the above JESS-based facts and rule

to infer the rule-based learning objects. The

inference can infer that there is a treeMode

relation from cu-1 to cu-3. The output result of

rule-based reasoning is shown in (C) of Figure

5 CONCLUSION

In this paper, an intelligent SLOAR prototype

system is implemented. SLOAR is developed based

on multi-layered semantic framework, including

URI layer (learning objects), XML layer (LOM),

Ontology layer (OWL), and Rule layer (RuleML).

This framework is embedded into the Semantic Web

stack and does not change the original schema of

LOM. It results in making LOM computer-

interpretable and hence enables automatically

relevant learning objects finding.

Novel Semantic Web solutions, integrated with

different types of high-level ontology-based

metadata and XML-base rules, can dynamically

tailor the knowledge base to take into account the

user preferences for personalization. Thus, one

future work is to extend the accessibility of the

SLOAR towards the personalization model for

individual-dependent dynamic courses according to

user preferences.

REFERENCES

IEEE LOM, 2002. Retrieved Sep. 18, 2005, from

http://ltsc.ieee.org/wg12/

JESS, 2003. Retrieved Sep. 18, 2005, from

http://herzberg.ca.sandia.gov/jess/

Nilsson, M., et al., 2003. IEEE LOM RDF binding,

Retrieved Sep. 18, 2005, from

http://kmr.nada.kth.se/el/ims/metadata.html

SCORM: Sharable Content Object Reference Model, 2004.

Retrieved Sep. 18, 2005, from http://www.adlnet.org

OWL2Jess, 2005. Retrieved Sep. 18, 2005, from

http://www.inf.fu-berlin.de/inst/ag-

nbi/research/owltrans/OWL2Jess.xsl

RuleML2Jess, 2002. Retrieved Sep. 18, 2005, from

http://www.ruleml.org/jess/RuleMLTransform.xsl.

ICEIS 2006 - SOFTWARE AGENTS AND INTERNET COMPUTING

184